1 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
4 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
6 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
8 * config/arm/arm.opt (mlra): New option.
9 * config/arm/arm.c (arm_lra_p): New function.
10 (TARGET_LRA_P): Define.
12 2013-10-16 Paulo Matos <pmatos@broadcom.com>
14 * tree-core.h (tree_code_name): Remove.
15 * tree.h (get_tree_code_name): New prototype.
16 * tree.c (tree_code_name): Make static.
17 (get_tree_code_name): New function.
18 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
19 tree_class_check_failed, tree_range_check_failed,
20 tree_not_class_check_failed, omp_clause_check_failed,
21 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
22 wrapper get_tree_code_name instead of calling tree_code_name directly.
23 * tree-vrp.c (dump_asserts_for): Likewise.
24 * tree-dump.c (dequeue_and_dump): Likewise.
25 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
26 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
27 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
28 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
29 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
30 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
31 dump_gimple_omp_for): Likewise.
32 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
33 * tree-ssa-pre.c (print_pre_expr): Likewise.
34 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
35 * print-tree.c (print_node_brief, print_node): Likewise.
36 * gimple.c (gimple_check_failed): Likewise.
37 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
38 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
39 * config/mep/mep.c (mep_validate_vliw): Likewise.
40 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
41 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
43 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
45 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
48 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
50 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
51 (MULTILIB_MATCHES): Add multilib for -march=v8.
53 2013-10-15 Sriraman Tallam <tmsriram@google.com>
56 * optc-save-gen.awk: Add extra parameter to the save and restore
58 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
59 * tree.c (build_optimization_node): New parameter. Add extra parameter
60 to call to cl_optimization_save.
61 (build_target_option_node): New parameter. Add extra parameter
62 to call to cl_target_option_save.
63 * tree.h (build_optimization_node): New parameter.
64 (build_target_option_node): New parameter.
65 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
66 build_optimization_node and build_target_option_node.
67 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
68 (handle_pragma_push_options): Ditto.
69 * toplev.c (process_options): Ditto.
70 * opts.c (init_options_struct): Check for opts_set non-null.
71 * target.def (target_option.save): New parameter.
72 (target_option.restore): New parameter.
74 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
75 (ix86_pragma_target_parse): Ditto.
76 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
78 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
79 to build_optimization_node and build_target_option_node.
80 (rs6000_valid_attribute_p): Ditto.
81 (rs6000_pragma_target_parse): Ditto.
82 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
85 TARGET_64BIT_P: New Macro
86 TARGET_MMX_P: New Macro.
87 TARGET_3DNOW_P: New Macro.
88 TARGET_3DNOW_A_P: New Macro.
89 TARGET_SSE_P: New Macro.
90 TARGET_SSE2_P: New Macro.
91 TARGET_SSE3_P: New Macro.
92 TARGET_SSSE3_P: New Macro.
93 TARGET_SSE4_1_P: New Macro.
94 TARGET_SSE4_2_P: New Macro.
95 TARGET_AVX_P: New Macro.
96 TARGET_AVX2_P: New Macro.
97 TARGET_AVX512F_P: New Macro.
98 TARGET_AVX512PF_P: New Macro.
99 TARGET_AVX512ER_P: New Macro.
100 TARGET_AVX512CD_P: New Macro.
101 TARGET_FMA_P: New Macro.
102 TARGET_SSE4A_P: New Macro.
103 TARGET_FMA4_P: New Macro.
104 TARGET_XOP_P: New Macro.
105 TARGET_LWP_P: New Macro.
106 TARGET_ABM_P: New Macro.
107 TARGET_BMI_P: New Macro.
108 TARGET_BMI2_P: New Macro.
109 TARGET_LZCNT_P: New Macro.
110 TARGET_TBM_P: New Macro.
111 TARGET_POPCNT_P: New Macro.
112 TARGET_SAHF_P: New Macro.
113 TARGET_MOVBE_P: New Macro.
114 TARGET_CRC32_P: New Macro.
115 TARGET_AES_P: New Macro.
116 TARGET_PCLMUL_P: New Macro.
117 TARGET_CMPXCHG16B_P: New Macro.
118 TARGET_FSGSBASE_P: New Macro.
119 TARGET_RDRND_P: New Macro.
120 TARGET_F16C_P: New Macro.
121 TARGET_RTM_P: New Macro.
122 TARGET_HLE_P: New Macro.
123 TARGET_RDSEED_P: New Macro.
124 TARGET_PRFCHW_P: New Macro.
125 TARGET_ADX_P: New Macro.
126 TARGET_FXSR_P: New Macro.
127 TARGET_XSAVE_P: New Macro.
128 TARGET_XSAVEOPT_P: New Macro.
129 TARGET_LP64_P: New Macro.
130 TARGET_X32_P: New Macro.
131 TARGET_FPMATH_DEFAULT_P: New Macro.
132 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
133 * config/i386/i386.c (ix86_option_override_internal): New parameters.
135 Change ix86_tune_string to access opts->x_ix86_tune_string.
136 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
137 Change ix86_arch_string to access opts->x_ix86_arch_string.
138 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
139 Change ix86_pmode to access opts->x_ix86_pmode.
140 Change ix86_abi to access opts->x_ix86_abi.
141 Change ix86_cmodel to access opts->x_ix86_cmodel.
142 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
143 Change ix86_isa_flags_explicit to access
144 opts->x_ix86_isa_flags_explicit.
145 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
146 Change ix86_regparm to access opts->x_ix86_regparm.
147 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
148 Change ix86_preferred_stack_boundary_arg to access
149 opts->x_ix86_preferred_stack_boundary_arg.
150 Change ix86_force_align_arg_pointer to access
151 opts->x_ix86_force_align_arg_pointer.
152 Change ix86_incoming_stack_boundar_arg to access
153 opts->x_ix86_incoming_stack_boundar_arg.
154 Change ix86_fpmath to access opts->x_ix86_fpmath.
155 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
156 Change ix86_recip_name to access opts->x_ix86_recip_name.
157 Change ix86_stack_protector_guard to access
158 opts->x_ix86_stack_protector_guard.
159 Change ix86_tune_memcpy_strategy to access
160 opts->x_ix86_tune_memcpy_strategy.
161 Change ix86_tune_memset_strategy to access
162 opts->x_ix86_tune_memset_strategy.
163 Change global_options to access opts.
164 Change global_options_set to access opts_set.
165 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
166 Change TARGET_MMX to TARGET_MMX_P (opts->...)
167 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...)
168 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...)
169 Change TARGET_SSE to TARGET_SSE_P (opts->...)
170 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
171 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...)
172 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...)
173 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...)
174 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...)
175 Change TARGET_AVX to TARGET_AVX_P (opts->...)
176 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...)
177 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...)
178 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...)
179 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...)
180 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...)
181 Change TARGET_FMA to TARGET_FMA_P (opts->...)
182 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...)
183 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...)
184 Change TARGET_XOP to TARGET_XOP_P (opts->...)
185 Change TARGET_LWP to TARGET_LWP_P (opts->...)
186 Change TARGET_ABM to TARGET_ABM_P (opts->...)
187 Change TARGET_BMI to TARGET_BMI_P (opts->...)
188 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...)
189 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...)
190 Change TARGET_TBM to TARGET_TBM_P (opts->...)
191 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...)
192 Change TARGET_SAHF to TARGET_SAHF_P (opts->...)
193 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...)
194 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...)
195 Change TARGET_AES to TARGET_AES_P (opts->...)
196 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...)
197 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...)
198 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...)
199 Change TARGET_RDRND to TARGET_RDRND_P (opts->...)
200 Change TARGET_F16C to TARGET_F16C_P (opts->...)
201 Change TARGET_RTM to TARGET_RTM_P (opts->...)
202 Change TARGET_HLE to TARGET_HLE_P (opts->...)
203 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...)
204 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...)
205 Change TARGET_ADX to TARGET_ADX_P (opts->...)
206 Change TARGET_FXSR to TARGET_FXSR_P (opts->...)
207 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...)
208 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...)
209 Change TARGET_LP64 to TARGET_LP64_P (opts->...)
210 Change TARGET_X32 to TARGET_X32_P (opts->...)
211 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...)
212 Change TARGET_FLOAT_RETURNS_IN_80387 to
213 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...)
214 (ix86_function_specific_save): New parameter. Use opts-> fields
215 to replace global fields.
216 (ix86_function_specific_restore): Ditto.
217 (ix86_valid_target_attribute_inner_p): New parameters.
219 Fix call to ix86_handle_option and set_option.
220 (ix86_valid_target_attribute_tree): New parameters.
221 Change global_options to access opts.
222 Change global_options_set to access opts_set.
223 Fix call to ix86_valid_target_attribute_inner_p.
224 Change ix86_tune_string to access opts->x_ix86_tune_string.
225 Change ix86_arch_string to access opts->x_ix86_arch_string.
226 Change ix86_fpmath to access opts->x_ix86_fpmath
227 Fix call to ix86_option_override_internal.
228 Fix call to ix86_add_new_builtins.
229 Fix calls to build_optimization_node and build_target_option_node.
230 (ix86_valid_target_attribute_p): Remove access to global_options.
231 Use new gcc_options structure func_options.
232 Fix call to ix86_valid_target_attribute_tree.
233 Fix call to build_optimization_node.
234 (get_builtin_code_for_version): Fix call to
235 ix86_valid_target_attribute_tree.
237 2013-10-15 David Malcolm <dmalcolm@redhat.com>
239 * Makefile.in (PICFLAG): New.
240 (enable_host_shared): New.
241 (INTERNAL_CFLAGS): Use PICFLAG.
242 (LIBIBERTY): Use pic build of libiberty.a if configured with
243 --enable-host-shared.
244 * configure.ac: Add --enable-host-shared, setting up new
246 * configure: Regenerate.
247 * doc/install.texi (--enable-shared): Add note contrasting it
249 (--enable-host-shared): New option.
251 2013-10-15 Richard Biener <rguenther@suse.de>
253 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
254 for built-in functions.
256 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
258 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
259 (optimize_range_tests_1): New function,
260 extracted from optimize_range_tests.
261 (optimize_range_tests_xor): Similarly.
262 (optimize_range_tests_diff): New function.
263 (optimize_range_tests): Use optimize_range_tests_1.
265 2013-10-15 Cong Hou <congh@google.com>
267 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
268 requirement of the reduction pattern so that one operand of the
269 reduction operation can come from outside of the loop.
271 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
273 * config/arm/neon-schedgen.ml: Remove.
274 * config/arm/cortex-a9-neon.md: Remove comment regarding
277 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
279 * config/arm/types: Remove old neon types.
281 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
283 * config/arm/cortex-a7.md
284 (cortex_a7_neon_type): New.
285 (cortex_a7_neon_mul): Update for new types.
286 (cortex_a7_neon_mla): Likewise.
287 (cortex_a7_neon): Likewise.
289 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
291 * config/arm/cortex-a15-neon.md
292 (cortex_a15_neon_type): New,
294 (cortex_a15_neon_int_1): Remove.
295 (cortex_a15_neon_int_2): Likewise.
296 (cortex_a15_neon_int_3): Likewise.
297 (cortex_a15_neon_int_4): Likewise.
298 (cortex_a15_neon_int_5): Likewise.
299 (cortex_a15_neon_vqneg_vqabs): Likewise.
300 (cortex_a15_neon_vmov): Likewise.
301 (cortex_a15_neon_vaba): Likewise.
302 (cortex_a15_neon_vaba_qqq): Likewise.
303 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
304 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
305 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
307 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
308 (cortex_a15_neon_mla_qqq_8_16): Likewise.
309 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
310 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
311 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
312 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
313 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
314 (cortex_a15_neon_shift_1): Likewise.
315 (cortex_a15_neon_shift_2): Likewise.
316 (cortex_a15_neon_shift_3): Likewise.
317 (cortex_a15_neon_vshl_ddd): Likewise.
318 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
319 (cortex_a15_neon_vsra_vrsra): Likewise.
320 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
321 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
322 (cortex_a15_neon_bp_3cycle): Likewise.
323 (cortex_a15_neon_ldm_2): Likewise.
324 (cortex_a15_neon_stm_2): Likewise.
325 (cortex_a15_neon_mcr): Likewise.
326 (cortex_a15_neon_mrc): Likewise.
327 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
328 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
329 (cortex_a15_neon_fp_vmul_ddd): Likewise.
330 (cortex_a15_neon_fp_vmul_qqd): Likewise.
331 (cortex_a15_neon_fp_vmla_ddd): Likewise.
332 (cortex_a15_neon_fp_vmla_qqq): Likewise.
333 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
334 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
335 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
336 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
337 (cortex_a15_neon_bp_simple): Likewise.
338 (cortex_a15_neon_bp_2cycle): Likewise.
339 (cortex_a15_neon_bp_3cycle): Likewise.
340 (cortex_a15_neon_vld1_1_2_regs): Likewise.
341 (cortex_a15_neon_vld1_3_4_regs): Likewise.
342 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
343 (cortex_a15_neon_vld2_4_regs): Likewise.
344 (cortex_a15_neon_vld3_vld4): Likewise.
345 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
346 (cortex_a15_neon_vst1_3_4_regs): Likewise.
347 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
348 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
349 (cortex_a15_neon_vst3_vst4): Rename to...
350 (cortex_a15_neon_vst4): This, update for new attributes.
351 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
352 (cortex_a15_neon_vld3_vld4_lane): Likewise.
353 (cortex_a15_neon_vst1_vst2_lane): Likewise.
354 (cortex_a15_neon_vst3_vst4_lane): Likewise.
355 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
356 (cortex_a15_neon_ldm_2): Likewise.
357 (cortex_a15_neon_stm_2): Likewise.
358 (cortex_a15_neon_mcr): Likewise.
359 (cortex_a15_neon_mcr_2_mcrr): Likewise.
360 (cortex_a15_neon_mrc): Likewise.
361 (cortex_a15_neon_mrrc): Likewise.
363 (cortex_a15_neon_abd): New.
364 (cortex_a15_neon_abd_q): Likewise.
365 (cortex_a15_neon_aba): Likewise.
366 (cortex_a15_neon_aba_q): Likewise.
367 (cortex_a15_neon_acc): Likewise.
368 (cortex_a15_neon_acc_q): Likewise.
369 (cortex_a15_neon_arith_basic): Likewise.
370 (cortex_a15_neon_arith_complex): Likewise.
371 (cortex_a15_neon_multiply): Likewise.
372 (cortex_a15_neon_multiply_q): Likewise.
373 (cortex_a15_neon_mla): Likewise.
374 (cortex_a15_neon_mla_q): Likewise.
375 (cortex_a15_neon_sat_mla_long): Likewise.
376 (cortex_a15_neon_shift_acc): Likewise.
377 (cortex_a15_neon_shift_imm_basic): Likewise.
378 (cortex_a15_neon_shift_imm_complex): Likewise.
379 (cortex_a15_neon_shift_reg_basic): Likewise.
380 (cortex_a15_neon_shift_reg_basic_q): Likewise.
381 (cortex_a15_neon_shift_reg_complex): Likewise.
382 (cortex_a15_neon_shift_reg_complex_q): Likewise.
383 (cortex_a15_neon_fp_negabs): Likewise
384 (cortex_a15_neon_fp_arith): Likewise
385 (cortex_a15_neon_fp_arith_q): Likewise
386 (cortex_a15_neon_fp_cvt_int): Likewise
387 (cortex_a15_neon_fp_cvt_int_q): Likewise
388 (cortex_a15_neon_fp_cvt_16): Likewise
389 (cortex_a15_neon_fp_mul): Likewise
390 (cortex_a15_neon_fp_mul_q): Likewise
391 (cortex_a15_neon_fp_mla): Likewise
392 (cortex_a15_neon_fp_mla_q): Likewise
393 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
394 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
395 (cortex_a15_neon_bitops): Likewise.
396 (cortex_a15_neon_bitops_q): Likewise.
397 (cortex_a15_neon_from_gp): Likewise.
398 (cortex_a15_neon_from_gp_q): Likewise.
399 (cortex_a15_neon_tbl3_tbl4): Likewise.
400 (cortex_a15_neon_zip_q): Likewise.
401 (cortex_a15_neon_to_gp): Likewise.
402 (cortex_a15_neon_load_a): Likewise.
403 (cortex_a15_neon_load_b): Likewise.
404 (cortex_a15_neon_load_c): Likewise.
405 (cortex_a15_neon_load_d): Likewise.
406 (cortex_a15_neon_load_e): Likewise.
407 (cortex_a15_neon_load_f): Likewise.
408 (cortex_a15_neon_store_a): Likewise.
409 (cortex_a15_neon_store_b): Likewise.
410 (cortex_a15_neon_store_c): Likewise.
411 (cortex_a15_neon_store_d): Likewise.
412 (cortex_a15_neon_store_e): Likewise.
413 (cortex_a15_neon_store_f): Likewise.
414 (cortex_a15_neon_store_g): Likewise.
415 (cortex_a15_neon_store_h): Likewise.
416 (cortex_a15_vfp_to_from_gp): Likewise.
418 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
420 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
422 (cortex_a9_neon_vshl_ddd): Remove.
423 (cortex_a9_neon_vst3_vst4): Likewise.
424 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
426 (cortex_a9_neon_bit_ops_q): New.
428 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
429 (cortex_a9_neon_int_2): Likewise.
430 (cortex_a9_neon_int_3): Likewise.
431 (cortex_a9_neon_int_4): Likewise.
432 (cortex_a9_neon_int_5): Likewise.
433 (cortex_a9_neon_vqneg_vqabs): Likewise.
434 (cortex_a9_neon_vmov): Likewise.
435 (cortex_a9_neon_vaba): Likewise.
436 (cortex_a9_neon_vaba_qqq): Likewise.
437 (cortex_a9_neon_shift_1): Likewise.
438 (cortex_a9_neon_shift_2): Likewise.
439 (cortex_a9_neon_shift_3): Likewise.
440 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
441 (cortex_a9_neon_vsra_vrsra): Likewise.
442 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
443 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
444 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
446 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
447 (cortex_a9_neon_mla_qqq_8_16): Likewise.
448 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
450 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
451 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
452 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
453 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
454 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
455 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
456 (cortex_a9_neon_fp_vsum): Likewise.
457 (cortex_a9_neon_fp_vmul_ddd): Likewise.
458 (cortex_a9_neon_fp_vmul_qqd): Likewise.
459 (cortex_a9_neon_fp_vmla_ddd): Likewise.
460 (cortex_a9_neon_fp_vmla_qqq): Likewise.
461 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
462 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
463 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
464 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
465 (cortex_a9_neon_bp_simple): Likewise.
466 (cortex_a9_neon_bp_2cycle): Likewise.
467 (cortex_a9_neon_bp_3cycle): Likewise.
468 (cortex_a9_neon_ldr): Likewise.
469 (cortex_a9_neon_str): Likewise.
470 (cortex_a9_neon_vld1_1_2_regs): Likewise.
471 (cortex_a9_neon_vld1_3_4_regs): Likewise.
472 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
473 (cortex_a9_neon_vld2_4_regs): Likewise.
474 (cortex_a9_neon_vld3_vld4): Likewise.
475 (cortex_a9_neon_vld1_vld2_lane): Likewise.
476 (cortex_a9_neon_vld3_vld4_lane): Likewise.
477 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
478 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
479 (cortex_a9_neon_vst1_3_4_regs): Likewise.
480 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
481 (cortex_a9_neon_vst1_vst2_lane): Likewise.
482 (cortex_a9_neon_vst3_vst4_lane): Likewise.
483 (cortex_a9_neon_mcr): Likewise.
484 (cortex_a9_neon_mcr_2_mcrr): Likewise.
485 (cortex_a9_neon_mrc): Likewise.
486 (cortex_a9_neon_mrrc): Likewise.
488 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
490 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
492 (cortex_a8_neon_vshl_ddd): Remove.
493 (cortex_a8_neon_vst3_vst4): Likewise.
494 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
496 (cortex_a8_neon_bit_ops_q): New.
498 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
499 (cortex_a8_neon_int_2): Likewise..
500 (cortex_a8_neon_int_3): Likewise.
501 (cortex_a8_neon_int_5): Likewise.
502 (cortex_a8_neon_vqneg_vqabs): Likewise.
503 (cortex_a8_neon_int_4): Likewise.
504 (cortex_a8_neon_vaba): Likewise.
505 (cortex_a8_neon_vaba_qqq): Likewise.
506 (cortex_a8_neon_shift_1): Likewise.
507 (cortex_a8_neon_shift_2): Likewise.
508 (cortex_a8_neon_shift_3): Likewise.
509 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
510 (cortex_a8_neon_vsra_vrsra): Likewise.
511 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
512 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
513 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
515 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
516 (cortex_a8_neon_mla_qqq_8_16): Likewise.
517 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
519 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
520 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
521 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
522 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
523 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
524 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
525 (cortex_a8_neon_fp_vsum): Likewise.
526 (cortex_a8_neon_fp_vmul_ddd): Likewise.
527 (cortex_a8_neon_fp_vmul_qqd): Likewise.
528 (cortex_a8_neon_fp_vmla_ddd): Likewise.
529 (cortex_a8_neon_fp_vmla_qqq): Likewise.
530 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
531 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
532 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
533 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
534 (cortex_a8_neon_bp_simple): Likewise.
535 (cortex_a8_neon_bp_2cycle): Likewise.
536 (cortex_a8_neon_bp_3cycle): Likewise.
537 (cortex_a8_neon_ldr): Likewise.
538 (cortex_a8_neon_str): Likewise.
539 (cortex_a8_neon_vld1_1_2_regs): Likewise.
540 (cortex_a8_neon_vld1_3_4_regs): Likewise.
541 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
542 (cortex_a8_neon_vld2_4_regs): Likewise.
543 (cortex_a8_neon_vld3_vld4): Likewise.
544 (cortex_a8_neon_vld1_vld2_lane): Likewise.
545 (cortex_a8_neon_vld3_vld4_lane): Likewise.
546 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
547 (cortex_a8_neon_vst1_3_4_regs): Likewise.
548 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
549 (cortex_a8_neon_vst1_vst2_lane): Likewise.
550 (cortex_a8_neon_vst3_vst4_lane): Likewise.
551 (cortex_a8_neon_mcr): Likewise.
552 (cortex_a8_neon_mcr_2_mcrr): Likewise.
553 (cortex_a8_neon_mrc): Likewise.
554 (cortex_a8_neon_mrrc): Likewise.
556 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
558 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
560 * config/aarch64/aarch64-simd.md (neon_type): Remove.
561 (aarch64_simd_dup<mode>): Add "type" attribute.
562 (aarch64_dup_lane<mode>): Likewise.
563 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
564 (*aarch64_simd_mov<mode>): Likewise.
565 (aarch64_simd_mov_from_<mode>low): Likewise.
566 (aarch64_simd_mov_from_<mode>high): Likewise.
567 (orn<mode>3): Likewise.
568 (bic<mode>3): Likewise.
569 (add<mode>3): Likewise.
570 (sub<mode>3): Likewise.
571 (mul<mode>3): Likewise.
572 (*aarch64_mul3_elt<mode>): Likewise.
573 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
574 (*aarch64_mul3_elt_to_128df): Likewise.
575 (*aarch64_mul3_elt_to_64v2df): Likewise.
576 (neg<mode>2): Likewise.
577 (abs<mode>2): Likewise.
578 (abd<mode>_3): Likewise.
579 (aba<mode>_3): Likewise.
580 (fabd<mode>_3): Likewise.
581 (*fabd_scalar<mode>3): Likewise.
582 (and<mode>3): Likewise.
583 (ior<mode>3): Likewise.
584 (xor<mode>3): Likewise.
585 (one_cmpl<mode>2): Likewise.
586 (aarch64_simd_vec_set<mode>): Likewise.
587 (aarch64_simd_lshr<mode>): Likewise.
588 (aarch64_simd_ashr<mode>): Likewise.
589 (aarch64_simd_imm_shl<mode>): Likewise.
590 (aarch64_simd_reg_sshl<mode): Likewise.
591 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
592 (aarch64_simd_reg_shl<mode>_signed): Likewise.
593 (aarch64_simd_vec_setv2di): Likewise.
594 (aarch64_simd_vec_set<mode>): Likewise.
595 (aarch64_mla<mode>): Likewise.
596 (*aarch64_mla_elt<mode>): Likewise.
597 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
598 (aarch64_mls<mode>): Likewise.
599 (*aarch64_mls_elt<mode>): Likewise.
600 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
601 (<su><maxmin><mode>3): Likewise.
602 (move_lo_quad_<mode>): Likewise.
603 (aarch64_simd_move_hi_quad_<mode>): Likewise.
604 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
605 (vec_pack_trunc_<mode>): Likewise.
606 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
607 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
608 (*aarch64_<su>mlal_lo<mode>): Likewise.
609 (*aarch64_<su>mlal_hi<mode>): Likewise.
610 (*aarch64_<su>mlsl_lo<mode>): Likewise.
611 (*aarch64_<su>mlsl_hi<mode>): Likewise.
612 (*aarch64_<su>mlal<mode>): Likewise.
613 (*aarch64_<su>mlsl<mode>): Likewise.
614 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
615 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
616 (add<mode>3): Likewise.
617 (sub<mode>3): Likewise.
618 (mul<mode>3): Likewise.
619 (div<mode>3): Likewise.
620 (neg<mode>2): Likewise.
621 (abs<mode>2): Likewise.
622 (fma<mode>4): Likewise.
623 (*aarch64_fma4_elt<mode>): Likewise.
624 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
625 (*aarch64_fma4_elt_to_128df): Likewise.
626 (*aarch64_fma4_elt_to_64v2df): Likewise.
627 (fnma<mode>4): Likewise.
628 (*aarch64_fnma4_elt<mode>): Likewise.
629 (*aarch64_fnma4_elt_<vswap_width_name><mode>
630 (*aarch64_fnma4_elt_to_128df): Likewise.
631 (*aarch64_fnma4_elt_to_64v2df): Likewise.
632 (<frint_pattern><mode>2): Likewise.
633 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
634 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
635 (vec_unpacks_lo_v4sf): Likewise.
636 (aarch64_float_extend_lo_v2df): Likewise.
637 (vec_unpacks_hi_v4sf): Likewise.
638 (aarch64_float_truncate_lo_v2sf): Likewise.
639 (aarch64_float_truncate_hi_v4sf): Likewise.
640 (aarch64_vmls<mode>): Likewise.
641 (<su><maxmin><mode>3): Likewise.
642 (<maxmin_uns><mode>3): Likewise.
643 (reduc_<sur>plus_<mode>): Likewise.
644 (reduc_<sur>plus_v2di): Likewise.
645 (reduc_<sur>plus_v2si): Likewise.
646 (reduc_<sur>plus_<mode>): Likewise.
647 (aarch64_addpv4sf): Likewise.
648 (clz<mode>2): Likewise.
649 (reduc_<maxmin_uns>_<mode>): Likewise.
650 (reduc_<maxmin_uns>_v2di): Likewise.
651 (reduc_<maxmin_uns>_v2si): Likewise.
652 (reduc_<maxmin_uns>_<mode>): Likewise.
653 (reduc_<maxmin_uns>_v4sf): Likewise.
654 (aarch64_simd_bsl<mode>_internal): Likewise.
655 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
656 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
657 (aarch64_get_lane<mode>): Likewise.
658 (*aarch64_combinez<mode>): Likewise.
659 (aarch64_combine<mode>): Likewise.
660 (aarch64_simd_combine<mode>): Likewise.
661 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
662 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
663 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
664 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
665 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
666 (aarch64_<sur>h<addsub><mode>): Likewise.
667 (aarch64_<sur><addsub>hn<mode>): Likewise.
668 (aarch64_<sur><addsub>hn2<mode>): Likewise.
669 (aarch64_pmul<mode>): Likewise.
670 (aarch64_<su_optab><optab><mode>): Likewise.
671 (aarch64_<sur>qadd<mode>): Likewise.
672 (aarch64_sqmovun<mode>): Likewise.
673 (aarch64_<sur>qmovn<mode>): Likewise.
674 (aarch64_s<optab><mode>): Likewise.
675 (aarch64_sq<r>dmulh<mode>): Likewise.
676 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
677 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
678 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
679 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
680 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
681 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
682 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
683 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
684 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
685 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
686 (aarch64_sqdmull<mode>): Likewise.
687 (aarch64_sqdmull_lane<mode>_internal): Likewise.
688 (aarch64_sqdmull_n<mode>): Likewise.
689 (aarch64_sqdmull2<mode>_internal): Likewise.
690 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
691 (aarch64_sqdmull2_n<mode>_internal): Likewise.
692 (aarch64_<sur>shl<mode>): Likewise.
693 (aarch64_<sur>q<r>shl<mode>
694 (aarch64_<sur>shll_n<mode>): Likewise.
695 (aarch64_<sur>shll2_n<mode>): Likewise.
696 (aarch64_<sur>shr_n<mode>): Likewise.
697 (aarch64_<sur>sra_n<mode>): Likewise.
698 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
699 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
700 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
701 (aarch64_cm<optab><mode>): Likewise.
702 (aarch64_cm<optab>di): Likewise.
703 (aarch64_cm<optab><mode>): Likewise.
704 (aarch64_cm<optab>di): Likewise.
705 (aarch64_cmtst<mode>): Likewise.
706 (aarch64_cmtstdi): Likewise.
707 (aarch64_cm<optab><mode>): Likewise.
708 (*aarch64_fac<optab><mode>): Likewise.
709 (aarch64_addp<mode>): Likewise.
710 (aarch64_addpdi): Likewise.
711 (sqrt<mode>2): Likewise.
712 (vec_load_lanesoi<mode>): Likewise.
713 (vec_store_lanesoi<mode>): Likewise.
714 (vec_load_lanesci<mode>): Likewise.
715 (vec_store_lanesci<mode>): Likewise.
716 (vec_load_lanesxi<mode>): Likewise.
717 (vec_store_lanesxi<mode>): Likewise.
718 (*aarch64_mov<mode>): Likewise.
719 (aarch64_ld2<mode>_dreg): Likewise.
720 (aarch64_ld2<mode>_dreg): Likewise.
721 (aarch64_ld3<mode>_dreg): Likewise.
722 (aarch64_ld3<mode>_dreg): Likewise.
723 (aarch64_ld4<mode>_dreg): Likewise.
724 (aarch64_ld4<mode>_dreg): Likewise.
725 (aarch64_tbl1<mode>): Likewise.
726 (aarch64_tbl2v16qi): Likewise.
727 (aarch64_combinev16qi): Likewise.
728 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
729 (aarch64_st2<mode>_dreg): Likewise.
730 (aarch64_st2<mode>_dreg): Likewise.
731 (aarch64_st3<mode>_dreg): Likewise.
732 (aarch64_st3<mode>_dreg): Likewise.
733 (aarch64_st4<mode>_dreg): Likewise.
734 (aarch64_st4<mode>_dreg): Likewise.
735 (*aarch64_simd_ld1r<mode>): Likewise.
736 (aarch64_frecpe<mode>): Likewise.
737 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
738 (aarch64_frecps<mode>): Likewise.
740 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
742 * config/arm/iterators.md (V_elem_ch): New.
744 (VQH_type): Likewise.
745 * config/arm/arm.md (is_neon_type): New.
746 (conds): Use is_neon_type.
747 (anddi3_insn): Update type attribute.
748 (xordi3_insn): Likewise.
749 (one_cmpldi2): Likewise.
750 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
751 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
752 (*movmisalign<mode>_neon_store): Likewise.
753 (*movmisalign<mode>_neon_load): Likewise.
754 (vec_set<mode>_internal): Likewise.
755 (vec_set<mode>_internal): Likewise.
756 (vec_setv2di_internal): Likewise.
757 (vec_extract<mode>): Likewise.
758 (vec_extract<mode>): Likewise.
759 (vec_extractv2di): Likewise.
760 (*add<mode>3_neon): Likewise.
761 (adddi3_neon): Likewise.
762 (*sub<mode>3_neon): Likewise.
763 (subdi3_neon): Likewise.
764 (fma<VCVTF:mode>4): Likewise.
765 (fma<VCVTF:mode>4_intrinsic): Likewise.
766 (*fmsub<VCVTF:mode>4): Likewise.
767 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
768 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
769 (ior<mode>3): Likewise.
770 (and<mode>3): Likewise.
771 (orn<mode>3_neon): Likewise.
772 (orndi3_neon): Likewise.
773 (bic<mode>3_neon): Likewise.
774 (bicdi3_neon): Likewise.
775 (xor<mode>3): Likewise.
776 (one_cmpl<mode>2): Likewise.
777 (abs<mode>2): Likewise.
778 (neg<mode>2): Likewise.
779 (negdi2_neon): Likewise.
780 (*umin<mode>3_neon): Likewise.
781 (*umax<mode>3_neon): Likewise.
782 (*smin<mode>3_neon): Likewise.
783 (*smax<mode>3_neon): Likewise.
784 (vashl<mode>3): Likewise.
785 (vashr<mode>3_imm): Likewise.
786 (vlshr<mode>3_imm): Likewise.
787 (ashl<mode>3_signed): Likewise.
788 (ashl<mode>3_unsigned): Likewise.
789 (neon_load_count): Likewise.
790 (ashldi3_neon_noclobber): Likewise.
791 (ashldi3_neon): Likewise.
792 (signed_shift_di3_neon): Likewise.
793 (unsigned_shift_di3_neon): Likewise.
794 (ashrdi3_neon_imm_noclobber): Likewise.
795 (lshrdi3_neon_imm_noclobber): Likewise.
796 (<shift>di3_neon): Likewise.
797 (widen_ssum<mode>3): Likewise.
798 (widen_usum<mode>3): Likewise.
799 (quad_halves_<code>v4si): Likewise.
800 (quad_halves_<code>v4sf): Likewise.
801 (quad_halves_<code>v8hi): Likewise.
802 (quad_halves_<code>v16qi): Likewise.
803 (reduc_splus_v2di): Likewise.
804 (neon_vpadd_internal<mode>): Likewise.
805 (neon_vpsmin<mode>): Likewise.
806 (neon_vpsmax<mode>): Likewise.
807 (neon_vpumin<mode>): Likewise.
808 (neon_vpumax<mode>): Likewise.
809 (*ss_add<mode>_neon): Likewise.
810 (*us_add<mode>_neon): Likewise.
811 (*ss_sub<mode>_neon): Likewise.
812 (*us_sub<mode>_neon): Likewise.
813 (neon_vadd<mode>_unspec): Likewise.
814 (neon_vaddl<mode>): Likewise.
815 (neon_vaddw<mode>): Likewise.
816 (neon_vhadd<mode>): Likewise.
817 (neon_vqadd<mode>): Likewise.
818 (neon_vaddhn<mode>): Likewise.
819 (neon_vmul<mode>): Likewise.
820 (neon_vfms<VCVTF:mode>): Likewise.
821 (neon_vmlal<mode>): Likewise.
822 (neon_vmls<mode>): Likewise.
823 (neon_vmlsl<mode>): Likewise.
824 (neon_vqdmulh<mode>): Likewise.
825 (neon_vqdmlal<mode>): Likewise.
826 (neon_vqdmlsl<mode>): Likewise.
827 (neon_vmull<mode>): Likewise.
828 (neon_vqdmull<mode>): Likewise.
829 (neon_vsub<mode>_unspec): Likewise.
830 (neon_vsubl<mode>): Likewise.
831 (neon_vsubw<mode>): Likewise.
832 (neon_vqsub<mode>): Likewise.
833 (neon_vhsub<mode>): Likewise.
834 (neon_vsubhn<mode>): Likewise.
835 (neon_vceq<mode>): Likewise.
836 (neon_vcge<mode>): Likewise.
837 (neon_vcgeu<mode>): Likewise.
838 (neon_vcgt<mode>): Likewise.
839 (neon_vcgtu<mode>): Likewise.
840 (neon_vcle<mode>): Likewise.
841 (neon_vclt<mode>): Likewise.
842 (neon_vcage<mode>): Likewise.
843 (neon_vcagt<mode>): Likewise.
844 (neon_vtst<mode>): Likewise.
845 (neon_vabd<mode>): Likewise.
846 (neon_vabdl<mode>): Likewise.
847 (neon_vaba<mode>): Likewise.
848 (neon_vabal<mode>): Likewise.
849 (neon_vmax<mode>): Likewise.
850 (neon_vmin<mode>): Likewise.
851 (neon_vpaddl<mode>): Likewise.
852 (neon_vpadal<mode>): Likewise.
853 (neon_vpmax<mode>): Likewise.
854 (neon_vpmin<mode>): Likewise.
855 (neon_vrecps<mode>): Likewise.
856 (neon_vrsqrts<mode>): Likewise.
857 (neon_vqabs<mode>): Likewise.
858 (neon_vqneg<mode>): Likewise.
859 (neon_vcls<mode>): Likewise.
860 (clz<mode>2): Likewise.
861 (popcount<mode>2): Likewise.
862 (neon_vrecpe<mode>): Likewise.
863 (neon_vrsqrte<mode>): Likewise.
864 (neon_vget_lane<mode>_sext_internal): Likewise.
865 (neon_vget_lane<mode>_zext_internal): Likewise.
866 (neon_vdup_n<mode>): Likewise.
867 (neon_vdup_n<mode>): Likewise.
868 (neon_vdup_nv2di): Likewise.
869 (neon_vdup_lane<mode>_interal): Likewise.
870 (*neon_vswp<mode>): Likewise.
871 (neon_vcombine<mode>): Likewise.
872 (float<mode><V_cvtto>2): Likewise.
873 (floatuns<mode><V_cvtto>2): Likewise.
874 (fix_trunc<mode><V_cvtto>2): Likewise.
875 (fixuns_trunc<mode><V_cvtto>2
876 (neon_vcvt<mode>): Likewise.
877 (neon_vcvt<mode>): Likewise.
878 (neon_vcvtv4sfv4hf): Likewise.
879 (neon_vcvtv4hfv4sf): Likewise.
880 (neon_vcvt_n<mode>): Likewise.
881 (neon_vcvt_n<mode>): Likewise.
882 (neon_vmovn<mode>): Likewise.
883 (neon_vqmovn<mode>): Likewise.
884 (neon_vqmovun<mode>): Likewise.
885 (neon_vmovl<mode>): Likewise.
886 (neon_vmul_lane<mode>): Likewise.
887 (neon_vmul_lane<mode>): Likewise.
888 (neon_vmull_lane<mode>): Likewise.
889 (neon_vqdmull_lane<mode>): Likewise.
890 (neon_vqdmulh_lane<mode>): Likewise.
891 (neon_vqdmulh_lane<mode>): Likewise.
892 (neon_vmla_lane<mode>): Likewise.
893 (neon_vmla_lane<mode>): Likewise.
894 (neon_vmlal_lane<mode>): Likewise.
895 (neon_vqdmlal_lane<mode>): Likewise.
896 (neon_vmls_lane<mode>): Likewise.
897 (neon_vmls_lane<mode>): Likewise.
898 (neon_vmlsl_lane<mode>): Likewise.
899 (neon_vqdmlsl_lane<mode>): Likewise.
900 (neon_vext<mode>): Likewise.
901 (neon_vrev64<mode>): Likewise.
902 (neon_vrev32<mode>): Likewise.
903 (neon_vrev16<mode>): Likewise.
904 (neon_vbsl<mode>_internal): Likewise.
905 (neon_vshl<mode>): Likewise.
906 (neon_vqshl<mode>): Likewise.
907 (neon_vshr_n<mode>): Likewise.
908 (neon_vshrn_n<mode>): Likewise.
909 (neon_vqshrn_n<mode>): Likewise.
910 (neon_vqshrun_n<mode>): Likewise.
911 (neon_vshl_n<mode>): Likewise.
912 (neon_vqshl_n<mode>): Likewise.
913 (neon_vqshlu_n<mode>): Likewise.
914 (neon_vshll_n<mode>): Likewise.
915 (neon_vsra_n<mode>): Likewise.
916 (neon_vsri_n<mode>): Likewise.
917 (neon_vsli_n<mode>): Likewise.
918 (neon_vtbl1v8qi): Likewise.
919 (neon_vtbl2v8qi): Likewise.
920 (neon_vtbl3v8qi): Likewise.
921 (neon_vtbl4v8qi): Likewise.
922 (neon_vtbl1v16qi): Likewise.
923 (neon_vtbl2v16qi): Likewise.
924 (neon_vcombinev16qi): Likewise.
925 (neon_vtbx1v8qi): Likewise.
926 (neon_vtbx2v8qi): Likewise.
927 (neon_vtbx3v8qi): Likewise.
928 (neon_vtbx4v8qi): Likewise.
929 (*neon_vtrn<mode>_insn): Likewise.
930 (*neon_vzip<mode>_insn): Likewise.
931 (*neon_vuzp<mode>_insn): Likewise.
932 (neon_vld1<mode>): Likewise.
933 (neon_vld1_lane<mode>): Likewise.
934 (neon_vld1_lane<mode>): Likewise.
935 (neon_vld1_dup<mode>): Likewise.
936 (neon_vld1_dup<mode>): Likewise.
937 (neon_vld1_dupv2di): Likewise.
938 (neon_vst1<mode>): Likewise.
939 (neon_vst1_lane<mode>): Likewise.
940 (neon_vst1_lane<mode>): Likewise.
941 (neon_vld2<mode>): Likewise.
942 (neon_vld2<mode>): Likewise.
943 (neon_vld2_lane<mode>): Likewise.
944 (neon_vld2_lane<mode>): Likewise.
945 (neon_vld2_dup<mode>): Likewise.
946 (neon_vst2<mode>): Likewise.
947 (neon_vst2<mode>): Likewise.
948 (neon_vst2_lane<mode>): Likewise.
949 (neon_vst2_lane<mode>): Likewise.
950 (neon_vld3<mode>): Likewise.
951 (neon_vld3qa<mode>): Likewise.
952 (neon_vld3qb<mode>): Likewise.
953 (neon_vld3_lane<mode>): Likewise.
954 (neon_vld3_lane<mode>): Likewise.
955 (neon_vld3_dup<mode>): Likewise.
956 (neon_vst3<mode>): Likewise.
957 (neon_vst3qa<mode>): Likewise.
958 (neon_vst3qb<mode>): Likewise.
959 (neon_vst3_lane<mode>): Likewise.
960 (neon_vst3_lane<mode>): Likewise.
961 (neon_vld4<mode>): Likewise.
962 (neon_vld4qa<mode>): Likewise.
963 (neon_vld4qb<mode>): Likewise.
964 (neon_vld4_lane<mode>): Likewise.
965 (neon_vld4_lane<mode>): Likewise.
966 (neon_vld4_dup<mode>): Likewise.
967 (neon_vst4<mode>): Likewise.
968 (neon_vst4qa<mode>): Likewise.
969 (neon_vst4qb<mode>): Likewise.
970 (neon_vst4_lane<mode>): Likewise.
971 (neon_vst4_lane<mode>): Likewise.
972 (neon_vec_unpack<US>_lo_<mode>): Likewise.
973 (neon_vec_unpack<US>_hi_<mode>): Likewise.
974 (neon_vec_<US>mult_lo_<mode>): Likewise.
975 (neon_vec_<US>mult_hi_<mode>): Likewise.
976 (neon_vec_<US>shiftl_<mode>): Likewise.
977 (neon_unpack<US>_<mode>): Likewise.
978 (neon_vec_<US>mult_<mode>): Likewise.
979 (vec_pack_trunc_<mode>): Likewise.
980 (neon_vec_pack_trunc_<mode>): Likewise.
981 (neon_vabd<mode>_2): Likewise.
982 (neon_vabd<mode>_3): Likewise.
984 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
986 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
987 (load_pair): Update type attribute.
988 (store_pair): Update type attribute.
989 * config/aarch64/iterators.md (q): New.
991 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
993 * config/arm/types.md: Add new types for Neon insns.
995 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
996 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
997 Sergey Lega <sergey.s.lega@intel.com>
998 Anna Tikhonova <anna.tikhonova@intel.com>
999 Ilya Tocar <ilya.tocar@intel.com>
1000 Andrey Turetskiy <andrey.turetskiy@intel.com>
1001 Ilya Verbin <ilya.verbin@intel.com>
1002 Kirill Yukhin <kirill.yukhin@intel.com>
1003 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1005 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1006 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1007 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1009 (srcp14<mode>): Ditto.
1010 (rsqrt14<mode>): Ditto.
1011 (rsqrt14<mode>): Ditto.
1012 (avx512f_vmscalef<mode>): Ditto.
1013 (avx512f_scalef<mode>): Ditto.
1014 (avx512f_getexp<mode>): Ditto.
1015 (avx512f_sgetexp<mode>): Ditto.
1016 (avx512f_fixupimm<mode>): Ditto.
1017 (avx512f_sfixupimm<mode>): Ditto.
1018 (avx512f_rndscale<mode>): Ditto.
1019 (*avx512er_exp2<mode>): Ditto.
1020 (*avx512er_rcp28<mode>): Ditto.
1021 (avx512er_rsqrt28<mode>): Ditto.
1022 (avx512f_getmant<mode>): Ditto.
1023 (avx512f_getmant<mode>): Ditto.
1024 (avx512f_rndscale<mode>): Fix formatting.
1026 2013-10-15 Martin Jambor <mjambor@suse.cz>
1028 * ipa-utils.h (ipa_edge_within_scc): Declare.
1029 * ipa-cp.c (edge_within_scc): Moved...
1030 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1032 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1033 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1034 Sergey Lega <sergey.s.lega@intel.com>
1035 Anna Tikhonova <anna.tikhonova@intel.com>
1036 Ilya Tocar <ilya.tocar@intel.com>
1037 Andrey Turetskiy <andrey.turetskiy@intel.com>
1038 Ilya Verbin <ilya.verbin@intel.com>
1039 Kirill Yukhin <kirill.yukhin@intel.com>
1040 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1042 * config/i386/predicates.md (const_8_to_15_operand): New.
1043 (const_16_to_31_operand): Ditto.
1044 * config/i386/sse.md (V8FI): New.
1046 (reduc_splus_v8df): Ditto.
1047 (reduc_splus_v16sf): Ditto.
1048 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1049 (vec_extract_hi_<mode>): Ditto.
1050 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1051 (vec_set_lo_<mode>): Ditto.
1052 (vec_set_hi_<mode>): Ditto.
1053 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1054 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1055 (avx512f_pshufd_1): Ditto.
1056 (avx512f_broadcast<mode>): Ditto.
1057 (avx512f_broadcast<mode>): Ditto.
1058 (define_split): Split vec_extract_lo into move.
1059 (ssequartermode): Ditto.
1060 (ssedoublemode): Extened with wider modes.
1061 (vec_extract_lo_<mode>): Ditto.
1063 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1064 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1065 Sergey Lega <sergey.s.lega@intel.com>
1066 Anna Tikhonova <anna.tikhonova@intel.com>
1067 Ilya Tocar <ilya.tocar@intel.com>
1068 Andrey Turetskiy <andrey.turetskiy@intel.com>
1069 Ilya Verbin <ilya.verbin@intel.com>
1070 Kirill Yukhin <kirill.yukhin@intel.com>
1071 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1073 * config/i386/predicates.md (register_or_constm1_operand): New.
1074 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1075 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1076 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1077 UNSPEC_SCATTER_PREFETCH
1079 (avx512f_ucmp<mode>3): Ditto.
1080 (avx512f_vternlog<mode>): Ditto.
1081 (avx512f_align<mode>): Ditto.
1082 (<shift_insn><mode>3): Ditto.
1083 (avx512f_<rotate>v<mode>): Ditto.
1084 (avx512f_<rotate><mode>): Ditto.
1085 (avx512f_eq<mode>3): Ditto.
1086 (avx512f_eq<mode>3_1): Ditto.
1087 (avx512f_gt<mode>3): Ditto.
1088 (avx512f_testm<mode>3): Ditto.
1089 (avx512f_testnm<mode>3): Ditto.
1090 (avx512pf_gatherpf<mode>): Ditto.
1091 (*avx512pf_gatherpf<mode>_mask): Ditto.
1092 (*avx512pf_gatherpf<mode>): Ditto.
1093 (avx512pf_scatterpf<mode>): Ditto.
1094 (*avx512pf_scatterpf<mode>_mask): Ditto.
1095 (*avx512pf_scatterpf<mode>): Ditto.
1096 (avx512f_vec_dup_gpr<mode>): Ditto.
1097 (clz<mode>2): Ditto.
1098 (conflict<mode>): Ditto.
1099 (REDUC_SMINMAX_MODE): Extened with wider modes.
1100 (reduc_<code>_<mode>): Ditto.
1101 (vlshr<mode>3): Ditto.
1102 (vashl<mode>3): Ditto.
1104 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1105 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1106 Sergey Lega <sergey.s.lega@intel.com>
1107 Anna Tikhonova <anna.tikhonova@intel.com>
1108 Ilya Tocar <ilya.tocar@intel.com>
1109 Andrey Turetskiy <andrey.turetskiy@intel.com>
1110 Ilya Verbin <ilya.verbin@intel.com>
1111 Kirill Yukhin <kirill.yukhin@intel.com>
1112 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1114 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1117 (avx512fmaskmode): Ditto.
1118 (bcstscalarsuff): Ditto.
1119 (avx512f_blendm<mode>): Ditto.
1120 (cmp_imm_predicate): Ditto.
1121 (avx512f_cmp<mode>3): Ditto.
1122 (avx512f_vec_dup<mode>): Ditto.
1123 (avx512f_vec_dup_mem<mode>): Ditto.
1124 (avx512f_vpermi2var<mode>3): Ditto.
1125 (avx512f_vpermt2var<mode>3): Ditto.
1126 (vec_init<mode>): Ditto.
1127 (avx512f_gathersi<mode>): Ditto.
1128 (*avx512f_gathersi<mode>): Ditto.
1129 (*avx512f_gathersi<mode>_2): Ditto.
1130 (avx512f_gatherdi<mode>): Ditto.
1131 (*avx512f_gatherdi<mode>): Ditto.
1132 (*avx512f_gatherdi<mode>_2): Ditto.
1133 (avx512f_scattersi<mode>): Ditto.
1134 (*avx512f_scattersi<mode>): Ditto.
1135 (avx512f_scatterdi<mode>): Ditto.
1136 (*avx512f_scatterdi<mode>): Ditto.
1137 (sseintprefix): Extened with wider modes.
1138 (VEC_GATHER_IDXSI): Ditto.
1139 (VEC_GATHER_IDXDI): Ditto.
1140 (VEC_GATHER_SRCDI): Ditto.
1142 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1143 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1145 * config/arm/t-aprofile: New file.
1146 * config.gcc: Handle --with-multilib-list option.
1148 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1150 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
1153 2013-10-15 Richard Biener <rguenther@suse.de>
1155 * tree-loop-distribution.c (build_empty_rdg): Inline into
1157 (rdg_flag_vertex): Inline into single user.
1158 (rdg_flag_vertex_and_dependent): Likewise.
1159 (build_rdg_partition_for_vertex): Remove processed bitmap.
1160 (rdg_build_partitions): Simplify.
1162 2013-10-15 Richard Biener <rguenther@suse.de>
1164 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1165 Restructure forwarding through conversions and copies to
1166 avoid performing copy-propagation the wrong way. Adjust
1167 recursion invocations.
1168 (forward_propagate_addr_expr): Add argument stating if we
1169 are recursing from a single-use.
1170 (ssa_forward_propagate_and_combine): Adjust.
1172 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1174 * dumpfile.h (gcc::dump_manager): New class, to hold state
1175 relating to dumpfile management.
1176 (get_dump_file_name): Remove in favor of method of dump_manager.
1177 (dump_initialized_p): Likewise.
1178 (dump_start): Likewise.
1179 (dump_finish): Likewise.
1180 (dump_switch_p): Likewise.
1181 (dump_register): Likewise.
1182 (get_dump_file_info): Likewise.
1183 * context.c (gcc::context::context): Construct the dump_manager
1185 * context.h (gcc::context::get_dumps): New.
1186 (gcc::context::m_dumps): New.
1187 * coverage.c (coverage_init): Port to dump_manager API.
1188 * dumpfile.c (extra_dump_files): Convert to field of
1190 (extra_dump_files_in_use): Likewise.
1191 (extra_dump_files_alloced): Likewise.
1192 (gcc::dump_manager::dump_manager): New.
1193 (dump_register): Convert to...
1194 (gcc::dump_manager::dump_register): ...method, replacing
1195 function-static next_dump with m_next_dump field.
1196 (get_dump_file_info): Convert to...
1197 (gcc::dump_manager::get_dump_file_info): ...method.
1198 (get_dump_file_name): Convert to...
1199 (gcc::dump_manager::get_dump_file_name): ...method.
1200 (dump_start): Convert to...
1201 (gcc::dump_manager::dump_start): ...method.
1202 (dump_finish): Convert to...
1203 (gcc::dump_manager::dump_finish): ...method.
1204 (dump_begin): Replace body with...
1205 (gcc::dump_manager::dump_begin): ...new method.
1206 (dump_phase_enabled_p): Convert to...
1207 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1208 (dump_phase_enabled_p): Convert to...
1209 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1210 (dump_initialized_p): Convert to...
1211 (gcc::dump_manager::dump_initialized_p): ...method.
1212 (dump_flag_name): Replace body with...
1213 (gcc::dump_manager::dump_flag_name): ...new method.
1214 (dump_enable_all): Convert to...
1215 (gcc::dump_manager::dump_enable_all): ...new method.
1216 (opt_info_enable_passes): Convert to...
1217 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1218 (dump_switch_p_1): Convert to...
1219 (gcc::dump_manager::dump_switch_p_1): ...new method.
1220 (dump_switch_p): Convert to...
1221 (gcc::dump_manager::dump_switch_p): ...new method.
1222 (opt_info_switch_p): Port to dump_manager API.
1223 (enable_rtl_dump_file): Likewise.
1224 * opts-global.c (handle_common_deferred_options): Port to new
1226 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1227 (pass_manager::register_one_dump_file): Likewise.
1228 (pass_manager::register_pass): Likewise.
1229 (pass_init_dump_file): Likewise.
1230 (pass_fini_dump_file): Likewise.
1231 * statistics.c (statistics_early_init): Likewise.
1233 2013-10-14 Richard Biener <rguenther@suse.de>
1235 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1236 iterative_hash_canonical_type, gimple_canonical_type_hash,
1237 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1238 gimple_register_canonical_type, print_gimple_types_stats,
1239 free_gimple_type_tables): Move to lto/lto.c
1240 (gt-gimple.h): Do not include.
1241 * gimple.h (gimple_register_canonical_type,
1242 print_gimple_types_stats, free_gimple_type_tables): Remove.
1243 * Makefile.in (GTFILES): Remove gimple.c.
1245 2013-10-14 Travis Snoozy <quandary@remstate.com>
1248 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1249 scanning for msp430x targets.
1251 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1254 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1255 (registers_ok_for_ldd_peep): Move around.
1256 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1257 * config/sparc/sparc.md (widening peepholes): Use it.
1259 2013-10-14 Richard Biener <rguenther@suse.de>
1263 * gimple.c (iterative_hash_canonical_type): Make sure to
1264 record the hash into the correct hashtable slot.
1266 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1268 PR rtl-optimization/58662
1269 * combine.c (try_combine): Take into account death nodes on I2 when
1270 splitting a PARALLEL of two independent SETs. Fix dump message.
1272 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1275 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1276 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1277 * config/sh/t-sh (sh_treg_combine.o): New entry.
1278 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1279 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1280 (register_sh_passes): New function. Register sh_treg_combine pass.
1281 (sh_option_override): Invoke it.
1282 (sh_canonicalize_comparison): Handle op0_preserve_value.
1283 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1284 opportunities. Canonicalize branch condition.
1285 (nott): Allow only if pseudos can be created for non-SH2A.
1287 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1290 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1291 (ix86_expand_movmem): Replace copy_addr_to_reg with
1292 ix86_copy_addr_to_reg.
1293 (ix86_expand_setmem): Likewise.
1295 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1297 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1298 provided by ix86_fp_compare_mode instead of CCFPUmode.
1300 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1302 * config/aarch64/arm_neon.h
1303 (vtbx<1,3>_<psu>8): Fix register constriants.
1305 2013-10-11 Jeff Law <law@redhat.com>
1307 PR tree-optimization/58640
1308 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1309 threading paths that cross over two loop entry points.
1311 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1313 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1314 handle vector float as well.
1315 (*vsx_le_perm_load_v4si): Likewise.
1316 (*vsx_le_perm_store_v2di): Likewise.
1317 (*vsx_le_perm_store_v4si): Likewise.
1319 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1321 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1322 directly to circumvent subtract from splat{31} workaround.
1323 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1325 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1326 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1327 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1328 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1329 separate big and little endian logic.
1330 (*altivec_vperm_<mode>_internal): New define_insn.
1331 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1332 separate big and little endian logic.
1333 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1334 (vec_permv16qi): Add little endian logic.
1336 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1338 * doc/extend.texi (returns_nonnull): Remove arguments.
1340 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1341 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1342 Sergey Lega <sergey.s.lega@intel.com>
1343 Anna Tikhonova <anna.tikhonova@intel.com>
1344 Ilya Tocar <ilya.tocar@intel.com>
1345 Andrey Turetskiy <andrey.turetskiy@intel.com>
1346 Ilya Verbin <ilya.verbin@intel.com>
1347 Kirill Yukhin <kirill.yukhin@intel.com>
1348 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1350 * config/i386/sse.md (VI48F_256_512): New.
1351 (avx2_permvar<mode>): Change to ...
1352 (<avx2_avx512f>_permvar<mode>): This.
1354 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1355 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1356 Sergey Lega <sergey.s.lega@intel.com>
1357 Anna Tikhonova <anna.tikhonova@intel.com>
1358 Ilya Tocar <ilya.tocar@intel.com>
1359 Andrey Turetskiy <andrey.turetskiy@intel.com>
1360 Ilya Verbin <ilya.verbin@intel.com>
1361 Kirill Yukhin <kirill.yukhin@intel.com>
1362 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1364 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1365 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1366 * config/i386/sse.md (VI4_AVX): New.
1367 (sf2simodelower): Ditto.
1368 (sse2_cvtps2dq): Change to ...
1369 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1371 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1372 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1373 Sergey Lega <sergey.s.lega@intel.com>
1374 Anna Tikhonova <anna.tikhonova@intel.com>
1375 Ilya Tocar <ilya.tocar@intel.com>
1376 Andrey Turetskiy <andrey.turetskiy@intel.com>
1377 Ilya Verbin <ilya.verbin@intel.com>
1378 Kirill Yukhin <kirill.yukhin@intel.com>
1379 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1381 * config/i386/sse.md (V_512): New.
1383 (vcond<V_512:mode><VF_512:mode>): Ditto.
1384 (vcond<V_512:mode><VI_512:mode>): Ditto.
1385 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1387 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1388 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1389 Sergey Lega <sergey.s.lega@intel.com>
1390 Anna Tikhonova <anna.tikhonova@intel.com>
1391 Ilya Tocar <ilya.tocar@intel.com>
1392 Andrey Turetskiy <andrey.turetskiy@intel.com>
1393 Ilya Verbin <ilya.verbin@intel.com>
1394 Kirill Yukhin <kirill.yukhin@intel.com>
1395 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1397 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1398 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1400 (fma<mode>4): Removed condition.
1401 (fms<mode>4): Ditto.
1402 (fnma<mode>4): Ditto.
1403 (fnms<mode>4): Ditto.
1404 (fma4i_fmadd_<mode>): Ditto.
1405 (*fma_fmadd_<mode>): Ditto.
1406 (*fma_fmsub_<mode>): Ditto.
1407 (*fma_fnmadd_<mode>): Ditto.
1408 (*fma_fnmsub_<mode>): Ditto.
1409 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1410 (*fma_fmaddsub_<mode>): Ditto.
1411 (*fma_fmsubadd_<mode>): Ditto.
1412 (*fmai_fmadd_<mode>): Ditto.
1413 (*fmai_fmsub_<mode>): Ditto.
1414 (*fmai_fnmadd_<mode>): Ditto.
1415 (*fmai_fnmsub_<mode>): Ditto.
1417 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1418 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1419 Sergey Lega <sergey.s.lega@intel.com>
1420 Anna Tikhonova <anna.tikhonova@intel.com>
1421 Ilya Tocar <ilya.tocar@intel.com>
1422 Andrey Turetskiy <andrey.turetskiy@intel.com>
1423 Ilya Verbin <ilya.verbin@intel.com>
1424 Kirill Yukhin <kirill.yukhin@intel.com>
1425 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1427 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1428 (VI124_256): Changed to ...
1429 (VI124_256_48_512): This.
1430 (ssepackmode): Extended with wider modes.
1431 (<code><mode>3): Changed iterator.
1432 (*avx2_<code><mode>3): Ditto.
1433 (vec_pack_trunc_<mode>): Ditto.
1435 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1436 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1437 Sergey Lega <sergey.s.lega@intel.com>
1438 Anna Tikhonova <anna.tikhonova@intel.com>
1439 Ilya Tocar <ilya.tocar@intel.com>
1440 Andrey Turetskiy <andrey.turetskiy@intel.com>
1441 Ilya Verbin <ilya.verbin@intel.com>
1442 Kirill Yukhin <kirill.yukhin@intel.com>
1443 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1445 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1446 (VI8F_256_512): Ditto.
1447 (abs<mode>2): Changed iterator.
1448 (avx2_perm<mode>): Changed to ...
1449 (<avx2_avx512f>_perm<mode>): This.
1450 (avx2_perm<mode>_1): Changed to ...
1451 (<avx2_avx512f>_perm<mode>_1): This.
1453 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1454 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1455 Sergey Lega <sergey.s.lega@intel.com>
1456 Anna Tikhonova <anna.tikhonova@intel.com>
1457 Ilya Tocar <ilya.tocar@intel.com>
1458 Andrey Turetskiy <andrey.turetskiy@intel.com>
1459 Ilya Verbin <ilya.verbin@intel.com>
1460 Kirill Yukhin <kirill.yukhin@intel.com>
1461 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1463 * config/i386/sse.md (VI48_AVX512F): New.
1464 (VI48_AVX2): Changed to ...
1465 (VI48_AVX2_48_AVX512F): This.
1466 (avx2_ashrv<mode>): Changed to ...
1467 (<avx2_avx512f>_ashrv<mode>): This.
1468 (avx2_<shift_insn>v<mode>): Changed to ...
1469 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1471 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1472 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1473 Sergey Lega <sergey.s.lega@intel.com>
1474 Anna Tikhonova <anna.tikhonova@intel.com>
1475 Ilya Tocar <ilya.tocar@intel.com>
1476 Andrey Turetskiy <andrey.turetskiy@intel.com>
1477 Ilya Verbin <ilya.verbin@intel.com>
1478 Kirill Yukhin <kirill.yukhin@intel.com>
1479 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1481 * config/i386/sse.md (VI4_AVX512F): New.
1482 (VI8_AVX2_AVX512F): Ditto.
1483 (mul<mode>3): Extended with wider modes.
1484 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1485 (mul<mode>3): Ditto.
1486 (vec_widen_<s>mult_odd_<mode>): Ditto.
1488 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1489 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1490 Sergey Lega <sergey.s.lega@intel.com>
1491 Anna Tikhonova <anna.tikhonova@intel.com>
1492 Ilya Tocar <ilya.tocar@intel.com>
1493 Andrey Turetskiy <andrey.turetskiy@intel.com>
1494 Ilya Verbin <ilya.verbin@intel.com>
1495 Kirill Yukhin <kirill.yukhin@intel.com>
1496 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1498 * config/i386/sse.md (VI2_AVX512F): New.
1499 (VI124_AVX512F): Ditto.
1500 (sseunpackmode): Extended with wider modes.
1501 (sseunpackfltmode): Ditto.
1502 (vec_unpacks_float_hi_<mode>): Ditto.
1503 (vec_unpacks_float_lo_<mode>): Ditto.
1504 (vec_unpacku_float_hi_<mode>): Ditto.
1505 (vec_unpacku_float_lo_<mode>): Ditto.
1506 (vec_unpacks_lo_<mode>): Ditto.
1507 (vec_unpacks_hi_<mode>): Ditto.
1508 (vec_unpacku_lo_<mode>): Ditto.
1509 (vec_unpacku_hi_<mode>): Ditto.
1511 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1512 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1513 Sergey Lega <sergey.s.lega@intel.com>
1514 Anna Tikhonova <anna.tikhonova@intel.com>
1515 Ilya Tocar <ilya.tocar@intel.com>
1516 Andrey Turetskiy <andrey.turetskiy@intel.com>
1517 Ilya Verbin <ilya.verbin@intel.com>
1518 Kirill Yukhin <kirill.yukhin@intel.com>
1519 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1521 * config/i386/i386.md (multdiv): New.
1522 (multdiv_mnemonic): Ditto.
1523 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1524 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1525 (<sse>_vmdiv<mode>3): Removed.
1527 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1528 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1529 Sergey Lega <sergey.s.lega@intel.com>
1530 Anna Tikhonova <anna.tikhonova@intel.com>
1531 Ilya Tocar <ilya.tocar@intel.com>
1532 Andrey Turetskiy <andrey.turetskiy@intel.com>
1533 Ilya Verbin <ilya.verbin@intel.com>
1534 Kirill Yukhin <kirill.yukhin@intel.com>
1535 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1537 * config/i386/sse.md (V): Extended with wider modes.
1539 (ssehalfvecmode): Ditto.
1541 (ssepackfltmode): Ditto.
1542 (avx_vec_concat<mode>): Ditto.
1543 (V_256_512): New iterator.
1544 (VF2_512_256): Ditto.
1545 (si2dfmode): New attribute.
1546 (si2dfmodelower): Ditto.
1548 (concat_tg_mode): Ditto.
1549 (floatv4siv4df2): Changed to ...
1550 (float<si2dfmodelower><mode>2): This.
1551 (avx_cvtps2pd256): Changed to ...
1552 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1553 (vec_pack_trunc_v4df): Changed to ...
1554 (vec_pack_trunc_<mode>): This.
1555 (avx_vpermil<mode>): Changed to ...
1556 (<sse2_avx_avx512f>_vpermil<mode>): This.
1557 (<fixsuffix>fix_truncv8dfv8si2): New.
1558 (vec_pack_sfix_trunc_v8df): Ditto.
1559 (avx512f_rndscale<mode>): Ditto.
1560 (avx512f_roundpd512): Ditto.
1561 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1563 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1564 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1565 Sergey Lega <sergey.s.lega@intel.com>
1566 Anna Tikhonova <anna.tikhonova@intel.com>
1567 Ilya Tocar <ilya.tocar@intel.com>
1568 Andrey Turetskiy <andrey.turetskiy@intel.com>
1569 Ilya Verbin <ilya.verbin@intel.com>
1570 Kirill Yukhin <kirill.yukhin@intel.com>
1571 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1573 * config/i386/i386.md (any_fix): New iterator.
1574 (fixsuffix): New attribute.
1575 * config/i386/sse.md (VF1): Extened with wider modes.
1579 (sseintvecmodelower): Ditto.
1580 (ssescalarmode): Ditto.
1581 (ssescalarnum): Ditto.
1583 (ssexmmmode): Ditto.
1584 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1585 (<sse>_rcp<mode>2): Change iterator.
1586 (rsqrt<mode>2): Ditto.
1587 (<sse>_rsqrt<mode>2): Ditto.
1588 (avx2_vec_dup<mode>): Ditto.
1589 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1590 (round<mode>2_sfix): Ditto.
1591 (avx2_pbroadcast<mode>): Ditto.
1592 (*andnot<mode>3): Handle XI mode.
1593 (*<code><mode>3): Ditto.
1594 (AVXTOSSEMODE): Removed.
1595 (avx_vpermil<mode>): Changed to ...
1596 (<sse2_avx_avx512f>_vpermil<mode>): This.
1598 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1599 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1600 Sergey Lega <sergey.s.lega@intel.com>
1601 Anna Tikhonova <anna.tikhonova@intel.com>
1602 Ilya Tocar <ilya.tocar@intel.com>
1603 Andrey Turetskiy <andrey.turetskiy@intel.com>
1604 Ilya Verbin <ilya.verbin@intel.com>
1605 Kirill Yukhin <kirill.yukhin@intel.com>
1606 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1608 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1609 (<sse>_comi): Ditto.
1610 (<sse>_ucomi): Ditto.
1611 (sse_cvtss2siq_2): Ditto.
1612 (sse2_cvtsd2si): Ditto.
1613 (sse2_cvtsd2siq): Ditto.
1614 (sse2_cvttsd2si): Ditto.
1615 (sse2_cvttsd2siq): Ditto.
1616 (<shift_insn><mode>3): Ditto.
1617 (sse2_cvtsi2sdq): Update constraint and prefix.
1618 (sse_cvtsi2ss): Update prefix.
1619 (sse_cvtsi2ssq): Ditto.
1621 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1623 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1624 ignore internal calls.
1626 2013-10-11 Richard Biener <rguenther@suse.de>
1628 * tree-pretty-print.c (dump_generic_node): Allow to dump
1629 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1630 (OVF) if TREE_OVERFLOW is set.
1632 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1634 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1636 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1637 * gimple.h: Likewise.
1639 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1641 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1644 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1646 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1647 * configure: Regenerate.
1649 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1651 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1652 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1653 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1654 similarly pp_right_paren.
1655 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1656 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1657 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1658 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1660 (omp_declare_simd_clauses_equal,
1661 omp_remove_redundant_declare_simd_attrs): New functions.
1662 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1663 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1664 * tree.h (OMP_LOOP_CHECK): Define.
1665 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1666 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1667 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1668 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1669 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1670 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1671 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1672 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1673 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1674 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1675 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1676 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1677 (omp_remove_redundant_declare_simd_attrs): New prototype.
1678 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1679 GIMPLE_OMP_TEAMS): New codes.
1680 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1681 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1683 (target_nesting_level): New variable.
1684 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1685 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1686 for collapse > 1 static schedule unless ordered.
1687 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1688 (determine_parallel_type): Adjust get_ws_args_for caller.
1689 (install_var_field): Handle mask & 4 for double indirection.
1690 (scan_sharing_clauses): Ignore shared clause on teams construct.
1691 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1692 (create_omp_child_function): If inside target or declare target
1693 constructs, set "omp declare target" attribute on the child function.
1694 (find_combined_for): New function.
1695 (scan_omp_parallel): Handle combined loops.
1696 (scan_omp_target, scan_omp_teams): New functions.
1697 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1698 restrictions and set ctx->cancellable for cancellable constructs.
1699 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1700 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1701 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1702 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1703 (omp_clause_aligned_alignment): New function.
1704 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1706 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1707 on teams constructs. Handle user defined reductions and new
1709 (lower_reduction_clauses): Don't set placeholder to address of ref
1710 if it has already the right type.
1711 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1712 (expand_parallel_call): Use the new non-_start suffixed builtins,
1713 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1714 and GOMP_parallel_end after the call.
1715 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1716 (expand_omp_for_init_counts): Handle combined loops.
1717 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1719 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1720 end of cancellable loops.
1721 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1722 Likewise. Handle collapse > 1 loops.
1723 (expand_omp_simd): Handle combined loops.
1724 (expand_omp_for): Add inner_stmt argument, adjust callers of
1725 expand_omp_for* functions, use expand_omp_for_static*chunk even
1726 for collapse > 1 unless ordered.
1727 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1728 of cancellable sections.
1729 (expand_omp_single): Remove need_barrier variable, just rely on
1730 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1731 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1732 (expand_omp_atomic_load, expand_omp_atomic_store,
1733 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1734 (expand_omp_target): New function.
1735 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1736 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1737 (build_omp_regions_1): Immediately close region for
1738 GF_OMP_TARGET_KIND_UPDATE.
1739 (maybe_add_implicit_barrier_cancel): New function.
1740 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1742 (lower_omp_single): Likewise. Add clobber after the barrier.
1743 (lower_omp_taskgroup): New function.
1744 (lower_omp_for): Handle combined loops. Adjust
1745 lower_rec_input_clauses caller. Handle cancellation.
1746 (lower_depend_clauses): New function.
1747 (lower_omp_taskreg): Lower depend clauses. Adjust
1748 lower_rec_input_clauses caller. Add clobber after the call. Handle
1750 (lower_omp_target, lower_omp_teams): New functions.
1751 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1752 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1753 and GOMP_cancellation_point calls.
1754 (lower_omp): Fold stmts inside of target region.
1755 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1756 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1757 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1758 (BT_FN_VOID_OMPFN_PTR_UINT,
1759 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1760 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1761 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1762 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1763 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1764 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1765 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1766 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1767 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1768 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1769 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1770 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1771 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1772 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1773 BUILT_IN_GOMP_PARALLEL_END.
1774 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1775 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1776 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1777 GF_OMP_FOR_KIND_DISTRIBUTE.
1778 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1779 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1780 (dump_gimple_omp_return): Print lhs if it has any.
1781 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1782 gimple_omp_atomic_seq_cst_p.
1783 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1784 and GIMPLE_OMP_TEAMS.
1785 * langhooks.c (lhd_omp_mappable_type): New function.
1786 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1787 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1789 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1790 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1791 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1793 (struct gimplify_omp_ctx): Add combined_loop field.
1794 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1795 on stmts inside of target region.
1796 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1797 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1798 ORT_TARGET and ORT_TARGET_DATA.
1799 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1801 (omp_notice_threadprivate_variable): Complain about threadprivate
1802 variables in target region.
1803 (omp_notice_variable): Complain about vars with non-mappable type
1804 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1805 (omp_check_private): Ignore ORT_TARGET* regions.
1806 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1807 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1808 (find_combined_omp_for): New function.
1809 (gimplify_omp_for): Handle gimplification of combined loops.
1810 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1812 (gimplify_omp_target_update): New function.
1813 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1814 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1815 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1816 (gimplify_body): If fndecl has "omp declare target" attribute, add
1817 implicit ORT_TARGET context around it.
1818 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1819 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1820 * tree-nested.c (convert_nonlocal_reference_stmt,
1821 convert_local_reference_stmt, convert_gimple_call): Handle
1822 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1823 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1824 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1825 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1826 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1827 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1828 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1829 BUILT_IN_GOMP_LOOP_END_CANCEL,
1830 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1831 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1832 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1833 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1834 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1835 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1836 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1837 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1838 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1839 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1840 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1841 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1842 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1843 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1844 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1845 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1846 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1847 gimple_build_omp_teams): New functions.
1848 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1849 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1850 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1851 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1852 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1853 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1854 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1855 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1856 GF_OMP_ATOMIC_SEQ_CST): New.
1857 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1858 gimple_build_omp_teams): New prototypes.
1859 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1860 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1861 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1862 GIMPLE_OMP_SINGLE as end of range.
1863 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1864 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1865 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1866 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1867 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1868 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1869 gimple_omp_target_kind, gimple_omp_target_set_kind,
1870 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1871 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1872 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1873 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1874 gimple_omp_teams_set_clauses): New inlines.
1875 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1876 and GIMPLE_OMP_TASKGROUP.
1877 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
1878 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1879 enum omp_clause_proc_bind_kind): New.
1880 (union omp_clause_subcode): Add depend_kind, map_kind and
1881 proc_bind_kind fields.
1882 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1883 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1884 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1885 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1886 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1888 2013-10-10 Teresa Johnson <tejohnson@google.com>
1890 * predict.c (tree_estimate_probability): Add new parameter
1891 for estimate_bb_frequencies.
1892 (estimate_bb_frequencies): Add new parameter to force estimation.
1893 (rebuild_frequencies): When max frequency in function is small,
1894 recompute counts from frequencies.
1895 * predict.h (estimate_bb_frequencies): New parameter.
1897 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1899 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1900 optimizations are disabled.
1902 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1904 * coverage.c (coverage_finish): Fix leak of da_file_name.
1906 2013-10-10 Jan Hubicka <jh@suse.cz>
1908 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1909 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1912 2013-10-10 Jakub Jelinek <jakub@redhat.com>
1915 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1916 if any labels are in FALLTHRU_BB, use a special label emitted
1917 immediately after the asm goto insn rather than label_rtx
1919 (expand_asm_stmt): Adjust caller.
1920 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1921 edge if the last insn in predecessor is a jump with single successor,
1922 but it isn't simplejump_p.
1924 2013-10-10 Richard Biener <rguenther@suse.de>
1926 PR tree-optimization/58656
1927 * tree-ssa-pre.c (phi_translate): Do not cache failed
1930 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1932 * gimplify.c: Include expr.h and tm_p.h for targets with special
1933 va-arg padding requirements.
1935 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1937 * tree-flow.h: Move some prototypes to gimple.h.
1938 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1939 * gimple.h: Relocate some prototypes from tree-flow.h
1940 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1942 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1943 (build_va_arg_indirect_ref): Relocate and make static.
1944 (std_gimplify_va_arg_expr): Relocate here.
1945 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1946 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1948 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1950 * doc/md.texi: Document the mnemonic attribute.
1952 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1955 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1956 (set_attr_alternative x ...) when searching for user defined
1959 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1961 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1963 2013-10-09 Easwaran Raman <eraman@google.com>
1965 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1966 * cfgexpand.c (defer_stack_allocation): ...use here
1967 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1969 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1971 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1972 (operand_equal_for_value_replacement): New function, extracted from
1973 value_replacement and enhanced to catch more cases.
1974 (value_replacement): Use operand_equal_for_value_replacement.
1976 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1978 * loop-doloop.c (doloop_modify, doloop_optimize): Use
1979 get_max_loop_iterations.
1981 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1983 * config/arm/aarch-common.c (arm_early_load_addr_dep):
1984 Place comment above function.
1986 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1988 * tree-flow.h: Remove all remaining prototypes, enums and structs that
1989 are not related to tree-cfg.c.
1990 * tree-ssa-address.h: New file. Relocate prototypes.
1991 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
1992 (addr_for_mem_ref): New. Combine call to get_address_description and
1993 return addr_for_mem_ref.
1994 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
1995 * tree-ssa-live.h: Adjust prototypes.
1996 * passes.c: Include tree-ssa-live.h.
1997 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
1998 * graphite.c (graphite_transform_loops): Make static.
1999 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2000 make_pass_graphite, pass_data_graphite_transforms,
2001 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2002 * ipa-pure-const.c (warn_function_noreturn): Make static.
2003 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2004 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2005 Relocate from tree-cfg.c
2006 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2008 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2009 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2010 Move to ipa-pure-const.c.
2011 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2012 Relocate from tree-optimize.c.
2013 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2014 make_pass_fixup_cfg): Move to tree-cfg.c.
2015 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2016 Relocate some prototypes.
2017 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2018 * tree-dump.c (dump_function_to_file): Remove prototype.
2019 Add tree-flow.h to the include file.
2020 * tree-dump.h: Remove prototype.
2021 * tree-parloops.h: New File. Add prototypes.
2022 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2023 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2024 from tree-ssa-loop.c.
2025 * tree-predcom.c (run_tree_predictive_commoning,
2026 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2027 Relocate here from tree-ssa-loop.c.
2028 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2029 ssa_name_values.release ().
2030 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2031 (ssa_name_values): Relocate from tree-flow.h.
2032 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2033 * tree-ssa-loop.c (run_tree_predictive_commoning,
2034 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2035 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2036 make_pass_graphite, pass_data_graphite_transforms,
2037 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2038 tree_parallelize_loops, pass_data_parallelize_loops,
2039 make_pass_parallelize_loops): Move to other files.
2040 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2042 * tree.h: Remove prototypes from tree-address.c.
2044 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2046 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2047 (struct int_tree_map): Move to tree-hasher.h
2048 (SCALE, LABEL, PERCENT): Move to gimple.h
2049 * tree-flow-inline.h: Delete. Move functions to other files.
2050 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2051 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2052 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2053 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2054 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2055 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2056 * tree-hasher.h: Don't include tree-flow.h.
2057 (struct int_tree_map): Relocate from tree-flow.h.
2058 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2059 tree-flow-inline.h and make static.
2060 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2062 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2063 tree-flow-inline.h and make static.
2064 * tree.h (is_global_var, may_be_aliased): Relocate from
2066 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2067 * value-prof.c: No longer include tree-flow-inline.h.
2068 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2070 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2072 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2073 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2074 (enum move_pos): Move to tree-ssa-loop-im.h
2075 * cfgloop.h: Move some prototypes.
2076 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2077 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2078 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2079 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2080 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2081 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2082 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2083 (enum move_pos): Relocate here.
2084 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2086 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2087 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2088 make_pass_lim): Relocate here from tree-ssa-loop.c.
2089 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2091 (loop_edge_to_cancel, unloop_loops): Make static.
2092 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2093 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2094 (tree_complete_unroll, gate_tree_complete_unroll,
2095 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2096 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2097 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2098 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2099 (stmt_invariant_in_loop_p): Remove unused function.
2100 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2101 * tree-ssa-loop-manip.h: New file. Add prototypes.
2102 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2103 (gcov_type_to_double_int): Move to cfgloop.h.
2104 (double_int_cmp, bound_index,
2105 estimate_numbers_of_iterations_loop): Make static.
2106 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2107 (max_loop_iterations): Factor out get_max_loop_iterations.
2108 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2110 * tree-ssa-loop-niter.h: New file. Add prototypes.
2111 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2112 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2113 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2114 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2115 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2116 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2117 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2118 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2119 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2120 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2121 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2122 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2123 pass_data_complete_unroll, make_pass_complete_unroll,
2124 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2125 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2126 tree-ssa-loop-ivcanon.c.
2127 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2128 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2129 tree-ssa-loop-prefetch.c.
2130 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2132 (get_lsm_tmp_name): Relocate and add suffix parameter.
2133 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2134 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2135 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2136 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2137 (get_estimated_loop_iterations): Factor out accessor from
2138 estimated_loop_iterations in tree-ssa-loop-niter.c.
2139 (get_max_loop_iterations): Factor out accessor from
2140 _max_loop_iterations in tree-ssa-niter.c.
2141 * loop-unroll.c (decide_unroll_constant_iterations,
2142 decide_unroll_runtime_iterations, decide_peel_simple,
2143 decide_unroll_stupid): Use new get_* accessors.
2145 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2147 PR tree-optimization/20318
2148 * doc/extend.texi (returns_nonnull): New function attribute.
2149 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2151 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2152 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2154 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2157 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2158 false if both components are bitfields.
2160 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2162 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2163 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2164 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2165 * config/aarch64/aarch64.h
2166 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2167 * config/aarch64/aarch64-simd-builtins.def
2168 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2170 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2172 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2174 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2176 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2178 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2180 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2181 (vneg_f64): New intrinsic.
2182 (vneg_s8): Asm replaced with C.
2183 (vneg_s16): Likewise.
2184 (vneg_s32): Likewise.
2185 (vneg_s64): New intrinsic.
2186 (vnegq_f32): Asm replaced with C.
2187 (vnegq_f64): Likewise.
2188 (vnegq_s8): Likewise.
2189 (vnegq_s16): Likewise.
2190 (vnegq_s32): Likewise.
2191 (vnegq_s64): Likewise.
2193 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2195 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2197 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2199 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2200 packed stack special handling.
2201 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2202 back to fixed stack slots for FPRs saved due to stdarg.
2204 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2206 * config/s390/s390.c (s390_frame_info): Restructure function.
2208 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2210 * config/s390/s390.c (struct s390_frame_layout): New field
2212 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2213 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2214 regs_ever_clobbered to char*.
2215 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2216 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2217 (s390_register_info): Enable FPR save slots. Move/Copy some
2218 functionality into ...
2219 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2220 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2222 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2223 not imply a stack frame.
2224 (s390_init_frame_layout): Remove variable clobbered_regs.
2225 (s390_update_register_info): Remove function.
2226 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2227 cannot be used for register renaming.
2228 (s390_hard_regno_scratch_ok): New function.
2229 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2230 (s390_initial_elimination_offset): Change offset calculation of
2231 the return address pointer.
2232 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2233 (restore_gprs): Set frame related flag.
2234 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2235 (s390_emit_prologue): Call s390_register_info instead of
2236 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2237 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2238 (s390_optimize_prologue): Call s390_optimize_register_info.
2239 Try to remove also FPR slot save/restore INSNs. Remove frame
2240 related flags from restore INSNs.
2242 2013-10-08 DJ Delorie <dj@redhat.com>
2244 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2247 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2248 avoid conflict with the MI use of %c.
2249 * config/rl78/rl78-real.md: change %c to %C throughout.
2250 * config/rl78/rl78-virt.md: Likewise.
2252 2013-10-08 Jan Hubicka <jh@suse.cz>
2254 * config/i386/i386.c (ix86_option_override_internal): Switch
2255 to SSE math for -ffast-math when target ISA supports SSE2.
2257 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2259 * tree-flow.h: Remove some prototypes.
2260 * tree.h: Remove some protypes, add a couple.
2261 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2262 using_eh_for_cleanups_p): Add interface routines for front ends.
2263 * tree-eh.h: New file. Add protoptyes.
2264 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2265 (add_stmt_to_eh_lp_fn): Make static.
2266 (lower_try_finally): Use new using_eh_for_cleanups_p.
2267 * emit-rtl.c: Include tree-eh.h.
2268 * gimple.h: Include tree-eh.h.
2270 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2272 PR tree-optimization/58480
2273 * tree-vrp.c (infer_nonnull_range): New function.
2274 (infer_value_range): Call infer_nonnull_range.
2276 2013-10-08 Dehao Chen <dehao@google.com>
2278 PR tree-optimization/58619
2279 * tree-inline.c (copy_phis_for_bb): Combine location data
2282 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2285 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2286 RTX_FRAME_RELATED_P on INSN.
2288 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2290 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2291 (altivec_expand_vec_perm_const): Call it.
2293 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2295 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2296 sequences for LE VSX loads and stores at expand time.
2297 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2299 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2300 (rs6000_gen_le_vsx_permute): New.
2301 (rs6000_gen_le_vsx_load): New.
2302 (rs6000_gen_le_vsx_store): New.
2303 (rs6000_gen_le_vsx_move): New.
2304 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2305 (*vsx_le_perm_load_v4si): New.
2306 (*vsx_le_perm_load_v8hi): New.
2307 (*vsx_le_perm_load_v16qi): New.
2308 (*vsx_le_perm_store_v2di): New.
2309 (*vsx_le_perm_store_v4si): New.
2310 (*vsx_le_perm_store_v8hi): New.
2311 (*vsx_le_perm_store_v16qi): New.
2312 (*vsx_xxpermdi2_le_<mode>): New.
2313 (*vsx_xxpermdi4_le_<mode>): New.
2314 (*vsx_xxpermdi8_le_V8HI): New.
2315 (*vsx_xxpermdi16_le_V16QI): New.
2316 (*vsx_lxvd2x2_le_<mode>): New.
2317 (*vsx_lxvd2x4_le_<mode>): New.
2318 (*vsx_lxvd2x8_le_V8HI): New.
2319 (*vsx_lxvd2x16_le_V16QI): New.
2320 (*vsx_stxvd2x2_le_<mode>): New.
2321 (*vsx_stxvd2x4_le_<mode>): New.
2322 (*vsx_stxvd2x8_le_V8HI): New.
2323 (*vsx_stxvd2x16_le_V16QI): New.
2325 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2327 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2329 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2331 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2332 loop to work also for 31bit ABI.
2333 Save the stack pointer for frame_size > 0.
2335 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2337 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2338 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2339 constraint letters from expanders.
2340 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2341 retry count to general_operand.
2342 ("tabort"): Give operand 0 a mode.
2343 ("tabort_1"): Add mode and constraint letter for operand 0.
2344 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2346 2013-10-04 Jeff Law <law@redhat.com>
2348 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2350 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2352 (thread_across_edge): Here. Call it.
2354 2013-10-04 Cary Coutant <ccoutant@google.com>
2356 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2357 discarding a location list expression (or a piece of one).
2359 2013-10-03 Jan Hubicka <jh@suse.cz>
2361 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2362 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2363 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2365 2013-10-03 Jan Hubicka <jh@suse.cz>
2367 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2368 accumulate-outgoing-args when producing unwind info.
2370 2013-10-03 Wei Mi <wmi@google.com>
2372 * lra-constraints.c (insert_move_for_subreg): New function
2373 extracted from simplify_operand_subreg.
2374 (simplify_operand_subreg): Add reload for paradoxical subreg.
2376 2013-10-03 Rong Xu <xur@google.com>
2378 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2379 the candidate of builtin_expect such that we should fix the
2380 size/time estimation.
2381 (estimate_function_body_sizes): Do the acutally size/time fix-up
2384 2013-10-03 Rong Xu <xur@google.com>
2386 * predict.c (tree_predict_by_opcode): Get the probability
2387 for builtin_expect from param builtin_expect_probability.
2388 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2389 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2390 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2392 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2395 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2396 * calls.c (alloca_call_p): Use get_callee_fndecl.
2397 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2398 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2400 (vrp_visit_stmt): Remove duplicated code.
2402 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2404 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2405 ceildf2, btruncdf2, instead of vsx_* name.
2407 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2408 iterators to only do V2DF and V4SF here. Move the DF code to
2409 rs6000.md where it is combined with SF mode. Replace <VSv> with
2410 just 'v' since only vector operations are handled with these insns
2411 after moving the DF support to rs6000.md.
2412 (vsx_sub<mode>3): Likewise.
2413 (vsx_mul<mode>3): Likewise.
2414 (vsx_div<mode>3): Likewise.
2415 (vsx_fre<mode>2): Likewise.
2416 (vsx_neg<mode>2): Likewise.
2417 (vsx_abs<mode>2): Likewise.
2418 (vsx_nabs<mode>2): Likewise.
2419 (vsx_smax<mode>3): Likewise.
2420 (vsx_smin<mode>3): Likewise.
2421 (vsx_sqrt<mode>2): Likewise.
2422 (vsx_rsqrte<mode>2): Likewise.
2423 (vsx_fms<mode>4): Likewise.
2424 (vsx_nfma<mode>4): Likewise.
2425 (vsx_copysign<mode>3): Likewise.
2426 (vsx_btrunc<mode>2): Likewise.
2427 (vsx_floor<mode>2): Likewise.
2428 (vsx_ceil<mode>2): Likewise.
2429 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2430 (vsx_sminsf3): Likewise.
2431 (vsx_fmadf4): Likewise.
2432 (vsx_fmsdf4): Likewise.
2433 (vsx_nfmadf4): Likewise.
2434 (vsx_nfmsdf4): Likewise.
2435 (vsx_cmpdf_internal1): Likewise.
2437 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2438 simpler to select whether a target has SPE or traditional floating
2439 point support in iterators.
2440 (TARGET_DF_SPE): Likewise.
2441 (TARGET_SF_FPR): Likewise.
2442 (TARGET_DF_FPR): Likewise.
2443 (TARGET_SF_INSN): Macros to say whether floating point support
2444 exists for a given operation for expanders.
2445 (TARGET_DF_INSN): Likewise.
2447 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2448 combining of SF/DF mode operations, using both traditional and VSX
2456 (abs<mode>2): Combine SF/DF modes using traditional floating point
2457 instructions. Add support for using the upper DF registers with
2458 VSX support, and SF registers with power8-vector support. Update
2459 expanders for operations supported by both the SPE and traditional
2460 floating point units.
2461 (abs<mode>2_fpr): Likewise.
2462 (nabs<mode>2): Likewise.
2463 (nabs<mode>2_fpr): Likewise.
2464 (neg<mode>2): Likewise.
2465 (neg<mode>2_fpr): Likewise.
2466 (add<mode>3): Likewise.
2467 (add<mode>3_fpr): Likewise.
2468 (sub<mode>3): Likewise.
2469 (sub<mode>3_fpr): Likewise.
2470 (mul<mode>3): Likewise.
2471 (mul<mode>3_fpr): Likewise.
2472 (div<mode>3): Likewise.
2473 (div<mode>3_fpr): Likewise.
2474 (sqrt<mode>3): Likewise.
2475 (sqrt<mode>3_fpr): Likewise.
2476 (fre<Fs>): Likewise.
2477 (rsqrt<mode>2): Likewise.
2478 (cmp<mode>_fpr): Likewise.
2479 (smax<mode>3): Likewise.
2480 (smin<mode>3): Likewise.
2481 (smax<mode>3_vsx): Likewise.
2482 (smin<mode>3_vsx): Likewise.
2483 (negsf2): Delete SF operations that are merged with DF.
2490 (fmasf4_fpr): Likewise.
2491 (fmssf4_fpr): Likewise.
2492 (nfmasf4_fpr): Likewise.
2493 (nfmssf4_fpr): Likewise.
2494 (sqrtsf2): Likewise.
2495 (rsqrtsf_internal1): Likewise.
2496 (smaxsf3): Likewise.
2497 (sminsf3): Likewise.
2498 (cmpsf_internal1): Likewise.
2499 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2500 (negdf2): Delete DF operations that are merged with SF.
2502 (nabsdf2): Likewise.
2508 (rsqrtdf_internal1): Likewise.
2509 (fmadf4_fpr): Likewise.
2510 (fmsdf4_fpr): Likewise.
2511 (nfmadf4_fpr): Likewise.
2512 (nfmsdf4_fpr): Likewise.
2513 (sqrtdf2): Likewise.
2514 (smaxdf3): Likewise.
2515 (smindf3): Likewise.
2516 (cmpdf_internal1): Likewise.
2517 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2518 (btrunc<mode>2): Delete separate expander, and combine with the
2519 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2520 (btrunc<mode>2_fpr): Likewise.
2521 (ceil<mode>2): Likewise.
2522 (ceil<mode>2_fpr): Likewise.
2523 (floor<mode>2): Likewise.
2524 (floor<mode>2_fpr): Likewise.
2525 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2526 Add support for using the upper registers with VSX and
2527 power8-vector. Move insns to be closer to the define_expands. On
2528 VSX systems, prefer the traditional form of FMA over the VSX
2529 version, since the traditional form allows the target not to
2530 overlap with the inputs.
2531 (fms<mode>4_fpr): Likewise.
2532 (nfma<mode>4_fpr): Likewise.
2533 (nfms<mode>4_fpr): Likewise.
2535 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2536 Richard Earnshaw <richard.earnshaw@arm.com>
2538 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2539 (struct mult_cost_table): Likewise.
2540 (struct mem_cost_table): Likewise.
2541 (struct fp_cost_table): Likewise.
2542 (struct vector_cost_table): Likewise.
2543 (cpu_cost_table): Likewise.
2544 * config/arm/arm.opt (mold-rts-costs): New option.
2545 (mnew-generic-costs): Likewise.
2546 * config/arm/arm.c (generic_extra_costs): New table.
2547 (cortexa15_extra_costs): Likewise.
2548 (arm_slowmul_tune): Use NULL as new costs.
2549 (arm_fastmul_tune): Likewise.
2550 (arm_strongarm_tune): Likewise.
2551 (arm_xscale_tune): Likewise.
2552 (arm_9e_tune): Likewise.
2553 (arm_v6t2_tune): Likewise.
2554 (arm_cortex_a5_tune): Likewise.
2555 (arm_cortex_a9_tune): Likewise.
2556 (arm_v6m_tune): Likewise.
2557 (arm_fa726te_tune): Likewise.
2558 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2559 (arm_cortex_tune): Use generict_extra_costs.
2560 (shifter_op_p): New function.
2561 (arm_unspec_cost): Likewise.
2562 (LIBCALL_COST): Define.
2563 (arm_new_rtx_costs): New function.
2564 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2565 table is available. Use old costs otherwise unless mnew-generic-costs
2567 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2568 (cpu_cost_table): Declare.
2570 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2573 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2574 (*subs_mul_imm_<mode>)
2575 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2576 (*sub_<shift>_<mode>)
2577 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2578 Remove k constraint.
2580 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2582 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2584 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2586 2013-10-02 Teresa Johnson <tejohnson@google.com>
2588 * predict.c (probably_never_executed): New function.
2589 (probably_never_executed_bb_p): Invoke probably_never_executed.
2590 (probably_never_executed_edge_p): Ditto.
2591 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2592 Treat profile insanities conservatively.
2594 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2596 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2598 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2600 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2601 use it. Use smaller increase for scratch. Don't increase reject
2602 for early clobber scratch.
2603 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2604 setting eliminated regs except setting fp from hfp.
2605 (lra_eliminate): Check lra_insn_recog_data on NULL.
2607 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2610 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2611 setting -mvsx-timode by default until the underlying problem is fixed.
2612 (RS6000_CPU, power7 defaults): Likewise.
2614 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2616 * config/x-linux (host-linux.o): Remove header dependencies.
2617 Use $(COMPILE) and $(POSTCOMPILE).
2618 * config/t-linux-android (linux-android.o): Ditto.
2620 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2622 * Makefile.in (expmed.o-warn): Remove.
2624 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2626 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2627 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2629 2013-10-02 Teresa Johnson <tejohnson@google.com>
2631 * dojump.c (do_jump_1): Divide probability between
2632 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2634 2013-10-02 Tom Tromey <tromey@redhat.com>
2636 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2638 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2640 * tree-flow.h: Remove some prototypes.
2641 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2642 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2643 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2644 mark_virtual_phi_result_for_renaming): Relocate here.
2645 * tree-into-ssa.h: Add prototypes.
2646 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2647 single_pred_before_succ_order.
2648 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2649 (nonfreeing_call_p) Move to gimple.c.
2650 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2652 * basic-block.h (single_pred_before_succ_order): Add prototype.
2653 * gimple.c (nonfreeing_call_p): Relocate here.
2654 * gimple.h: Add prototype.
2655 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2656 * tree-ssa-dom.h: New file. Relocate prototypes here.
2657 * tree-ssa.h: Include tree-ssa-dom.h.
2659 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2661 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2662 Use $(COMPILE) and $(POSTCOMPILE).
2664 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2666 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2668 * tree-flow.h: Remove some prototypes.
2669 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2670 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2671 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2672 propagate_tree_value*): Move from here to...
2673 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2674 propagate_tree_value*): Relocate here.
2675 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2676 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2677 * gimple-fold.c: Remove gimple-fold.h from include list.
2678 * tree-vrp.c: Remove gimple-fold.h from include list.
2679 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2680 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2681 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2682 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2683 * sese.c: Add tree-ssa-propagate.h to include list.
2685 2013-10-02 Richard Biener <rguenther@suse.de>
2687 * tree-loop-distribution.c: Include tree-vectorizer.h for
2689 (enum partition_kind): Remove PKIND_REDUCTION.
2690 (struct partition_s): Remove has_writes member, add reduction_p member.
2691 (partition_alloc): Adjust.
2692 (partition_builtin_p): Likewise.
2693 (partition_has_writes): Remove.
2694 (partition_reduction_p): New function.
2695 (partition_merge_into): Likewise.
2696 (generate_code_for_partition): Commonize builtin partition
2698 (rdg_cannot_recompute_vertex_p): Remove.
2699 (already_processed_vertex_p): Likewise.
2700 (rdg_flag_vertex): Do not set has_writes.
2701 (classify_partition): Adjust.
2702 (rdg_build_partitions): Do not set has_writes, treat all
2703 partitions as useful.
2704 (distribute_loop): Record number of library calls generated. Adjust.
2705 (tree_loop_distribution): Report number of loops and library
2706 calls generated as opt-info.
2708 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2710 * tree-flow.h: Include new .h files. Move prototypes.
2711 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2712 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2713 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2714 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2715 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2716 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2717 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2718 * tree.h (get_ref_base_and_extent): Move prototype out.
2719 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2721 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2722 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2723 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2724 * tree-scalar-evolution.c: Include tree.h.
2725 * sese.c: Include tree.h.
2726 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2727 * dwarf2out.c: Include tree-dfa.h.
2728 * tree-chrec.c: Include tree.h.
2729 * tree-data-ref.c: Include tree.h.
2731 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2733 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2736 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2738 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2739 Use $(COMPILE) and $(POSTCOMPILE).
2740 (sol2-cxx.o): Likewise.
2741 (sol2-stubs.o): Likewise.
2743 * config/x-solaris (host-solaris.o): Likewise.
2745 * config/sparc/t-sparc (sparc.o): Remove.
2746 (sparc-c.o): Remove header dependencies.
2747 Use $(COMPILE) and $(POSTCOMPILE).
2748 * config/sparc/x-sparc: Likewise.
2750 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2752 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2753 * config/arc/arc700.md: Likewise.
2754 * config/arc/arc-modes.def: Likewise.
2755 * config/arc/arc-simd.h: Likewise.
2756 * config/arc/t-arc-uClibc: Likewise.
2757 * config/arc/t-arc-newlib: Likewise.
2759 2013-10-02 Renlin Li <renlin.li@arm.com>
2761 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2763 (aarch64_expand_epilogue): Likewise.
2765 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2766 Yufeng Zhang <yufeng.zhang@arm.com>
2768 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2770 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2771 'base_in' represent a conversion and legal_cast_p_1 holds; set
2772 'base_in' with the returned value from get_unwidened.
2774 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2776 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2777 plus_constant is not used.
2779 2013-10-01 Wei Mi <wmi@google.com>
2781 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2782 * config/i386/i386.md: Add define_peephole2 to
2783 break partial reg stall for cvtss2sd/cvtsd2ss.
2785 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2787 * config/arc/arc.c (pass_arc_ifcvt::clone):
2788 Update for ctxt_ -> m_ctxt change.
2790 2013-10-01 Jeff Law <law@redhat.com>
2792 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2793 outgoing_edge and intermediate_edge fields. Instead store the path.
2794 (redirection_data::hash): Hash on the last edge's destination index.
2795 (redirection_data::equal): Check the entire thread path.
2796 (lookup_redirectio_data): Corresponding changes.
2797 (create_edge_and_update_destination_phis): Likewise.
2798 (thread_single_edge): Likewise.
2800 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2801 Diego Novillo <dnovillo@google.com>
2803 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2804 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2805 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2806 versions of these patterns.
2808 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2809 (__builtin_arc_aligned): Likewise.
2811 * config/arc/arc.md: Expand adc_0 comment stating the intended
2812 purpose and why it isn't ready.
2813 Replace commented out call_value_via_label_mixed with a
2814 plain comment about bl_s.
2816 * config/arc/arc.c (stdio.h): Don't include directly.
2817 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2818 Remove if (1) condition.
2819 (arc_encode_section_info): Fix comment.
2821 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2823 * config/arc/arc.c (arc_conditional_register_usage):
2824 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2825 Also set reg_alloc_order for DMA config regs.
2827 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2828 Jeremy Bennett <jeremy.bennett@embecosm.com>
2830 * doc/install.texi (--with-cpu): Mention ARC.
2831 (arc-*-elf32): New paragraph.
2832 (arc-linux-uclibc): Likewise.
2833 * doc/md.texi (Machine Constraints): Add ARC part.
2834 * doc/invoke.texi: (menu): Add ARC Options.
2835 (Machine Dependent Options) <ARC Options>: Add synopsis.
2836 (node ARC Options): Add.
2837 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2838 (ARC Built-in Functions): New section defining
2839 generic ARC built-in functions.
2840 (ARC SIMD Built-in Functions): New section defining SIMD specific
2842 (Declaring Attributes of Functions): Extended
2843 description of short_call and long_call attributes for ARC and
2844 added index entries.
2846 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2847 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2848 Joern Rennecke <joern.rennecke@embecosm.com>
2849 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2850 Brendan Kehoe <brendan@zen.org>
2851 Michael Eager <eager@eagercon.com>
2852 Simon Cook <simon.cook@embecosm.com>
2853 Jeremy Bennett <jeremy.bennett@embecosm.com>
2855 * config/arc, common/config/arc: New directories.
2857 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2858 Brendan Kehoe <brendan@zen.org>
2859 Simon Cook <simon.cook@embecosm.com>
2861 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2863 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2865 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2866 * tree-ssa-coalesce.h: New. Move prototype to here.
2867 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2868 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2869 (gimple_can_coalesce_p): Move to...
2870 * gimple.c (gimple_can_coalesce_p): Here.
2872 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2874 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2875 (dump_decl_set): Move to gimple.c.
2876 * gimple.h: Don't include tree-ssa-operands.h.
2877 (dump_decl_set): Add prototype.
2878 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2879 Move to gimple-ssa.h.
2880 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2881 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2882 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2883 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2884 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2886 (dump_decl_set): Relocate here.
2887 * gimple-ssa.h: New file.
2888 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2889 Relocate from gimple.h.
2890 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2891 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2892 swap_tree_operands and remove non-ssa path.
2893 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2894 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2896 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2897 vect_is_simple_reduction_1): Use swap_ssa_operands.
2898 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2899 (enum need_phi_state): Move to tree-into-ssa.c.
2900 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2901 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2902 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2903 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2904 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2905 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2906 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2907 num_imm_uses): Move to ssa-iterators.h.
2908 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2909 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2911 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2912 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2913 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2914 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2915 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2916 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2917 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2918 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2919 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2920 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2921 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2922 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2923 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2924 (set_phi_nodes): Move to tree-phinodes.h.
2925 * tree-ssa-operands.h (enum ssa_op_iter_type,
2926 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2927 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2928 (dump_decl_set): Remove prototype.
2929 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2930 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2931 (set_phi_nodes): Relocate from tree-flow-inline.h.
2932 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2934 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2935 include list. Temporarily add gimple.h to include list.
2936 * ssa-iterators.h: New file.
2937 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2938 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2939 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2940 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2941 Relocate from tree-ssa-operands.h.
2942 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2943 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2944 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2945 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2946 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2947 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2948 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2949 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2950 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2951 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2952 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2953 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2954 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2955 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2956 Relocate from tree-flow-inline.h.
2957 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2959 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2962 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2963 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2964 Insert '\t' to output template.
2965 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2966 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2967 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2968 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2970 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
2972 * doc/install.texi (Host/target specific installation notes for GCC):
2973 Put @anchor before @heading.
2974 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2975 Use @email for email addresses.
2977 2013-10-01 Jeff Law <law@redhat.com>
2979 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
2980 a vec. Only delete the path if we create one without successfully
2981 registering a jump thread.
2982 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
2984 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
2985 (paths): New vector of jump threading paths.
2986 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
2987 (THREAD_PATH): New accessor macro for the entire thread path.
2988 (lookup_redirection_data): Get intermediate and final outgoing edge
2989 from the thread path.
2990 (create_edge_and_update_destination_phis): Copy the threading path.
2991 (ssa_fix_duplicate_block_edges): Get edges and block types from the
2992 jump threading path.
2993 (ssa_redirect_edges): Get edges and block types from the jump threading
2994 path. Free the path vector.
2995 (thread_block): Get edges from the jump threading path. Look at the
2996 entire path to see if we thread to a loop exit. If we cancel a jump
2997 thread request, then free the path vector.
2998 (thread_single_edge): Get edges and block types from the jump threading
2999 path. Free the path vector.
3000 (thread_through_loop_header): Get edges and block types from the jump
3001 threading path. Free the path vector.
3002 (mark_threaded_blocks): Iterate over the vector of paths and store
3003 the path on the appropriate edge. Get edges and block types from the
3004 jump threading path.
3005 (mark_threaded_blocks): Get edges and block types from the jump
3006 threading path. Free the path vector.
3007 (thread_through_all_blocks): Use the vector of paths rather than
3008 a vector of 3-edge sets.
3009 (register_jump_thread): Accept pointer to a path vector rather
3010 than the path vector itself. Store the path vector for later use.
3013 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3014 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3017 * config/s390/s390.c (s390_split_branches): Modify check for table
3019 (s390_chunkify_start): Rearrange table jump insn check in order to
3020 deal with compare and branch insns correctly.
3022 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3026 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3027 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3028 define_insn_and_split.
3029 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3030 (shiftsi3_compare): New pattern.
3032 * config/arm/unspecs.md (UNSPEC_RRX): New.
3034 2013-10-01 Alan Modra <amodra@gmail.com>
3036 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3037 casing inout operands.
3039 2013-10-01 Richard Biener <rguenther@suse.de>
3041 PR tree-optimization/58553
3042 * tree-loop-distribution.c (struct partition_s): Add niter member.
3043 (classify_partition): Populate niter member for the partition
3044 and properly identify whether the relevant store happens before
3045 or after the loop exit.
3046 (generate_memset_builtin): Use niter member from the partition.
3047 (generate_memcpy_builtin): Likewise.
3049 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3051 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3052 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3054 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3056 * basic-block.h (edge_list): Prefix member names with "m_".
3057 * context.h (context): Likewise.
3058 * domwalk.h (dom_walker): Likewise.
3059 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3060 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3061 * hash-table.h (hash_table): Likewise.
3062 * machmode.h (bit_field_mode_iterator): Likewise.
3063 * pass_manager.h (pass_list): Likewise.
3064 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3065 * tree-pass.h (pass_data): Likewise.
3066 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3067 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3068 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3069 * asan.c (pass_data_asan): Update accordingly.
3070 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3071 (control_dependences::control_dependences): Likewise.
3072 (control_dependences::~control_dependences): Likewise.
3073 (control_dependences::~control_dependences): Likewise.
3074 (control_dependences::get_edges_dependent_on): Likewise.
3075 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3076 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3077 * context.c (gcc::context::context): Likewise.
3078 * cprop.c (pass_rtl_cprop::clone): Likewise.
3079 * domwalk.c (dom_walker::walk): Likewise.
3080 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3081 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3082 * mode-switching.c (pass_mode_switching::clone): Likewise.
3083 * passes.c (opt_pass::opt_pass): Likewise.
3084 (pass_manager::pass_manager): Likewise.
3085 * predict.c (pass_strip_predict_hints::clone): Likewise.
3086 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3087 (pass_split_all_insns::clone): Likewise.
3088 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3090 (bit_field_mode_iterator::next_mode): Likewise.
3091 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3092 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3093 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3094 * tree-complex.c (pass_lower_complex::clone): Likewise.
3095 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3096 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3097 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3098 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3099 (pass_fold_builtins::clone): Likewise.
3100 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3101 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3102 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3103 (pass_cd_dce::clone): Likewise.
3104 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3105 (pass_phi_only_cprop::clone): Likewise.
3106 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3107 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3108 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3109 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3110 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3111 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3112 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3113 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3114 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3115 * tree-vrp.c (pass_vrp::clone): Likewise.
3116 * tsan.c (pass_tsan::clone): Likewise.
3118 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3121 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3122 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3125 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3126 optimization, punt if sign_bit_p looked through any zero extension.
3128 2013-09-30 Teresa Johnson <tejohnson@google.com>
3130 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3131 Update redirected out edge count in joiner case.
3132 (ssa_redirect_edges): Common the joiner and non-joiner cases
3133 so that joiner case gets profile updates.
3135 2013-09-30 Richard Biener <rguenther@suse.de>
3137 PR tree-optimization/58554
3138 * tree-loop-distribution.c (classify_partition): Require
3139 unconditionally executed stores for memcpy and memset recognition.
3140 (tree_loop_distribution): Calculate dominance info.
3142 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3144 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3145 (NO_PROFILE_COUNTERS): Likewise.
3146 (PROFILE_HOOK): Likewise.
3147 (FUNCTION_PROFILER): Likewise.
3148 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3150 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3152 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3153 calls and defines in TARGET_MACHO conditional.
3154 (load_macho_picbase_di): Likewise.
3155 (reload_macho_picbase): Likewise.
3156 (reload_macho_picbase_si): Likewise.
3157 (reload_macho_picbase_di): Likewise.
3158 (nonlocal_goto_receiver): Likewise.
3160 2013-09-30 Nick Clifton <nickc@redhat.com>
3162 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3163 that use the MSP430X ISA.
3164 (msp430_option_override): Scan -mmcu command line option for any
3165 MCU name that supports the MSP430X ISA.
3166 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3167 -mmcu options which enable the MSP430X ISA.
3169 2013-09-30 Richard Biener <rguenther@suse.de>
3172 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3173 before looking for setjmp-like calls.
3175 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3178 * config/darwin-protos.h (machopic_get_function_picbase): New.
3179 * config/darwin.c (machopic_get_function_picbase): New.
3180 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3181 label for a new func. (load_macho_picbase_di): Likewise.
3182 (reload_macho_picbase): New expand.
3183 (reload_macho_picbase_si): New insn.
3184 (reload_macho_picbase_di): New insn.
3185 (nonlocal_goto_receiver): New define and split.
3186 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3187 (unspecv enum): Add UNSPECV_NLGR.
3189 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3191 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3192 that altivec registers are correctly sized on Darwin.
3194 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3196 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3197 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3198 * config/x-darwin (host-darwin.o): Likewise.
3199 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3200 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3201 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3203 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3205 * doc/invoke.texi: Fix usage of @tie{} command.
3207 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3209 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3211 2013-09-28 Jan Hubicka <jh@suse.cz>
3213 * config/i386/x86-tune.def: Add documentation for each of the options;
3216 2013-09-28 Jan Hubicka <jh@suse.cz>
3218 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3220 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3221 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3222 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3223 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3224 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3226 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3228 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3229 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3230 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3231 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3232 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3233 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3234 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3235 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3236 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3237 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3238 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3239 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3240 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3241 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3242 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3243 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3244 gimple.h, godump.c, graphite-clast-to-gimple.c,
3245 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3246 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3247 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3248 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3249 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3250 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3251 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3252 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3253 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3254 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3255 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3256 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3257 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3258 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3259 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3260 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3261 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3262 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3263 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3264 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3265 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3266 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3267 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3268 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3269 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3270 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3271 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3272 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3273 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3274 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3275 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3276 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3277 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3278 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3279 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3280 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3281 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3282 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3283 whitespace before "(".
3285 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3287 * expr.h (extract_bit_field): Remove packedp parameter.
3288 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3289 from forward declaration.
3290 (store_split_bit_field): Remove packedp arg from calls to
3291 extract_fixed_bit_field.
3292 (extract_bit_field_1): Remove packedp parameter and packedp
3293 argument from recursive calls and calls to extract_fixed_bit_field.
3294 (extract_bit_field): Remove packedp parameter and corresponding
3295 arg to extract_bit_field_1.
3296 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3298 (extract_split_bit_field): Remove packedp arg from call to
3299 extract_fixed_bit_field.
3300 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3301 (copy_blkmode_from_reg): Likewise.
3302 (copy_blkmode_to_reg): Likewise.
3303 (read_complex_part): Likewise.
3304 (store_field): Likewise.
3305 (expand_expr_real_1): Likewise.
3306 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3307 to extract_bit_field.
3308 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3309 call to extract_bit_field.
3310 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3311 call to extract_bit_field.
3312 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3313 and special packedp behavior from -fstrict-volatile-bitfields
3316 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3318 * lra-eliminations.c (init_elim_table): Guard value_p.
3320 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3322 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3323 DFmode, DImode, and SFmode in the upper VSX registers based on the
3324 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3325 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3326 of the VSX register handling.
3328 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3330 (zero_extendsidi2_lfiwzx): Add support for loading into the
3331 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3332 only do VSX memory options on Altivec registers.
3333 (extendsidi2_lfiwax): Likewise.
3334 (extendsfdf2_fpr): Likewise.
3335 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3336 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3337 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3338 (movdi_internal64): Likewise.
3340 2013-09-27 Xinliang David Li <davidxl@google.com>
3342 * opts.c (finish_options): Adjust parameters
3343 according to vect cost model.
3344 (common_handle_option): Set dynamic vect cost
3346 targhooks.c (default_add_stmt_cost): Compute stmt cost
3348 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3349 Use helper function.
3350 * tree-vectorizer.h (unlimited_cost_model): New function.
3351 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3352 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3354 (vect_enhance_data_refs_alignment): Ditto.
3355 * flag-types.h: New enum.
3356 * common/config/i386/i386-common.c (ix86_option_init_struct):
3357 No need to initialize vect_cost_model flag.
3358 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3361 2013-09-27 Diego Novillo <dnovillo@google.com>
3363 * gimple.h (enum ssa_mode): Remove.
3365 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3367 * cfgloop.h (number_of_loops): Fix typo in check for null.
3369 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3372 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3373 are children of outermost saved_cfun's loop, and set it up to
3374 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3375 if loop != loop0 and SESE region contains bbs that belong to loop0.
3377 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3379 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3380 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3381 (set_address_segment, set_address_base, set_address_index)
3382 (set_address_disp): Accept the argument unconditionally.
3383 (baseness): Remove must_be_base_p and must_be_index_p checks.
3384 (decompose_normal_address): Classify as much as possible in the
3387 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3389 * cse.c (count_reg_usage): Handle INT_LIST.
3390 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3391 * reginfo.c (reg_scan_mark_refs): Likewise.
3392 * reload1.c (eliminate_regs_1): Likewise.
3394 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3397 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3400 2013-09-27 Richard Biener <rguenther@suse.de>
3402 PR tree-optimization/58459
3403 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3404 restriction not propagating into loops.
3406 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3408 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3409 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3410 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3412 2013-09-26 Richard Biener <rguenther@suse.de>
3414 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3416 2013-09-26 Richard Biener <rguenther@suse.de>
3418 * alias.h (component_uses_parent_alias_set): Rename to ...
3419 (component_uses_parent_alias_set_from): ... this.
3420 * alias.c (component_uses_parent_alias_set): Rename to ...
3421 (component_uses_parent_alias_set_from): ... this and return
3423 (reference_alias_ptr_type_1): Use the result from
3424 component_uses_parent_alias_set_from instead of stripping
3425 components one at a time.
3426 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3428 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3430 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3431 Move prototypes to...
3432 * tree-ssa-ter.h: New File. Move prototypes here.
3433 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3434 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3435 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3436 from is_replaceable_p.
3437 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3438 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3439 (process_replaceable): Use ter_is_replaceable_p.
3440 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3441 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3442 newly refactored ssa_is_replaceable_p.
3443 * cfgexpand.c: Include tree-outof-ssa.h.
3444 * ssaexpand.h: Delete.
3446 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3448 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3449 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3451 (create_gimple_tmp): Delete.
3452 (get_expr_type, build_assign, build_type_cast): Move to...
3453 * gimple-builder.c: New File.
3454 (get_expr_type): Relocate from gimple.c.
3455 (build_assign, build_type_cast): Change to only create ssanames.
3456 * gimple.h: Move prototypes to...
3457 * gimple-builder.h: New File. Here.
3458 * tree-ssa.h: And here.
3459 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3460 count_uses_and_derefs): Relocate from gimple.c.
3461 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3462 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3463 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3464 gimple_replace_ssa_lhs.
3465 * asan.c: Include gimple-builder.h.
3466 * Makefile.in: Add gimple-builder.o.
3468 2013-09-26 Richard Biener <rguenther@suse.de>
3470 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3471 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3472 (loe_visit_block): Use gcc_checking_assert.
3473 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3474 gimple_assign_ssa_name_copy_p.
3475 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3478 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3480 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3481 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3483 2013-09-26 Richard Biener <rguenther@suse.de>
3485 PR tree-optimization/58539
3486 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3487 the fact that debug statements are not taking part in loop-closed
3490 2013-09-26 Nick Clifton <nickc@redhat.com>
3492 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3493 time warning message.
3494 (msp430_print_operand_raw): Delete unused letter parameter.
3495 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3496 (msp430_print_operand_address): New function.
3497 (msp430_print_operand): Move address printing code from here to
3499 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3501 (zero_extendpsisi2): Likewise.
3502 (extendpsisi2): New pattern.
3503 (andneghi3): New pattern.
3505 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3507 * config/aarch64/aarch64.opt (mlra): New option.
3508 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3509 (TARGET_LRA_P): Define.
3511 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3513 * expr.c (expand_assignment): Remove obsolete comment.
3515 2013-09-25 Jeff Law <law@redhat.com>
3517 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3518 tree-ssa-threadupdate.h.
3519 (register_jump_thread): Similarly.
3520 * tree-ssa-threadupdate.h: New header file.
3521 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3522 * tree-vrp.c: Likewise.
3523 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3524 (thread_around_empty_blocks): Change type of path vector argument to
3525 an edge,type pair from just an edge. Initialize both elements when
3526 appending to a jump threading path. Tweak references to elements
3528 (thread_across_edge): Similarly. Release memory for the elements
3530 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3531 (dump_jump_thread_path): New function broken out from
3532 register_jump_thread.
3533 (register_jump_thread): Use dump_jump_thread_path. Change type of
3534 path vector entries. Search the path for NULL edges and dump
3535 the path if one is found. Tweak the conversion of path to 3-edge
3536 form to use the block copy type information embedded in the path.
3538 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3540 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3542 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3543 Vladimir Makarov <vmakarov@redhat.com>
3545 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3546 from the least significant bit.
3547 (strip_address_mutations): Add bitfield operations handling.
3548 (must_be_index_p): Add shifting and rotate operations handling.
3549 (set_address_base): Use must_be_base_p predicate.
3550 (set_address_index): Use must_be_index_p predicate.
3552 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3553 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3554 Sergey Lega <sergey.s.lega@intel.com>
3555 Anna Tikhonova <anna.tikhonova@intel.com>
3556 Ilya Tocar <ilya.tocar@intel.com>
3557 Andrey Turetskiy <andrey.turetskiy@intel.com>
3558 Ilya Verbin <ilya.verbin@intel.com>
3559 Kirill Yukhin <kirill.yukhin@intel.com>
3560 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3562 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3564 (ix86_expand_vector_move_misalign): Support new unaligned load and
3565 stores and use new names.
3566 (CODE_FOR_sse2_storedqu): Rename to ...
3567 (CODE_FOR_sse2_storedquv16qi): ... this.
3568 (CODE_FOR_sse2_loaddqu): Rename to ...
3569 (CODE_FOR_sse2_loaddquv16qi): ... this.
3570 (CODE_FOR_avx_loaddqu256): Rename to ...
3571 (CODE_FOR_avx_loaddquv32qi): ... this.
3572 (CODE_FOR_avx_storedqu256): Rename to ...
3573 (CODE_FOR_avx_storedquv32qi): ... this.
3574 * config/i386/i386.md (fpint_logic): New.
3575 * config/i386/sse.md (VMOVE): Extend for AVX512.
3579 (VI_UNALIGNED_LOADSTORE): Ditto.
3580 (sse2_avx_avx512f): Ditto.
3581 (sse2_avx2): Extend for AVX512.
3582 (sse4_1_avx2): Ditto.
3583 (avx2_avx512f): New.
3584 (sse): Extend for AVX512.
3587 (avxsizesuffix): Ditto.
3588 (sseintvecmode): Ditto.
3590 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3591 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3592 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3593 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3594 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3595 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3596 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3597 (STORENT_MODE): Extend for AVX512.
3598 (*absneg<mode>2): Replace constraint "x" with "v".
3599 (*mul<mode>3): Ditto.
3600 (*ieee_smin<mode>3): Ditto.
3601 (*ieee_smax<mode>3): Ditto.
3602 (avx_cmp<mode>3): Replace VF with VF_128_256.
3603 (*<sse>_maskcmp<mode>3_comm): Ditto.
3604 (<sse>_maskcmp<mode>3): Ditto.
3605 (<sse>_andnot<mode>3): Extend for AVX512.
3606 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3607 (<code><mode>3, fpint_logic): New.
3608 (*<code><mode>3): Extend for AVX512.
3609 (avx512flogicsuff): New.
3610 (avx512f_<logic><mode>): Ditto.
3611 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3613 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3614 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3615 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3616 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3617 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3618 (xop_vpermil2<mode>3): Ditto.
3619 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3620 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3621 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3622 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3624 2013-09-25 Tom Tromey <tromey@redhat.com>
3626 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3627 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3628 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3629 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3630 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3631 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3632 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3633 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3634 (GRAPHITE_HTAB_H): Remove.
3636 2013-09-25 Tom Tromey <tromey@redhat.com>
3638 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3640 2013-09-25 Tom Tromey <tromey@redhat.com>
3642 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3644 2013-09-25 Tom Tromey <tromey@redhat.com>
3646 * config/i386/t-i386 (i386.o): Remove.
3647 (i386-c.o): Use COMPILE and POSTCOMPILE.
3649 2013-09-25 Tom Tromey <tromey@redhat.com>
3651 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3653 2013-09-25 Tom Tromey <tromey@redhat.com>
3655 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3656 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3657 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3658 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3659 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3660 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3661 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3662 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3663 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3664 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3665 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3666 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3667 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3668 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3669 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3670 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3671 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3672 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3673 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3674 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3675 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3676 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3677 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3678 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3679 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3680 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3681 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3682 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3683 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3684 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3685 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3686 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3687 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3688 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3689 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3690 (graphite-clast-to-gimple.o, graphite-dependences.o)
3691 (graphite-interchange.o, graphite-poly.o)
3692 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3693 (graphite-optimize-isl.o, tree-vect-loop.o)
3694 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3695 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3696 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3697 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3698 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3699 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3700 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3701 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3702 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3703 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3704 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3705 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3706 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3707 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3708 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3709 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3710 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3711 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3712 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3713 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3714 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3715 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3716 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3717 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3718 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3719 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3720 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3721 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3722 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3723 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3724 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3725 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3726 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3727 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3728 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3729 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3730 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3731 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3732 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3733 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3734 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3735 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3736 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3737 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3738 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3739 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3740 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3741 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3742 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3743 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3744 (gcov-dump.o): Remove.
3745 (default-c.o): Use COMPILE and POSTCOMPILE.
3746 (CFLAGS-gcc.o): New variable.
3747 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3749 2013-09-25 Tom Tromey <tromey@redhat.com>
3751 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3752 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3753 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3754 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3755 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3756 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3757 (c-family/array-notation-common.o, c-family/stub-objc.o)
3758 (c-family/c-ubsan.o): Remove.
3760 2013-09-25 Tom Tromey <tromey@redhat.com>
3762 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3764 2013-09-25 Tom Tromey <tromey@redhat.com>
3766 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3767 to add -DENABLE_SHARED_LIBGCC.
3768 (gcc.o): Don't use subshell.
3770 2013-09-25 Tom Tromey <tromey@redhat.com>
3772 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3773 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3774 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3775 * configure, config.in: Rebuild.
3777 2013-09-25 Tom Tromey <tromey@redhat.com>
3779 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3780 (COMPILE, POSTCOMPILE): New variables.
3781 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3782 (DEPFILES): New variable.
3783 Include ".Po" files.
3784 * configure.ac: Add checks for dependency checking.
3785 * configure, aclocal.m4: Regenerate.
3787 2013-09-25 Tom Tromey <tromey@redhat.com>
3789 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3790 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3792 2013-09-25 Tom Tromey <tromey@redhat.com>
3794 * Makefile.in (generated_files): Add options.h,
3795 target-hooks-def.h, insn-opinit.h,
3796 common/common-target-hooks-def.h, pass-instances.def,
3797 c-family/c-target-hooks-def.h.
3799 2013-09-25 Jeff Law <law@redhat.com>
3801 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3802 than foo[foo.length () - 1] to access last member in a vec.
3803 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3805 2013-09-25 Richard Biener <rguenther@suse.de>
3808 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3810 2013-09-25 Jan Hubicka <jh@suse.cz>
3812 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3815 2013-09-25 Marek Polacek <polacek@redhat.com>
3818 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3819 when determining the type name.
3821 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3823 * config/sh/sh.md: Fix formatting.
3825 2013-09-24 Xinliang David Li <davidxl@google.com>
3827 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3828 max peel iterations parameter.
3829 * param.def: New parameter.
3830 * doc/invoke.texi: Document New parameter.
3832 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3834 * gimple-pretty-print.c: Various whitespace tweaks.
3835 * tree-core.h: Likewise.
3836 * tree-pretty-print.c: Likewise.
3837 * tree-ssa-alias.c: Likewise.
3838 * tree-ssa-copy.c: Likewise.
3839 * tree-ssanames.c: Likewise.
3840 * tree-ssanames.h: Likewise.
3841 * tree-vrp.c: Likewise.
3843 2013-09-24 Alan Modra <amodra@gmail.com>
3847 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3848 for output operands that disallow regs. Don't use EXPAND_WRITE on
3851 2013-09-24 Richard Biener <rguenther@suse.de>
3854 * tree.c (reference_alias_ptr_type): Move ...
3855 * alias.c (reference_alias_ptr_type): ... here and implement
3856 in terms of the new reference_alias_ptr_type_1.
3857 (ref_all_alias_ptr_type_p): New helper.
3858 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3859 use ref_all_alias_ptr_type_p.
3860 (get_deref_alias_set): Add flag_strict_aliasing check here.
3861 (reference_alias_ptr_type_1): New function, split out from ...
3862 (get_alias_set): ... here.
3863 (alias_ptr_types_compatible_p): New function.
3864 * alias.h (reference_alias_ptr_type): Declare.
3865 (alias_ptr_types_compatible_p): Likewise.
3866 * tree.h (reference_alias_ptr_type): Remove.
3867 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3868 to compare MEM_REF alias types.
3870 2013-09-24 Richard Biener <rguenther@suse.de>
3872 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3874 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3876 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3877 reload helper function arrays into a single array reg_addr.
3878 (reload_fpr_gpr): Likewise.
3879 (reload_gpr_vsx): Likewise.
3880 (reload_vsx_gpr): Likewise.
3881 (struct rs6000_reg_addr): Likewise.
3882 (reg_addr): Likewise.
3883 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3884 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3885 (rs6000_init_hard_regno_mode_ok): Likewise.
3886 (rs6000_secondary_reload_direct_move): Likewise.
3887 (rs6000_secondary_reload): Likewise.
3889 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3890 constraints: wu, ww, and wy. Repurpose wv constraint added during
3891 power8 changes. Put wg constraint in alphabetical order.
3893 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3894 for future work to add ISA 2.07 VSX single precision support.
3895 (-mvsx-scalar-double): Change default from -1 to 1, update
3896 documentation comment.
3897 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3898 (-mupper-regs-df): New debug switch to control whether DF values
3899 can go in the traditional Altivec registers.
3900 (-mupper-regs-sf): New debug switch to control whether SF values
3901 can go in the traditional Altivec registers.
3903 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3905 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3906 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3907 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3908 (rs6000_debug_legitimate_address_p): Print if we are running
3909 before, during, or after reload.
3910 (rs6000_secondary_reload): Add a comment.
3911 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3913 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3914 constraints. Update documentation string.
3915 (wd constraint): Likewise.
3916 (wf constraint): Likewise.
3917 (wg constraint): Likewise.
3918 (wn constraint): Likewise.
3919 (ws constraint): Likewise.
3920 (wt constraint): Likewise.
3921 (wx constraint): Likewise.
3922 (wz constraint): Likewise.
3923 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3925 (ww constraint): Likewise.
3926 (wy constraint): Likewise.
3927 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3928 the previous submissions.
3929 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3931 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3933 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3934 (REG_BR_PROB): Say that the probability is stored in an int_list.
3935 * reg-notes.def: Update commentary to mention INT_LIST.
3936 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3937 (INT_LIST): New rtx.
3938 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3939 * rtlanal.c (int_reg_note_p): New function.
3940 (alloc_reg_note): Assert that the note does not have an int argument.
3941 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3942 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3943 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3944 rather than an INSN_LIST. Handle INT_LIST.
3945 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3946 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3947 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3948 Manipulate them as ints rather than rtxes.
3949 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3950 * regmove.c (copy_src_to_dest): Likewise.
3951 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3953 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3954 into the cases that need it.
3955 * config/arm/arm.c (arm_unwind_emit): Likewise.
3957 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3958 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3959 * loop-doloop.c (add_test, doloop_modify): Likewise.
3960 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3961 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3962 * predict.c (combine_predictions_for_insn): Likewise.
3963 * print-rtl.c (print_rtx): Handle INT_LIST.
3964 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3965 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3966 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3967 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3968 (ix86_split_fp_branch, predict_jump): Likewise.
3969 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3970 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3971 * config/spu/spu.c (ea_load_store_inline): Likewise.
3973 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3974 value of a REG_BR_PROB note.
3975 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3976 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3977 * emit-rtl.c (try_split): Likewise.
3978 * predict.c (br_prob_note_reliable_p): Likewise.
3979 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
3980 * reorg.c (mostly_true_jump): Likewise.
3981 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
3982 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
3983 * config/i386/i386.c (ix86_print_operand): Likewise.
3984 * config/ia64/ia64.c (ia64_print_operand): Likewise.
3985 * config/mmix/mmix.c (mmix_print_operand): Likewise.
3986 * config/rs6000/rs6000.c (output_cbranch): Likewise.
3987 * config/s390/s390.c (s390_expand_tbegin): Likewise.
3988 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
3989 * config/sparc/sparc.c (output_cbranch): Likewise.
3990 * config/spu/spu.c (get_branch_target): Likewise.
3991 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
3992 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
3994 2013-09-23 Jan Hubicka <jh@suse.cz>
3996 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
3998 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
3999 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4001 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4002 (gimple_fold_call): Dump inconsistent devirtualizations; add
4003 sanity check for type based devirtualizations.
4004 * ipa-prop.c: Include ipa-utils.h
4005 (ipa_intraprocedural_devirtualization): Add sanity check.
4006 (try_make_edge_direct_virtual_call): Likewise.
4008 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4010 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4011 assignment statements.
4013 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4015 * gimple-pretty-print.c (dump_ssaname_info): New function.
4016 (dump_gimple_phi): Call it.
4017 (pp_gimple_stmt_1): Likewise.
4018 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4019 (range_info_def): Declare.
4020 * tree-pretty-print.c (pp_double_int): New function.
4021 (dump_generic_node): Call it.
4022 * tree-pretty-print.h (pp_double_int): Declare.
4023 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4024 * tree-ssanames.h (range_info_def): New structure.
4025 (value_range_type): Move definition here.
4026 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4028 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4030 (set_range_info): New function.
4031 (get_range_info): Likewise.
4032 (duplicate_ssa_name_range_info): Likewise.
4033 (duplicate_ssa_name_fn): Check pointer type and call
4034 duplicate_ssa_name_range_info.
4035 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4036 * tree-vrp.c (value_range_type): Remove definition, now in
4038 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4039 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4040 (SSA_NAME_RANGE_INFO): New macro.
4042 2013-09-23 Richard Biener <rguenther@suse.de>
4044 PR tree-optimization/58464
4045 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4046 (phi_trans_add): Change to add conditionally on being not present.
4047 (phi_translate_1): Remove recursion detection here.
4048 (phi_translate): Pre-seed the cache with NULL to catch
4049 recursion here in a more generic way.
4050 (bitmap_find_leader): Adjust comment.
4051 (get_representative_for): Dump value-numbers.
4052 (create_expression_by_pieces): Likewise.
4053 (insert_into_preds_of_block): Likewise.
4055 2013-09-23 Christian Bruel <christian.bruel@st.com>
4058 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4059 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4061 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4064 2013-09-20 Renlin Li <renlin.li@arm.com>
4066 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4068 (aarch64_expand_epilogue): Likewise.
4069 (aarch64_legitimize_reload_address): Likewise.
4071 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4073 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4074 NULL_TREE before pushing them onto the vector. Likewise for labels.
4076 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4078 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4079 comparison operators when -fno-trapping-math is in effect.
4080 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4081 comparison operators in TFmode and assert that unsupported operators
4083 (ia64_print_operand): Likewise.
4085 2013-09-21 Jan Hubicka <jh@suse.cz>
4087 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4088 (sse_typeless_stores): Enable for core
4089 (sse_load0_by_pxor): Likewise.
4090 (four_jump_limit): Disable for core.
4091 (pad_returns): Likewise.
4092 (avoid_vector_decode): Likewise.
4093 (fuse_cmp_and_branch): Enable for cores.
4094 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4096 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4099 * config/pa/pa.c (pa_option_override): Disable auto increment and
4100 decrement instructions until reload is completed.
4102 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4103 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4104 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4106 2013-09-20 DJ Delorie <dj@redhat.com>
4107 Nick Clifton <nickc@redhat.com>
4109 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4110 (need_to_save): Save bank 0 on interrupts.
4111 (characterize_address): Strip far address wrappers.
4112 (rl78_as_legitimate_address): Likewise.
4113 (transcode_memory_rtx): Likewise.
4114 (rl78_peep_movhi_p): Disable this peephole after devirt.
4115 (rl78_propogate_register_origins): Forget all origins when a
4117 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4118 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4120 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4121 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4122 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4124 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4126 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4128 * config/pa/pa.md: In "scc" insn patterns, change output template to
4129 handle const0_rtx in reg_or_0_operand operands.
4131 2013-09-20 Martin Husemann <martin@NetBSD.org>
4134 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4135 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4137 2013-09-20 Richard Biener <rguenther@suse.de>
4140 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4141 remembering SSA name version and block index.
4142 (new_scev_info_str): Adjust.
4143 (hash_scev_info): Likewise. Also hash the block index.
4144 (eq_scev_info): Adjust.
4145 (find_var_scev_info): Likewise.
4146 (struct instantiate_cache_entry): Remove.
4147 (struct instantiate_cache_type): Use a htab to map name, block
4149 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4150 (get_instantiated_value_entry): Likewise.
4151 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4152 (instantiate_scev_name): Adjust.
4154 2013-09-20 Jeff Law <law@redhat.com>
4156 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4158 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4160 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4161 Call aarch64_simd_expand_args to update op[argc].
4163 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4165 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4168 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4170 * gengtype.c (file_rules): Added rule for *.cc files.
4171 (get_output_file_with_visibility): Give fatal message when no
4174 2013-09-20 Renlin Li <renlin.li@arm.com>
4176 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4177 (aarch64_expand_epilogue): Likewise.
4178 (aarch64_legitimize_reload_address): Likewise.
4180 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4183 * expr.c (expand_assignment): Remove misalignp code path.
4185 2013-09-20 Marek Polacek <polacek@redhat.com>
4188 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4189 TYPE_PRECISION. Add asserts.
4191 2013-09-20 Richard Biener <rguenther@suse.de>
4193 PR tree-optimization/58453
4194 * tree-loop-distribution.c (distribute_loop): Apply the cost
4195 model for -ftree-loop-distribute-patterns, too.
4197 2013-09-20 Richard Biener <rguenther@suse.de>
4200 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4201 make type comparison less strict.
4203 2013-09-20 Alan Modra <amodra@gmail.com>
4205 * configure: Regenerate.
4206 * aclocal.m4: Regenerate.
4208 2013-09-20 Marek Polacek <polacek@redhat.com>
4211 * doc/extend.texi: Document that attribute used is meant to be used
4212 on variables with static storage duration.
4214 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4216 PR tree-optimization/58472
4217 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4218 simd_lane_access set inv_p = false.
4219 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4220 the simduid magic VAR_DECL.
4222 2013-09-19 Jan Hubicka <jh@suse.cz>
4224 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4226 2013-09-17 Jeff Law <law@redhat.com>
4228 * tree-ssa-dom.c (record_temporary_equivalences): New function
4229 split out of dom_opt_dom_walker::after_dom_children.
4230 (dom_opt_dom_walker::thread_across_edge): Move common code
4231 in here from dom_opt_dom_walker::after_dom_children.
4232 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4234 2013-09-19 Jan Hubicka <jh@suse.cz>
4236 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4237 (TARGET_GENERIC): Use PROCESOR_GENERIC
4238 (enum processor_type): Unify generic32 and 64.
4239 * i386.md (cpu): Likewise.
4240 * x86-tune.def (use_leave): Enable for generic32.
4241 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4242 * athlon.md: Change generic64 to generic in all occurences.
4243 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4244 (ix86_target_macros_internal): Likewise.
4245 * driver-i386.c (host_detect_local_cpu): Likewise.
4246 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4248 (generic_memcpy, generic_memset, generic_cost): This one.
4249 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4250 (m_GENERIC32, m_GENERIC64): Remove.
4251 (m_GENERIC): Turn into one flag.
4252 (processor_target): Unify generic tunnings.
4253 (ix86_option_override_internal): Replace generic32/64 by generic.
4254 (ix86_issue_rate): Likewise.
4255 (ix86_adjust_cost): Likewise.
4257 2013-09-19 Jan Hubicka <jh@suse.cz>
4259 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4260 of speculative flag.
4262 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4264 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4265 GOMP_sections_start, even if !exit_reachable.
4267 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4269 * lra-constraints.c (need_for_all_save_p): Use macro
4270 HARD_REGNO_CALL_PART_CLOBBERED.
4271 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4272 set up pseudo conflict hard regs.
4274 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4277 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4280 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4282 PR rtl-optimization/58438
4283 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4284 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4285 from previous subpasses.
4287 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4289 * arm.c (arm_get_frame_offsets): Validate architecture supports
4290 LDRD/STRD before accepting the tuning preference.
4291 (arm_expand_prologue): Likewise.
4292 (arm_expand_epilogue): Likewise.
4294 2013-09-18 Richard Biener <rguenther@suse.de>
4296 PR tree-optimization/58417
4297 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4298 have chrecs with symbols defined in the loop as operands.
4299 (chrec_fold_multiply): Likewise.
4300 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4301 parameters before folding binary operations.
4302 (struct instantiate_cache_entry_hasher): Remove.
4303 (struct instantiate_cache_type): Use a pointer-map.
4304 (instantiate_cache_type::instantiate_cache_type): New function.
4305 (instantiate_cache_type::get): Likewise.
4306 (instantiate_cache_type::set): Likewise.
4307 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4308 (get_instantiated_value_entry): Likewise.
4309 (global_cache): New global.
4310 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4311 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4312 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4313 (instantiate_scev_name): Adjust.
4314 (instantiate_scev): Construct global instead of local cache.
4315 (resolve_mixers): Likewise.
4317 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4318 Paolo Carlini <paolo.carlini@oracle.com>
4321 * doc/implement-cxx.texi: Fix references to the C++ standards.
4323 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4325 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4326 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4327 simd_lane_access drs, update also DR_ALIGNED_TO.
4329 2013-09-18 Marek Polacek <polacek@redhat.com>
4332 * doc/extend.texi: Document no_sanitize_undefined attribute.
4333 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4334 no_sanitize_undefined attribute.
4336 2013-09-18 Nick Clifton <nickc@redhat.com>
4338 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4339 (ASM_DECLARE_FUNCTION_NAME): Define.
4341 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4343 * compare-elim.c (find_comparison_dom_walker): New class
4344 (find_comparisons_in_bb): Rename to
4345 find_comparison_dom_walker::before_dom_children
4346 (find_comparisons): Adjust
4347 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4349 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4350 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4351 (init_walk_dominator_tree): Remove declaration.
4352 (fini_walk_dominator_tree): Remove declaration.
4353 * fwprop.c (single_def_use_dom_walker): New class
4354 (single_def_use_enter_block): Convert to
4355 single_def_use_dom_walker::before_dom_children.
4356 (single_def_use_leave_block): Convert to
4357 single_def_use_dom_walker::after_dom_children.
4358 (build_single_def_use_links): Adjust.
4359 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4361 (find_candidates_in_block): Convert to
4362 find_candidates_dom_walker::before_dom_children.
4363 (execute_strength_reduction): Adjust.
4364 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4365 (sese_dom_walker): New class.
4366 (sese_dom_walker::sese_dom_walker): New constructor.
4367 (sese_dom_walker::~sese_dom_walker): New destructor.
4368 (build_sese_conditions_before): Convert to
4369 sese_dom_walker::before_dom_children.
4370 (build_sese_conditions_after): Convert to
4371 sese_dom_walker::after_dom_children.
4372 (build_poly_scop): Adjust
4373 * tree-into-ssa.c (rewrite_dom_walker): New class
4374 (rewrite_enter_block): Convert to
4375 rewrite_dom_walker::before_dom_children.
4376 (rewrite_leave_block): Convert to
4377 rewrite_dom_walker::after_dom_children.
4378 (rewrite_update_dom_walker): New class.
4379 (rewrite_update_enter_block): Convert to
4380 rewrite_update_dom_walker::before_dom_children.
4381 (rewrite_update_leave_block): Convert to
4382 rewrite_update_dom_walker::after_dom_children.
4383 (rewrite_blocks, rewrite_into_ssa): Adjust.
4384 (mark_def_dom_walker): New class.
4385 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4386 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4387 (mark_def_sites_blocks): Convert to
4388 mark_def_dom_walker::before_dom_children.
4389 (mark_def_site_blocks): Remove.
4390 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4391 (tree_ssa_dominator_optimize): Adjust.
4392 (dom_thread_across_edge): Convert to method
4393 dom_opt_dom_walker::thread_across_edge.
4394 (dom_opt_enter_block): Convert to member function
4395 dom_opt_dom_walker::before_dom_children.
4396 (dom_opt_leave_block): Convert to member function
4397 dom_opt_dom_walker::after_dom_children.
4398 * tree-ssa-dse.c (dse_dom_walker): New class.
4399 (dse_enter_block): Convert to member function
4400 dse_dom_walker::before_dom_children.
4401 (tree_ssa_dse): Adjust.
4402 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4403 (determine_invariantness_stmt): Convert to method
4404 invariantness_dom_walker::before_dom_children.
4405 (determine_invariantness): Remove
4406 (move_computations_dom_walker): New class.
4407 (move_computations_stmt): Convert to method
4408 move_computations_dom_walker::before_dom_children.
4409 (move_computations, tree_ssa_lim): Adjust.
4410 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4411 (nt_init_block): Make method
4412 notrappping_dom_walker::before_dom_children.
4413 (nt_fini_block): Make
4414 method nontrapping_dom_walker::after_dom_children.
4415 (get_non_trapping): Adjust.
4416 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4417 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4418 (eliminate_leave_block): Make method.
4419 eliminate_dom_walker::after_dom_children.
4421 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4422 (strlen_enter_block): Make method
4423 strlen_dom_walker::before_dom_children.
4424 (strlen_leave_block): Make
4425 method strlen_dom_walker::after_dom_children.
4426 (tree_ssa_strlen): Adjust.
4427 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4428 (tree_ssa_uncprop): Adjust.
4429 (uncprop_leave_block): Make method
4430 uncprop_dom_walker::after_dom_children.
4431 (uncprop_leave_block): Make method
4432 uncprop_dom_walker::before_dom_children.
4434 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4436 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4437 branch in same basic block. Check both src and dest of the move insn.
4439 2013-09-17 Nick Clifton <nickc@redhat.com>
4441 * config/rl78/rl78-real.md (bf): New pattern.
4443 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4444 (rl78_print_operand): Do not put a # before a %B.
4445 * config/rl78/rl78.opt: Tweak doc strings.
4447 2013-09-17 DJ Delorie <dj@redhat.com>
4449 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4450 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4451 (rl78_option_override): Likewise, if -mallregs.
4452 (is_virtual_register): Likewise.
4453 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4454 (REGNO_OK_FOR_BASE_P): Likewise.
4455 * config/rl78/rl78.opt (-mallregs): New.
4457 2013-09-17 Nick Clifton <nickc@redhat.com>
4459 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4460 For interrupt functions: save all call clobbered registers if the
4461 interrupt handler is not a leaf function.
4462 (rl78_expand_prologue): Always recompute the frame information.
4463 For interrupt functions: only select bank 0 if one of the bank 0
4464 registers is going to be psuhed.
4466 2013-09-17 DJ Delorie <dj@redhat.com>
4468 * config/rl78/constraints.md: For each W* constraint, rename to C*
4469 and create a W* constraint that checks for an optional ES: prefix
4471 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4472 (es_addr): New. Used to wrap far addresses.
4473 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4474 (rl78_es_base): New.
4475 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4476 wrapped far addresses.
4477 (rl78_print_operand_1): Unwrap far addresses before processing.
4478 (rl78_lo16): Wrap far addresses in unspecs.
4479 (rl78_es_addr): New.
4480 (rl78_es_base): New.
4481 (insn_ok_now): Check for not-yet-wrapped far addresses.
4482 (transcode_memory_rtx): Properly re-wrap far addresses.
4484 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4486 * config/sparc/t-rtems: Add leon3 multilibs.
4488 2013-09-17 Cong Hou <congh@google.com>
4490 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4491 when checking the dot production pattern. The type of rhs operand
4492 of multiply is now checked correctly.
4494 2013-09-17 Jeff Law <law@redhat.com>
4496 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4497 edge implied equivalences into successor phis.
4498 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4499 here from tree-ssa-threadedge.c.
4500 (mark_threaded_blocks): When threading through a joiner, if both
4501 successors of the joiner's clone reach the same block, verify the
4502 PHI arguments are equal. If not, cancel the jump threading request.
4503 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4504 tree-ssa-threadupdate.c
4505 (thread_across_edge): Don't check PHI argument equality when
4506 threading through joiner block here.
4508 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4510 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4511 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4512 (warn_uninit, warn_uninitialized_vars,
4513 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4514 Move to tree-ssa-uninit.c.
4515 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4516 (has_undefined_value_p): New. Pass dependant parts of
4517 ssa_undefined_value_p.
4518 (uninit_undefined_value_p): Use has_undefined_value_p.
4519 (warn_uninit, warn_uninitialized_vars,
4520 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4521 Move from tree-ssa.c.
4522 * tree-ssa.h: Adjust prototypes.
4524 2013-09-17 Jan Hubicka <jh@suse.cz>
4527 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4528 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4529 FUNCTION_NOT_OPTIMIZED.
4530 * ipa-inline-analysis.c (compute_inline_parameters): Function
4531 not optimized is not inlinable unless it is alwaysinline.
4532 (inline_analyze_function): Force calls in not optimized
4533 function not inlinable.
4535 2013-09-17 Jan Hubicka <jh@suse.cz>
4538 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4540 * ipa.c (function_and_variable_visibility): Likewise.
4541 * ipa-profile.c (ipa_profile): Likewise.
4543 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4546 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4547 attribute, and don't inline the resolver in this case.
4549 2013-09-17 Teresa Johnson <tejohnson@google.com>
4551 * coverage.c (get_coverage_counts): Add missing newline.
4553 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4555 PR tree-optimization/58088
4556 * fold-const.c (mask_with_trailing_zeros): New function.
4557 (fold_binary_loc): Make sure we don't recurse infinitely
4558 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4559 Use mask_with_trailing_zeros where appropriate.
4561 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4563 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4564 of 'prev' var to get better distance estimation.
4566 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4568 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4569 * tree-inline.c (is_parameter_of): New predicate.
4570 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4571 a parameter has been remapped.
4572 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4573 (optimize_inline_calls): Initialize transform_parameter.
4574 (copy_gimple_seq_and_replace_locals): Likewise.
4575 (tree_function_versioning): Likewise.
4576 (maybe_inline_call_in_expr): Likewise.
4578 2013-09-17 Nick Clifton <nickc@redhat.com>
4580 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4581 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4583 (is_attr_func): New function.
4584 (msp430_is_interrupt_func): New function.
4585 (is_naked_func): New function.
4586 (is_reentrant_func): New function.
4587 (is_critical_func): New function.
4588 (msp430_start_function): Add annotations for function attributes.
4589 (msp430_attr): New function.
4590 (msp430_attribute_table): New.
4591 (msp430_function_section): New function.
4592 (TARGET_ASM_FUNCTION_SECTION): Define.
4593 (msp430_builtin): New enum.
4594 (msp430_init_builtins): New function.
4595 (msp430_builtin_devl): New function.
4596 (msp430_expand_builtin): New function.
4597 (TARGET_INIT_BUILTINS): Define.
4598 (TARGET_EXPAND_BUILTINS): Define.
4599 (TARGET_BUILTIN_DECL): Define.
4600 (msp430_expand_prologue): Add support for naked, interrupt,
4601 critical and reentrant functions.
4602 (msp430_expand_epilogue): Likewise.
4603 (msp430_print_operand): Handle 'O' character.
4604 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4606 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4607 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4608 (pushm): Use a 'n' rather than an 'i' constraint.
4609 (msp_return): Add generation of the interrupt return instruction.
4610 (disable_interrupts): New pattern.
4611 (enable_interrupts): New pattern.
4612 (push_intr_state): New pattern.
4613 (pop_intr_state): New pattern.
4614 (bic_SR): New pattern.
4615 (bis_SR): New pattern.
4616 * doc/extend.texi: Document MSP430 function attributes and builtin
4619 2013-09-17 Richard Biener <rguenther@suse.de>
4621 PR tree-optimization/58432
4622 * tree-loop-distribution.c (tree_loop_distribution): Also
4623 scan PHIs for outside loop uses and seed a partition from them.
4625 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4627 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4628 (restructure_reference): Call backtrace_base_for_ref.
4630 2013-09-17 Alan Modra <amodra@gmail.com>
4633 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4636 2013-09-16 DJ Delorie <dj@redhat.com>
4638 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4639 vregs location for RL78/G10.
4640 (rl78_expand_prologue): Avoid SEL on G10.
4641 (rl78_expand_epilogue): Likewise.
4642 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4643 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4644 __RL78_G10__ when appropriate.
4645 (ASM_SPEC): Pass -mg10 along to the assembler.
4646 * config/rl78/rl78.md (sel_rb): Disable for G10.
4647 * config/rl78/rl78.opt: Add -mg10 option.
4648 * config/rl78/t-rl78: Add -mg10 multilib.
4650 2013-09-16 Xinliang David Li <davidxl@google.com>
4652 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4653 * omp-low.c (omp_max_vf): Ditto.
4654 (expand_omp_simd): Ditto.
4655 * tree-vectorizer.c (vectorize_loops): Ditto.
4656 (gate_vect_slp): Ditto.
4657 (gate_increase_alignment): Ditto.
4658 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4659 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4660 (gate_tree_vectorize): Name change.
4661 (tree_vectorize): Ditto.
4662 (pass_vectorize::gate): Call new function.
4663 (pass_vectorize::execute): Ditto.
4664 * opts.c: O3 default setting change.
4665 (finish_options): Check new flag.
4666 * doc/invoke.texi: Document new flags.
4667 * common.opt: New flags.
4669 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4671 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4672 (Condition Code): Don't reference it.
4673 * doc/tm.texi: Regenerate.
4675 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4678 * lra-constraints.c (undo_optional_reloads): Consider all optional
4679 reload even if it did not get a hard reg.
4681 2013-09-16 Teresa Johnson <tejohnson@google.com>
4683 * dumpfile.c (dump_loc): Remove newline emission.
4684 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4685 emission to dump_printf_loc calls where missing.
4686 (vect_mark_for_runtime_alias_test): Ditto.
4687 (vect_analyze_data_ref_dependence): Ditto.
4688 (vect_analyze_data_ref_dependences): Ditto.
4689 (vect_slp_analyze_data_ref_dependence): Ditto.
4690 (vect_slp_analyze_data_ref_dependences): Ditto.
4691 (vect_compute_data_ref_alignment): Ditto.
4692 (vect_update_misalignment_for_peel): Ditto.
4693 (vect_verify_datarefs_alignment): Ditto.
4694 (vector_alignment_reachable_p): Ditto.
4695 (vect_get_data_access_cost): Ditto.
4696 (vect_enhance_data_refs_alignment): Ditto.
4697 (vect_find_same_alignment_drs): Ditto.
4698 (vect_analyze_data_refs_alignment): Ditto.
4699 (vect_analyze_group_access): Ditto.
4700 (vect_analyze_data_ref_access): Ditto.
4701 (vect_analyze_data_ref_accesses): Ditto.
4702 (vect_prune_runtime_alias_test_list): Ditto.
4703 (vect_analyze_data_refs): Ditto.
4704 (vect_create_addr_base_for_vector_ref): Ditto.
4705 (vect_create_data_ref_ptr): Ditto.
4706 (vect_grouped_store_supported): Ditto.
4707 (vect_grouped_load_supported): Ditto.
4708 * value-prof.c (check_counter): Ditto.
4709 (check_ic_target): Ditto.
4710 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4711 (vect_recog_widen_mult_pattern): Ditto.
4712 (vect_recog_widen_sum_pattern): Ditto.
4713 (vect_recog_over_widening_pattern): Ditto.
4714 (vect_recog_widen_shift_pattern): Ditto.
4715 (vect_recog_rotate_pattern): Ditto.
4716 (vect_recog_vector_vector_shift_pattern): Ditto.
4717 (vect_recog_divmod_pattern): Ditto.
4718 (vect_recog_mixed_size_cond_pattern): Ditto.
4719 (vect_recog_bool_pattern): Ditto.
4720 (vect_pattern_recog_1): Ditto.
4721 (vect_pattern_recog): Ditto.
4722 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4723 (vect_is_simple_iv_evolution): Ditto.
4724 (vect_analyze_scalar_cycles_1): Ditto.
4725 (vect_get_loop_niters): Ditto.
4726 (vect_analyze_loop_1): Ditto.
4727 (vect_analyze_loop_form): Ditto.
4728 (vect_analyze_loop_operations): Ditto.
4729 (vect_analyze_loop_2): Ditto.
4730 (vect_analyze_loop): Ditto.
4731 (report_vect_op): Ditto.
4732 (vect_is_slp_reduction): Ditto.
4733 (vect_is_simple_reduction_1): Ditto.
4734 (vect_get_known_peeling_cost): Ditto.
4735 (vect_estimate_min_profitable_iters): Ditto.
4736 (vect_model_reduction_cost): Ditto.
4737 (vect_model_induction_cost): Ditto.
4738 (get_initial_def_for_induction): Ditto.
4739 (vect_create_epilog_for_reduction): Ditto.
4740 (vectorizable_reduction): Ditto.
4741 (vectorizable_induction): Ditto.
4742 (vectorizable_live_operation): Ditto.
4743 (vect_loop_kill_debug_uses): Ditto.
4744 (vect_transform_loop): Ditto.
4745 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4746 (vect_stmt_relevant_p): Ditto.
4747 (process_use): Ditto.
4748 (vect_mark_stmts_to_be_vectorized): Ditto.
4749 (vect_model_simple_cost): Ditto.
4750 (vect_model_promotion_demotion_cost): Ditto.
4751 (vect_model_store_cost): Ditto.
4752 (vect_get_store_cost): Ditto.
4753 (vect_model_load_cost): Ditto.
4754 (vect_get_load_cost): Ditto.
4755 (vect_init_vector_1): Ditto.
4756 (vect_get_vec_def_for_operand): Ditto.
4757 (vect_finish_stmt_generation): Ditto.
4758 (vectorizable_call): Ditto.
4759 (vectorizable_conversion): Ditto.
4760 (vectorizable_assignment): Ditto.
4761 (vectorizable_shift): Ditto.
4762 (vectorizable_operation): Ditto.
4763 (vectorizable_store): Ditto.
4764 (vectorizable_load): Ditto.
4765 (vectorizable_condition): Ditto.
4766 (vect_analyze_stmt): Ditto.
4767 (vect_transform_stmt): Ditto.
4768 (vect_is_simple_use): Ditto.
4769 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4770 (vect_can_advance_ivs_p): Ditto.
4771 (vect_update_ivs_after_vectorizer): Ditto.
4772 (vect_do_peeling_for_loop_bound): Ditto.
4773 (vect_gen_niters_for_prolog_loop): Ditto.
4774 (vect_update_inits_of_drs): Ditto.
4775 (vect_create_cond_for_alias_checks): Ditto.
4776 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4777 (vect_build_slp_tree_1): Ditto.
4778 (vect_supported_load_permutation_p): Ditto.
4779 (vect_analyze_slp_instance): Ditto.
4780 (vect_analyze_slp): Ditto.
4781 (vect_make_slp_decision): Ditto.
4782 (vect_detect_hybrid_slp): Ditto.
4783 (vect_bb_vectorization_profitable_p): Ditto.
4784 (vect_slp_analyze_bb_1): Ditto.
4785 (vect_update_slp_costs_according_to_vf): Ditto.
4786 (vect_get_mask_element): Ditto.
4787 (vect_transform_slp_perm_load): Ditto.
4788 (vect_schedule_slp_instance): Ditto.
4789 (vect_schedule_slp): Ditto.
4790 (vect_slp_transform_bb): Ditto.
4791 * profile.c (read_profile_edge_counts): Ditto.
4792 (compute_branch_probabilities): Ditto.
4793 * coverage.c (get_coverage_counts): Ditto.
4795 2013-09-16 Diego Novillo <dnovillo@google.com>
4797 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4799 2013-09-16 Jan Hubicka <jh@suse.cz>
4801 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4802 abstract functions; for static functions check the presence of body.
4804 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4806 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4807 * config/aarch64/aarch64-simd.md
4808 (aarch64_mla_elt<mode>): New.
4809 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4810 (aarch64_mls_elt<mode>): Likewise.
4811 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4812 (aarch64_fma4_elt<mode>): Likewise.
4813 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4814 (aarch64_fma4_elt_to_128v2df): Likewise.
4815 (aarch64_fma4_elt_to_64df): Likewise.
4816 (fnma<mode>4): Likewise.
4817 (aarch64_fnma4_elt<mode>): Likewise.
4818 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4819 (aarch64_fnma4_elt_to_128v2df): Likewise.
4820 (aarch64_fnma4_elt_to_64df): Likewise.
4821 * config/aarch64/iterators.md (VDQSF): New.
4822 * config/aarch64/arm_neon.h
4823 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4824 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4826 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4828 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4829 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4830 (aarch64_mul3_elt_to_128df): Likewise.
4831 (aarch64_mul3_elt_to_64v2df): Likewise.
4832 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4834 (VMUL_CHANGE_NLANES) Likewise.
4837 * config/aarch64/arm_neon.h
4838 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4840 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4842 * config/aarch64/arm_neon.h
4843 (vcvtx_high_f32_f64): Fix parameters.
4845 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4846 Uros Bizjak <ubizjak@gmail.com>
4848 * config/alpha.c: Include tree-ssanames.h.
4850 2013-09-16 Richard Biener <rguenther@suse.de>
4852 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4853 (dot_rdg_1): Handle control_dd.
4854 (create_edge_for_control_dependence): New function.
4855 (create_rdg_edges): Add control dependences if asked for.
4856 (build_rdg): Likewise.
4857 (generate_loops_for_partition): If there are not necessary
4858 control stmts remove all their dependencies.
4859 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4860 (distribute_loop): Pass on control dependences.
4861 (tree_loop_distribution): Compute control dependences and remove
4862 restriction on number of loop nodes.
4864 2013-09-16 Jakub Jelinek <jakub@redhat.com>
4866 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4869 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4871 * cse.c (try_const_anchors): Punt on CC modes.
4873 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4875 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4877 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
4880 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4883 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4886 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
4888 (darwin_objc1_section): Likewise.
4889 (darwin_file_end): Emit Image Info section when required.
4891 2013-09-14 Jan Hubicka <jh@suse.cz>
4893 * tree-into-ssa.c (gate_into_ssa): New.
4894 (pass_data_build_ssa): Use it.
4895 * cgraph.h (expand_thunk): Update prototype.
4896 * cgraphunit.c (analyze_function): Expand thunks early.
4897 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4898 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4899 set lowered flag; do not add new function.
4900 (assemble_thunks_and_aliases): Update.
4901 * tree-ssa.c (gate_init_datastructures): New gate.
4902 (pass_data_init_datastructures): Use it.
4904 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4907 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
4908 xmm register set described in the psABI.
4910 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4912 * dwarf2out.c (should_emit_struct_debug): Add check
4913 for type_decl variable is not NULL.
4915 2013-09-13 Jacek Caban <jacek@codeweavers.com>
4917 * config.gcc: Use new winnt-c.c target hooks
4918 * config/t-winnt: New file
4919 * config/winnt-c.c: New file
4920 * doc/tm.texi.in: Document new hook
4921 * doc/tm.texi: Regenerated
4923 2013-09-13 Jan Hubicka <jh@suse.cz>
4926 * ipa-inline.c (check_callers): New function.
4927 (check_caller_edge): Remove.
4928 (want_inline_function_to_all_callers_p): Also permit alises that are
4930 (inline_to_all_callers): Terminate the walk when devirtualization
4931 introduce new calls.
4933 2013-09-13 Jan Hubicka <jh@suse.cz>
4935 * ipa-inline-analysis.c (struct growth_data): Add node.
4936 (do_estimate_growth_1): Fix detection of recursion.
4938 2013-09-13 Jakub Jelinek <jakub@redhat.com>
4940 PR tree-optimization/58392
4941 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4942 to avoid shadowing of outer loop variable. If
4943 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4944 replace_by_duplicate_decl simduid of loops that have it set and
4945 set dest_cfun->has_simduid_loops and/or
4946 dest_cfun->has_force_vect_loops.
4947 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4948 instead of maybe_lookup_decl.
4949 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4950 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4951 call. Copy over force_vect and copy and remap simduid. Set
4952 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4953 (copy_cfg_body): Remove blocks_to_copy argument. Use
4954 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4955 caller. Don't set cfun->has_simduid_loops and/or
4956 cfun->has_force_vect_loops here.
4957 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4959 (expand_call_inline, tree_function_versioning): Adjust copy_body
4962 2013-09-13 Martin Jambor <mjambor@suse.cz>
4965 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4966 the assert if the edge was a speculative one.
4968 2013-09-13 Richard Biener <rguenther@suse.de>
4970 * tree-data-ref.h (known_dependences_p): Move here ...
4971 * tree-loop-distribution.c (known_dependences_p): ... from here.
4972 (dump_rdg_component, debug_rdg_component): Remove.
4974 (generate_loops_for_partition): Use gimple_uid instead of
4975 relying on matching stmt visit order.
4976 (rdg_build_partitions): Take starting stmt vector.
4977 (ldist_gen): Merge into ...
4978 (distribute_loop): ... this function. Do not compute starting
4980 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
4982 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4984 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
4985 Set type attribute correctly. Set predicable_short_it attribute.
4986 (cmpsi_shiftsi): Remove %? from output template.
4988 2013-09-13 Richard Biener <rguenther@suse.de>
4990 * tree-loop-distribution.c (struct rdg_component,
4991 rdg_defs_used_in_other_loops_p, free_rdg_components,
4992 rdg_build_components): Remove.
4993 (stmts_from_loop): Do not record virtual PHIs.
4994 (generate_loops_for_partition): Skip virtual PHIs.
4995 (build_rdg_partition_for_component): Rename to ...
4996 (build_rdg_partition_for_vertex): ... this and adjust.
4997 (rdg_build_partitions): Take a vector of starting vertices
4998 instead of components. Remove unnecessary leftover handling.
4999 (ldist_gen): Do not build components or record other stores.
5000 (distribute_loop): Do not distribute loops containing stmts
5003 2013-09-13 Christian Bruel <christian.bruel@st.com>
5006 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5008 2013-09-13 Kai Tietz <ktietz@redhat.com>
5010 * config.gcc: Separate cases for mingw and cygwin targets,
5011 and add 64-bit cygwin target case.
5013 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5014 dll-export inline-functions.
5015 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5018 2013-09-13 Jeff Law <law@redhat.com>
5022 2013-09-06 Jeff Law <law@redhat.com>
5024 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5025 edge implied equivalences into successor phis.
5027 2013-09-12 DJ Delorie <dj@redhat.com>
5029 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5032 2013-09-12 Brooks Moses <bmoses@google.com>
5035 * Makefile.in: Do not install driver binaries in $(target)/bin.
5037 2013-09-12 DJ Delorie <dj@redhat.com>
5039 * config/rl78/rl78.opt (mrelax): New.
5040 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5041 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5043 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5045 (rl78_expand_epilogue): Likewise.
5047 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5050 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5051 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5052 pointer elimination with using remove_reg_equal_offset_note.
5054 2013-09-12 DJ Delorie <dj@redhat.com>
5056 * config/msp430/: New port.
5057 * config.gcc (msp430): Added.
5058 * doc/invoke.texi: Document MSP430 options.
5059 * doc/install.texi: Document msp430-elf
5060 * doc/md.texi: Document msp430-elf
5061 * doc/contrib.texi: Document msp430-elf
5063 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5066 2013-09-12 Martin Jambor <mjambor@suse.cz>
5069 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5070 reference descriptor is NULL.
5071 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5074 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5076 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5077 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5078 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5079 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5080 make_temp_ssa_name): move to tree-ssanames.h
5081 * tree-ssa-alias.h: Move prototype.
5082 * tree-ssa.h: Include tree-ssanames.h.
5083 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5084 * tree-ssanames.h: New. Move items from tree-flow*.h
5085 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5087 2013-09-12 Richard Biener <rguenther@suse.de>
5089 PR tree-optimization/58404
5090 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5091 propagate non-invariant addresses into dereferences wrapped
5092 in component references.
5094 2013-09-12 Richard Biener <rguenther@suse.de>
5096 PR tree-optimization/58402
5097 * passes.def: Move pass_late_warn_uninitialized later.
5099 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5101 * tree-ssa.h: New. Move content from tree-flow.h and
5103 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5104 Move prototypes belonging to tree-ssa.c.
5105 * tree-flow-inline.h (redirect_edge_var_map_def,
5106 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5108 * gimple.h: Adjust prototypes.
5109 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5111 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5112 * tree.h: Move prototype to tree-ssa.h.
5113 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5114 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5115 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5116 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5117 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5118 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5119 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5120 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5121 graphite-blocking.c, graphite-clast-to-gimple.c,
5122 graphite-dependences.c, graphite-interchange.c,
5123 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5124 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5125 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5126 ipa-reference.c, ipa-split.c, ipa-utils.c,
5127 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5128 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5129 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5130 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5131 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5132 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5133 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5134 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5135 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5136 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5137 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5138 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5139 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5140 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5141 value-prof.c, var-tracking.c,
5142 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5144 2013-09-12 Richard Biener <rguenther@suse.de>
5146 PR tree-optimization/58396
5147 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5148 (build_rdg): Take a loop-nest parameter, fix memleaks.
5149 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5151 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5153 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5156 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5158 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5159 The FMA instruction names should have a 'v' prefix.
5161 2013-09-12 Richard Biener <rguenther@suse.de>
5163 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5164 (dot_rdg): Use popen instead of system in optional code.
5165 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5166 (already_processed_vertex_p): Adjust.
5167 (has_anti_or_output_dependence, predecessor_has_mem_write,
5168 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5169 rdg_flag_uses): Remove.
5170 (rdg_flag_vertex): Simplify.
5171 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5173 (build_rdg_partition_for_component): Process the first vertex
5174 of a component only.
5175 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5177 2013-09-12 Alan Modra <amodra@gmail.com>
5179 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5181 2013-09-11 DJ Delorie <dj@redhat.com>
5182 Nick Clifton <nickc@redhat.com>
5184 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5185 (rl78_stack_based_mem): New.
5186 * config/rl78/constraints.md (Iv08): New.
5195 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5197 (movsi): Change from expand to insn-and-split.
5198 (ashrsi3): Clobber AX.
5202 * config/rl78/rl78.md (CC_REG): Fix.
5203 (addsi3): Allow memory and immediate operands.
5204 (addsi3_internal): Split into...
5205 (addsi3_internal_virt): ...new, and ...
5206 (addsi3_internal_real): ...new.
5208 (subsi3_internal_virt): New.
5209 (subsi3_internal_real): New.
5210 (mulsi3): Add memory operand.
5211 (mulsi3_rl78): Likewise.
5212 (mulsi3_g13): Likewise.
5213 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5214 (cbranchqi4_real): Add more constraint options.
5215 (cbranchhi4_real): Expand pattern.
5216 (cbranchhi4_real_signed): New.
5217 (cbranchhi4_real_inverted): New.
5218 (cbranchsi4_real_lt): New.
5219 (cbranchsi4_real_ge): New.
5220 (cbranchsi4_real_signed): New.
5221 (cbranchsi4_real): New.
5223 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5225 (lshrsi3_virt): Likewise.
5226 (ashlsi3_virt): Likewise.
5227 (cbranchqi4_virt_signed): New.
5228 (cbranchhi4_virt_signed): New.
5229 (cbranchsi4_virt): New.
5230 * config/rl78/rl78.c: Whitespace fixes throughout.
5231 (move_elim_pass): New.
5232 (pass_data_rl78_move_elim): New.
5233 (pass_rl78_move_elim): New.
5234 (make_pass_rl78_move_elim): New.
5235 (rl78_devirt_info): Run devirt earlier.
5236 (rl78_move_elim_info): New.
5237 (rl78_asm_file_start): Register it.
5238 (rl78_split_movsi): New.
5239 (rl78_as_legitimate_address): Allow virtual base registers when
5241 (rl78_addr_space_convert): Remove spurious debug stuff.
5242 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5243 (rl78_print_operand): More cases for not printing '#'.
5244 (rl78_expand_compare): Remove most of the logic.
5245 (content_memory): New.
5246 (clear_content_memory): New.
5247 (get_content_index): New.
5248 (get_content_name): New.
5249 (display_content_memory): New.
5250 (update_content): New.
5251 (record_content): New.
5252 (already_contains): New.
5253 (insn_ok_now): Re-recog insns with virtual registers.
5254 (add_postponed_content_update): New.
5255 (process_postponed_content_update): New.
5256 (gen_and_emit_move): New.
5257 (transcode_memory_rtx): Record new location content.
5258 Use gen_and_emit_move.
5259 (force_into_acc): New.
5260 (move_to_acc): Use gen_and_emit_move.
5261 (move_from_acc): Likewise.
5262 (move_acc_to_reg): Likewise.
5263 (move_to_x): Likewise.
5264 (move_to_hl): Likewise.
5265 (move_to_de): Likewise.
5266 (rl78_alloc_physical_registers_op1): Record location content.
5267 (has_constraint): New.
5268 (rl78_alloc_physical_registers_op2): Record location content.
5270 (rl78_alloc_physical_registers_ro1): Likewise.
5271 (rl78_alloc_physical_registers_cmp): Likewise.
5272 (rl78_alloc_physical_registers_umul): Likewise.
5273 (rl78_alloc_address_registers_macax): New.
5274 (rl78_alloc_physical_registers): Initialize and set location
5275 content memory as needed.
5276 (rl78_reorg): Make sure split2 is called.
5277 (rl78_rtx_costs): New.
5279 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5281 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5282 for (not (neg ...)) and (neg (not ...)) cases.
5284 2013-09-11 Richard Biener <rguenther@suse.de>
5287 * passes.def: Split critical edges before late uninit warning passes.
5288 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5290 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5292 PR tree-optimization/58385
5293 * fold-const.c (build_range_check): If both low and high are NULL,
5294 use omit_one_operand_loc to preserve exp side-effects.
5296 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5298 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5300 2013-09-11 Richard Biener <rguenther@suse.de>
5302 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5303 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5304 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5305 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5306 stmts_from_loop, known_dependences_p, build_empty_rdg,
5307 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5308 * tree-loop-distribution.c: ... here.
5309 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5310 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5311 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5312 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5313 * tree-loop-distribution.c: ... here.
5314 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5315 (struct partition_s): Add loops member.
5316 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5317 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5318 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5320 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5321 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5322 Sergey Lega <sergey.s.lega@intel.com>
5323 Anna Tikhonova <anna.tikhonova@intel.com>
5324 Ilya Tocar <ilya.tocar@intel.com>
5325 Andrey Turetskiy <andrey.turetskiy@intel.com>
5326 Ilya Verbin <ilya.verbin@intel.com>
5327 Kirill Yukhin <kirill.yukhin@intel.com>
5328 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5330 * config/i386/constraints.md (k): New.
5332 * config/i386/i386.c (const regclass_map): Add new mask registers.
5333 (dbx_register_map): Ditto.
5334 (dbx64_register_map): Ditto.
5335 (svr4_dbx_register_map): Ditto.
5336 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5338 (ix86_preferred_reload_class): Disable constants for mask registers.
5339 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5340 (ix86_hard_regno_mode_ok): Support new mask registers.
5341 (x86_order_regs_for_local_alloc): Ditto.
5342 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5343 (FIXED_REGISTERS): Add new mask registers.
5344 (CALL_USED_REGISTERS): Ditto.
5345 (REG_ALLOC_ORDER): Ditto.
5346 (VALID_MASK_REG_MODE): New.
5347 (FIRST_MASK_REG): Ditto.
5348 (LAST_MASK_REG): Ditto.
5349 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5350 (MAYBE_MASK_CLASS_P): New.
5351 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5352 (REG_CLASS_CONTENTS): Ditto.
5353 (MASK_REGNO_P): New.
5354 (ANY_MASK_REG_P): Ditto.
5355 (HI_REGISTER_NAMES): Add new mask registers.
5356 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5357 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5359 (attribute "type"): Add mskmov, msklog.
5360 (attribute "length_immediate"): Support them.
5361 (attribute "memory"): Ditto.
5362 (attribute "prefix_0f"): Ditto.
5363 (*movhi_internal): Support new mask registers.
5364 (*movqi_internal): Ditto.
5365 (define_split): Split out clobber pattern is a logic
5366 insn on mask registers.
5367 (*k<logic><mode>): New.
5368 (*andhi_1): Extend to support mask regs.
5369 (*andqi_1): Extend to support mask regs.
5371 (define_split): Split and-not to and and not if operands
5373 (*<code><mode>_1): Separate HI mode to new pattern...
5374 (*<code>hi_1): This.
5375 (*<code>qi_1): Extend to support mask regs.
5377 (kortestzhi): Ditto.
5378 (kortestchi): Ditto.
5380 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5381 (*one_cmplhi2_1): ...Here, now with mask registers support.
5382 (*one_cmplqi2_1): Support new mask registers.
5383 (HI/QImode arithmetics splitter): Don't split if mask registers
5385 (HI/QImode not splitter): Ditto.
5386 * config/i386/predicated.md (mask_reg_operand): New.
5387 (general_reg_operand): Ditto.
5389 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5391 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5392 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5394 2013-09-10 Jeff Law <law@redhat.com>
5396 PR tree-optimization/58380
5397 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5398 of threading through a buried loop header.
5400 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5401 return value for single successor case.
5403 2013-09-10 Jan Hubicka <jh@suse.cz>
5405 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5407 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5410 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5411 support conditional execution.
5412 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5414 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5416 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5418 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5419 Check destination too to check move insn.
5420 (undo_optional_reloads): Add check that the original peudo did not
5421 changed its allocation and the optional reload was inherited on last
5422 inheritance pass. Break loop after deciding to keep optional reload.
5423 (lra_undo_inheritance): Add check that inherited pseudo still in
5426 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5428 * config/aarch64/aarch64.md (generic_sched): New.
5429 * config/aarch64/aarch64-generic.md (load): Make conditional
5430 on generic_sched attribute.
5431 (nonload): Likewise.
5433 2013-09-10 Jan Hubicka <jh@suse.cz>
5435 * lto-cgraph.c: Include ipa-utils.h.
5436 (compute_ltrans_boundary): Also add possible targets into the boundary.
5438 2013-09-10 Jan Hubicka <jh@suse.cz>
5440 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5441 VAR_DECL of vtable rather than full expression.
5443 2013-09-10 Jan Hubicka <jh@suse.cz>
5444 Paolo Carlini <paolo.carlini@oracle.com>
5446 * cgraphunit.c (analyze_functions): Save input_location, set it
5447 to UNKNOWN_LOCATION and restore it at the end.
5449 2013-09-10 Martin Jambor <mjambor@suse.cz>
5451 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5452 represented by a thunk.
5454 2013-09-10 Jeff Law <law@redhat.com>
5456 PR tree-optimization/58343
5457 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5458 jump threading requests through loop headers buried in the middle
5459 of a jump threading path.
5461 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5462 in return value/type.
5464 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5466 PR rtl-optimization/58365
5467 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5468 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5471 2013-09-10 Richard Biener <rguenther@suse.de>
5473 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5474 * tree-data-ref.c (create_rdg_vertices): Collect all data
5475 references, signal failure to the caller, use data-ref API.
5476 (build_rdg): Compute data references only once. Maintain lifetime
5477 of data references and data dependences from within RDG.
5478 (free_rdg): Free dependence relations.
5479 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5480 inventing extra dependences.
5481 (distribute_loop): Update for RDG API changes.
5483 2013-09-10 Kai Tietz <ktietz@redhat.com>
5485 * doc/invoke.texi (fms-extensions): Document changed
5486 behavior for ms-abi targets.
5487 * config/i386/i386.c (ix86_option_override_internal):
5488 Set default value of option -fms-extension for ms-abi targets.
5490 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5492 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5494 2013-09-10 Alan Modra <amodra@gmail.com>
5497 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5499 2013-09-10 Alan Modra <amodra@gmail.com>
5501 * config/rs6000/predicates.md (add_cint_operand): New.
5502 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5503 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5504 using add_cint_operand.
5505 (largetoc_high_plus_aix): Likewise.
5507 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5509 PR tree-optimization/58364
5510 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5511 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5512 the current range can't be an unconditional true or false.
5514 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5516 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5518 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5520 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5522 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5525 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5527 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5529 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5530 NO_REGS for immediate that can't be moved directly into FP_REGS.
5532 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5534 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5535 comparison with negated operand.
5536 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5539 2013-09-09 Richard Biener <rguenther@suse.de>
5542 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5543 of a subloop record all its block as affecting loop-closed SSA form.
5545 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5547 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5548 of an rtx/bitpos pair.
5549 (store_fixed_bit_field): Update accordingly.
5551 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5553 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5555 * builtins.c (expand_errno_check): Likewise.
5556 * dwarf2cfi.c (init_return_column_size): Likewise.
5557 * except.c (sjlj_mark_call_sites): Likewise.
5558 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5559 * lra-constraints.c (emit_inc): Likewise.
5560 * ree.c (combine_set_extension): Likewise.
5561 * regmove.c (fixup_match_2): Likewise.
5562 * reload1.c (inc_for_reload): Likewise.
5564 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5566 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5567 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5568 Use gen_int_mode with the mode of the associated simplify_* call.
5569 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5570 * expmed.c (expand_shift_1): Likewise.
5571 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5572 * loop-iv.c (iv_number_of_iterations): Likewise.
5573 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5574 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5576 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5578 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5579 of the associated expand_* call.
5580 (asan_emit_stack_protection): Likewise.
5581 * builtins.c (round_trampoline_addr): Likewise.
5582 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5583 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5584 (emit_store_flag): Likewise.
5585 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5586 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5588 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5589 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5590 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5591 * modulo-sched.c (generate_prolog_epilog): Likewise.
5592 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5593 (expand_ctz, expand_ffs, expand_unop): Likewise.
5595 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5597 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5598 of the associated gen_rtx_* call.
5599 * caller-save.c (init_caller_save): Likewise.
5600 * combine.c (find_split_point, make_extraction): Likewise.
5601 (make_compound_operation): Likewise.
5602 * dwarf2out.c (mem_loc_descriptor): Likewise.
5603 * explow.c (plus_constant, probe_stack_range): Likewise.
5604 * expmed.c (expand_mult_const): Likewise.
5605 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5606 * reload1.c (init_reload): Likewise.
5607 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5608 * var-tracking.c (adjust_mems): Likewise.
5609 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5610 rather than gen_rtx_fmt_ee.
5612 2013-09-09 Jan Hubicka <jh@suse.cz>
5615 * value-prof.c (gimple_ic): Copy also abnormal edges.
5617 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5619 * asan.c (asan_shadow_cst): Use gen_int_mode.
5621 2013-09-08 Jan Hubicka <jh@suse.cz>
5623 * ipa-profile.c: Add toplevel comment.
5624 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5625 (contains_hot_call_p): New function.
5626 (ipa_propagate_frequency): Set frequencies based on counts when
5628 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5629 profile; do not tamper with profile after inlining if it is read.
5631 2013-09-08 Jan Hubicka <jh@suse.cz>
5633 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5636 2013-09-08 Jan Hubicka <jh@suse.cz>
5638 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5641 2013-09-08 Jeff Law <law@redhat.com>
5644 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5647 2013-09-08 Andi Kleen <ak@linux.intel.com>
5649 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5651 2013-09-08 Jan Hubicka <jh@suse.cz>
5653 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5655 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5657 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5658 for non-debug insns.
5659 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5660 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5661 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5664 2013-09-08 Jan Hubicka <jh@suse.cz>
5666 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5667 targets and devirtualize to BUILT_IN_UNREACHABLE.
5668 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5669 * ipa.c (walk_polymorphic_call_targets): New function.
5670 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5671 functions; use the new timevar.
5672 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5673 was removed from the program.
5674 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5675 not consider it in the walk when its vtable is dead.
5676 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5678 (devirt_variable_node_removal_hook): New function.
5679 (possible_polymorphic_call_targets): Also register
5680 devirt_variable_node_removal_hook.
5681 (ipa_devirt): Do not do non-speculative devirtualization.
5682 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5684 2013-09-08 Jan Hubicka <jh@suse.cz>
5686 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5687 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5688 varpool_remove_variable_insertion_hook): Declare.
5689 * varpool.c (varpool_node_hook_list): New structure.
5690 (first_varpool_node_removal_hook,
5691 first_varpool_variable_insertion_hook): New variables.
5692 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5693 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5694 varpool_remove_variable_insertion_hook,
5695 varpool_call_variable_insertion_hooks): New functions.
5696 (varpool_remove_node): Use it.
5698 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5701 * diagnostic.c (diagnostic_build_prefix): When s.file is
5702 "<built-in>" don't output line and column numbers.
5704 2013-09-06 Jan Hubicka <jh@suse.cz>
5706 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5707 * lto-streamer-out.c: Stream thunks, too.
5708 * lto-streamer-in.c (input_function): Pop cfun here
5709 (lto_read_body): Instead of here.
5711 2013-09-06 Caroline Tice <cmtice@google.com>
5713 * doc/install.texi: Add documentation for the --enable-vtable-verify
5714 and the --disable-libvtv configure options.
5716 2013-09-06 Jeff Law <law@redhat.com>
5718 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5719 edge implied equivalences into successor phis.
5721 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5723 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5725 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5727 * resource.c (mark_target_live_regs): Compute resources taking
5728 into account if a call is predicated or not.
5730 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5732 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5733 the compiler in the function names.
5735 2013-09-06 Jan Hubicka <jh@suse.cz>
5738 * ipa-inline.c (has_caller_p): New function.
5739 (want_inline_function_to_all_callers_p): Use it.
5740 (sum_callers, inline_to_all_callers): Break out from ...
5741 (ipa_inline): ... here.
5743 2013-09-06 Jan Hubicka <jh@suse.cz>
5745 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5746 only when AVX is enabled.
5748 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5750 * config/aarch64/aarch64.md
5751 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5752 is fpsimd_<load/store>2.
5753 (load_pair<mode>): Likewise.
5754 (store_pair<mode>): Likewise.
5756 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5758 * config/arm/types.md (type): Add "mrs" type.
5759 * config/aarch64/aarch64.md
5760 (aarch64_load_tp_hard): Make type "mrs".
5762 (load_tp_hard): Make type "mrs".
5763 * config/arm/cortex-a15.md: Update with new attributes.
5764 * config/arm/cortex-a5.md: Update with new attributes.
5765 * config/arm/cortex-a53.md: Update with new attributes.
5766 * config/arm/cortex-a7.md: Update with new attributes.
5767 * config/arm/cortex-a8.md: Update with new attributes.
5768 * config/arm/cortex-a9.md: Update with new attributes.
5769 * config/arm/cortex-m4.md: Update with new attributes.
5770 * config/arm/cortex-r4.md: Update with new attributes.
5771 * config/arm/fa526.md: Update with new attributes.
5772 * config/arm/fa606te.md: Update with new attributes.
5773 * config/arm/fa626te.md: Update with new attributes.
5774 * config/arm/fa726te.md: Update with new attributes.
5776 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5778 * config/aarch64/aarch64.md
5779 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5780 (*movtf_aarch64): Likewise.
5782 (thumb1_movdi_insn): Use "multiple" for type where more than one
5783 instruction is used for a move.
5784 (*arm32_movhf): Likewise.
5785 (*thumb_movdf_insn): Likewise.
5787 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5789 * config/arm/types.md (type): Rename fcpys to fmov.
5791 (*arm_movsi_vfp): Rename type fcpys as fmov.
5792 (*thumb2_movsi_vfp): Likewise
5793 (*movhf_vfp_neon): Likewise
5794 (*movhf_vfp): Likewise
5795 (*movsf_vfp): Likewise
5796 (*thumb2_movsf_vfp): Likewise
5797 (*movsfcc_vfp): Likewise
5798 (*thumb2_movsfcc_vfp): Likewise
5799 * config/aarch64/aarch64-simd.md
5800 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5801 * config/aarch64/aarch64.md
5802 (*movsi_aarch64): Replace type mov_reg with fmovs.
5803 (*movdi_aarch64): Likewise
5804 (*movsf_aarch64): Likewise
5805 (*movdf_aarch64): Likewise
5807 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5808 * config/arm/iwmmxt.md
5809 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5810 * config/arm/arm1020e.md: Update with new attributes.
5811 * config/arm/cortex-a15-neon.md: Update with new attributes.
5812 * config/arm/cortex-a5.md: Update with new attributes.
5813 * config/arm/cortex-a53.md: Update with new attributes.
5814 * config/arm/cortex-a7.md: Update with new attributes.
5815 * config/arm/cortex-a8-neon.md: Update with new attributes.
5816 * config/arm/cortex-a9.md: Update with new attributes.
5817 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5818 * config/arm/cortex-r4f.md: Update with new attributes.
5819 * config/arm/marvell-pj4.md: Update with new attributes.
5820 * config/arm/vfp11.md: Update with new attributes.
5822 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5824 * config/aarch64/aarch64.md
5825 (*madd<mode>): Fix type attribute.
5826 (*maddsi_uxtw): Likewise.
5827 (*msub<mode>): Likewise.
5828 (*msubsi_uxtw): Likewise.
5829 (<su_optab>maddsidi4): Likewise.
5830 (<su_optab>msubsidi4): Likewise.
5832 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5834 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5835 * config/arm/arm.md (core_cycles): Remove fdiv.
5836 * config/arm/vfp.md:
5837 (*sqrtsf2_vfp): Update for attribute changes.
5838 (*sqrtdf2_vfp): Likewise.
5839 * config/aarch64/aarch64.md:
5840 (sqrt<mode>2): Update for attribute changes.
5841 * config/arm/arm1020e.md: Update with new attributes.
5842 * config/arm/cortex-a15-neon.md: Update with new attributes.
5843 * config/arm/cortex-a5.md: Update with new attributes.
5844 * config/arm/cortex-a53.md: Update with new attributes.
5845 * config/arm/cortex-a7.md: Update with new attributes.
5846 * config/arm/cortex-a8-neon.md: Update with new attributes.
5847 * config/arm/cortex-a9.md: Update with new attributes.
5848 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5849 * config/arm/cortex-r4f.md: Update with new attributes.
5850 * config/arm/marvell-pj4.md: Update with new attributes.
5851 * config/arm/vfp11.md: Update with new attributes.
5853 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5855 * config/arm/types.md
5856 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5857 * config/aarch64/aarch64.md
5858 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5860 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5861 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5862 (float<GPI:mode><GPF:mode>2): Likewise.
5864 (*truncsisf2_vfp): Update with new attributes.
5865 (*truncsidf2_vfp): Likewise.
5866 (fixuns_truncsfsi2): Likewise.
5867 (fixuns_truncdfsi2): Likewise.
5868 (*floatsisf2_vfp): Likewise.
5869 (*floatsidf2_vfp): Likewise.
5870 (floatunssisf2): Likewise.
5871 (floatunssidf2): Likewise.
5872 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5873 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5874 * config/arm/arm1020e.md: Update with new attributes.
5875 * config/arm/cortex-a15-neon.md: Update with new attributes.
5876 * config/arm/cortex-a5.md: Update with new attributes.
5877 * config/arm/cortex-a53.md: Update with new attributes.
5878 * config/arm/cortex-a7.md: Update with new attributes.
5879 * config/arm/cortex-a8-neon.md: Update with new attributes.
5880 * config/arm/cortex-a9.md: Update with new attributes.
5881 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5882 * config/arm/cortex-r4f.md: Update with new attributes.
5883 * config/arm/marvell-pj4.md: Update with new attributes.
5884 * config/arm/vfp11.md: Update with new attributes.
5886 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5888 * config/aarch64/arm_neon.h
5889 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5890 (vqtbx<1,2,3,4><q>_s8): Likewise.
5892 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5894 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5895 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
5896 (add<mode>3): Add type attribute.
5897 (add<mode>3): Likewise.
5898 (usadd<mode>3): Likewise.
5899 (ssadd<mode>3): Likewise.
5900 (sub<mode>3): Likewise.
5901 (sub<mode>3): Likewise.
5902 (ussub<mode>3): Likewise.
5903 (sssub<mode>3): Likewise.
5904 (ssmulsa3): Likewise.
5905 (usmulusa3): Likewise.
5906 (arm_usatsihi): Likewise.
5908 (*movdi_vfp): Add types for all instructions.
5909 (*movdi_vfp_cortexa8): Likewise.
5910 (*movhf_vfp_neon): Likewise.
5911 (*movhf_vfp): Likewise.
5912 (*movdf_vfp): Likewise.
5913 (*thumb2_movdf_vfp): Likewise.
5914 (*thumb2_movdfcc_vfp): Likewise.
5915 * config/arm/arm.md: Add type attribute to all insn patterns.
5916 (*thumb1_adddi3): Add type attribute.
5917 (*arm_adddi3): Likewise.
5918 (*adddi_sesidi_di): Likewise.
5919 (*adddi_zesidi_di): Likewise.
5920 (*thumb1_addsi3): Likewise.
5921 (addsi3_compare0): Likewise.
5922 (*addsi3_compare0_scratch): Likewise.
5923 (*compare_negsi_si): Likewise.
5924 (cmpsi2_addneg): Likewise.
5925 (*addsi3_carryin_<optab>): Likewise.
5926 (*addsi3_carryin_alt2_<optab>): Likewise.
5927 (*addsi3_carryin_clobercc_<optab>): Likewise.
5928 (*subsi3_carryin): Likewise.
5929 (*subsi3_carryin_const): Likewise.
5930 (*subsi3_carryin_compare): Likewise.
5931 (*subsi3_carryin_compare_const): Likewise.
5932 (*arm_subdi3): Likewise.
5933 (*thumb_subdi3): Likewise.
5934 (*subdi_di_zesidi): Likewise.
5935 (*subdi_di_sesidi): Likewise.
5936 (*subdi_zesidi_di): Likewise.
5937 (*subdi_sesidi_di): Likewise.
5938 (*subdi_zesidi_ze): Likewise.
5939 (thumb1_subsi3_insn): Likewise.
5940 (*arm_subsi3_insn): Likewise.
5941 (*anddi3_insn): Likewise.
5942 (*anddi_zesidi_di): Likewise.
5943 (*anddi_sesdi_di): Likewise.
5944 (*ne_zeroextracts): Likewise.
5945 (*ne_zeroextracts): Likewise.
5946 (*ite_ne_zeroextr): Likewise.
5947 (*ite_ne_zeroextr): Likewise.
5948 (*anddi_notdi_di): Likewise.
5949 (*anddi_notzesidi): Likewise.
5950 (*anddi_notsesidi): Likewise.
5951 (andsi_notsi_si): Likewise.
5952 (thumb1_bicsi3): Likewise.
5953 (*iordi3_insn): Likewise.
5954 (*iordi_zesidi_di): Likewise.
5955 (*iordi_sesidi_di): Likewise.
5956 (*thumb1_iorsi3_insn): Likewise.
5957 (*xordi3_insn): Likewise.
5958 (*xordi_zesidi_di): Likewise.
5959 (*xordi_sesidi_di): Likewise.
5960 (*arm_xorsi3): Likewise.
5961 (*andsi_iorsi3_no): Likewise.
5962 (*smax_0): Likewise.
5963 (*smax_m1): Likewise.
5964 (*arm_smax_insn): Likewise.
5965 (*smin_0): Likewise.
5966 (*arm_smin_insn): Likewise.
5967 (*arm_umaxsi3): Likewise.
5968 (*arm_uminsi3): Likewise.
5969 (*minmax_arithsi): Likewise.
5970 (*minmax_arithsi_): Likewise.
5971 (*satsi_<SAT:code>): Likewise.
5972 (arm_ashldi3_1bit): Likewise.
5973 (arm_ashrdi3_1bit): Likewise.
5974 (arm_lshrdi3_1bit): Likewise.
5975 (*arm_negdi2): Likewise.
5976 (*thumb1_negdi2): Likewise.
5977 (*arm_negsi2): Likewise.
5978 (*thumb1_negsi2): Likewise.
5979 (*negdi_extendsid): Likewise.
5980 (*negdi_zero_extend): Likewise.
5981 (*arm_abssi2): Likewise.
5982 (*thumb1_abssi2): Likewise.
5983 (*arm_neg_abssi2): Likewise.
5984 (*thumb1_neg_abss): Likewise.
5985 (one_cmpldi2): Likewise.
5986 (extend<mode>di2): Likewise.
5987 (*compareqi_eq0): Likewise.
5988 (*arm_extendhisi2addsi): Likewise.
5989 (*arm_movdi): Likewise.
5990 (*thumb1_movdi_insn): Likewise.
5991 (*arm_movt): Likewise.
5992 (*thumb1_movsi_insn): Likewise.
5993 (pic_add_dot_plus_four): Likewise.
5994 (pic_add_dot_plus_eight): Likewise.
5995 (tls_load_dot_plus_eight): Likewise.
5996 (*thumb1_movhi_insn): Likewise.
5997 (*thumb1_movsf_insn): Likewise.
5998 (*movdf_soft_insn): Likewise.
5999 (*thumb_movdf_insn): Likewise.
6000 (cbranchsi4_insn): Likewise.
6001 (cbranchsi4_scratch): Likewise.
6002 (*negated_cbranchsi4): Likewise.
6003 (*tbit_cbranch): Likewise.
6004 (*tlobits_cbranch): Likewise.
6005 (*tstsi3_cbranch): Likewise.
6006 (*cbranchne_decr1): Likewise.
6007 (*addsi3_cbranch): Likewise.
6008 (*addsi3_cbranch_scratch): Likewise.
6009 (*arm_cmpdi_insn): Likewise.
6010 (*arm_cmpdi_unsig): Likewise.
6011 (*arm_cmpdi_zero): Likewise.
6012 (*thumb_cmpdi_zero): Likewise.
6013 (*deleted_compare): Likewise.
6014 (*mov_scc): Likewise.
6015 (*mov_negscc): Likewise.
6016 (*mov_notscc): Likewise.
6017 (*cstoresi_eq0_thumb1_insn): Likewise.
6018 (cstoresi_nltu_thumb1): Likewise.
6019 (cstoresi_ltu_thu): Likewise.
6020 (thumb1_addsi3_addgeu): Likewise.
6021 (*arm_jump): Likewise.
6022 (*thumb_jump): Likewise.
6023 (*check_arch2): Likewise.
6024 (arm_casesi_internal): Likewise.
6025 (thumb1_casesi_dispatch): Likewise.
6026 (*arm_indirect_jump): Likewise.
6027 (*thumb1_indirect_jump): Likewise.
6029 (*and_scc): Likewise.
6030 (*ior_scc): Likewise.
6031 (*compare_scc): Likewise.
6032 (*cond_move): Likewise.
6033 (*cond_arith): Likewise.
6034 (*cond_sub): Likewise.
6035 (*cmp_ite0): Likewise.
6036 (*cmp_ite1): Likewise.
6037 (*cmp_and): Likewise.
6038 (*cmp_ior): Likewise.
6039 (*ior_scc_scc): Likewise.
6040 (*ior_scc_scc_cmp): Likewise.
6041 (*and_scc_scc): Likewise.
6042 (*and_scc_scc_cmp): Likewise.
6043 (*and_scc_scc_nod): Likewise.
6044 (*negscc): Likewise.
6045 (movcond_addsi): Likewise.
6046 (movcond): Likewise.
6047 (*ifcompare_plus_move): Likewise.
6048 (*if_plus_move): Likewise.
6049 (*ifcompare_move_plus): Likewise.
6050 (*if_move_plus): Likewise.
6051 (*ifcompare_arith_arith): Likewise.
6052 (*if_arith_arith): Likewise.
6053 (*ifcompare_arith_move): Likewise.
6054 (*if_arith_move): Likewise.
6055 (*ifcompare_move_arith): Likewise.
6056 (*if_move_arith): Likewise.
6057 (*ifcompare_move_not): Likewise.
6058 (*if_move_not): Likewise.
6059 (*ifcompare_not_move): Likewise.
6060 (*if_not_move): Likewise.
6061 (*ifcompare_shift_move): Likewise.
6062 (*if_shift_move): Likewise.
6063 (*ifcompare_move_shift): Likewise.
6064 (*if_move_shift): Likewise.
6065 (*ifcompare_shift_shift): Likewise.
6066 (*ifcompare_not_arith): Likewise.
6067 (*ifcompare_arith_not): Likewise.
6068 (*if_arith_not): Likewise.
6069 (*ifcompare_neg_move): Likewise.
6070 (*if_neg_move): Likewise.
6071 (*ifcompare_move_neg): Likewise.
6072 (*if_move_neg): Likewise.
6073 (prologue_thumb1_interwork): Likewise.
6074 (*cond_move_not): Likewise.
6075 (*sign_extract_onebit): Likewise.
6076 (*not_signextract_onebit): Likewise.
6077 (stack_tie): Likewise.
6078 (align_4): Likewise.
6079 (align_8): Likewise.
6080 (consttable_end): Likewise.
6081 (consttable_1): Likewise.
6082 (consttable_2): Likewise.
6083 (consttable_4): Likewise.
6084 (consttable_8): Likewise.
6085 (consttable_16): Likewise.
6086 (*thumb1_tablejump): Likewise.
6087 (prefetch): Likewise.
6088 (force_register_use): Likewise.
6089 (thumb_eh_return): Likewise.
6090 (load_tp_hard): Likewise.
6091 (load_tp_soft): Likewise.
6092 (tlscall): Likewise.
6093 (*arm_movtas_ze): Likewise.
6094 (*arm_rev): Likewise.
6095 (*arm_revsh): Likewise.
6096 (*arm_rev16): Likewise.
6097 * config/arm/thumb2.md
6098 (*thumb2_smaxsi3): Likewise.
6099 (*thumb2_sminsi3): Likewise.
6100 (*thumb32_umaxsi3): Likewise.
6101 (*thumb2_uminsi3): Likewise.
6102 (*thumb2_negdi2): Likewise.
6103 (*thumb2_abssi2): Likewise.
6104 (*thumb2_neg_abss): Likewise.
6105 (*thumb2_movsi_insn): Likewise.
6106 (tls_load_dot_plus_four): Likewise.
6107 (*thumb2_movhi_insn): Likewise.
6108 (*thumb2_mov_scc): Likewise.
6109 (*thumb2_mov_negs): Likewise.
6110 (*thumb2_mov_negs): Likewise.
6111 (*thumb2_mov_nots): Likewise.
6112 (*thumb2_mov_nots): Likewise.
6113 (*thumb2_movsicc_): Likewise.
6114 (*thumb2_movsfcc_soft_insn): Likewise.
6115 (*thumb2_indirect_jump): Likewise.
6116 (*thumb2_and_scc): Likewise.
6117 (*thumb2_ior_scc): Likewise.
6118 (*thumb2_ior_scc_strict_it): Likewise.
6119 (*thumb2_cond_move): Likewise.
6120 (*thumb2_cond_arith): Likewise.
6121 (*thumb2_cond_ari): Likewise.
6122 (*thumb2_cond_sub): Likewise.
6123 (*thumb2_negscc): Likewise.
6124 (*thumb2_movcond): Likewise.
6125 (thumb2_casesi_internal): Likewise.
6126 (thumb2_casesi_internal_pic): Likewise.
6127 (*thumb2_alusi3_short): Likewise.
6128 (*thumb2_mov<mode>_shortim): Likewise.
6129 (*thumb2_addsi_short): Likewise.
6130 (*thumb2_subsi_short): Likewise.
6131 (thumb2_addsi3_compare0): Likewise.
6132 (*thumb2_cbz): Likewise.
6133 (*thumb2_cbnz): Likewise.
6134 (*thumb2_one_cmplsi2_short): Likewise.
6135 (*thumb2_negsi2_short): Likewise.
6136 (*orsi_notsi_si): Likewise.
6137 * config/arm/arm1020e.md: Update with new attributes.
6138 * config/arm/arm1026ejs.md: Update with new attributes.
6139 * config/arm/arm1136jfs.md: Update with new attributes.
6140 * config/arm/arm926ejs.md: Update with new attributes.
6141 * config/arm/cortex-a15.md: Update with new attributes.
6142 * config/arm/cortex-a5.md: Update with new attributes.
6143 * config/arm/cortex-a53.md: Update with new attributes.
6144 * config/arm/cortex-a7.md: Update with new attributes.
6145 * config/arm/cortex-a8.md: Update with new attributes.
6146 * config/arm/cortex-a9.md: Update with new attributes.
6147 * config/arm/cortex-m4.md: Update with new attributes.
6148 * config/arm/cortex-r4.md: Update with new attributes.
6149 * config/arm/fa526.md: Update with new attributes.
6150 * config/arm/fa606te.md: Update with new attributes.
6151 * config/arm/fa626te.md: Update with new attributes.
6152 * config/arm/fa726te.md: Update with new attributes.
6154 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6156 * config/aarch64/aarch64-simd.md
6157 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6158 <vwx> iterator to ensure correct register choice.
6159 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6160 (aarch64_sqdmull_n<mode>): Likewise.
6161 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6162 * config/aarch64/arm_neon.h
6163 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6164 (vml<as><q>_n_<su>16): Likewise.
6165 (vml<as>l_high_lane<q>_<su>16): Likewise.
6166 (vml<as>l_high_n_<su>16): Likewise.
6167 (vml<as>l_lane<q>_<su>16): Likewise.
6168 (vml<as>l_n_<su>16): Likewise.
6169 (vmul<q>_lane<q>_<su>16): Likewise.
6170 (vmul<q>_n_<su>16): Likewise.
6171 (vmull_lane<q>_<su>16): Likewise.
6172 (vmull_n_<su>16): Likewise.
6173 (vmull_high_lane<q>_<su>16): Likewise.
6174 (vmull_high_n_<su>16): Likewise.
6175 (vqrdmulh<q>_n_s16): Likewise.
6177 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6179 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6180 have the correct lane parameter.
6182 2013-09-06 Richard Biener <rguenther@suse.de>
6184 * cfganal.c (control_dependences::~control_dependences):
6185 Properly free all of the vector.
6187 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6190 * config/i386/i386.c (ix86_conditional_register_usage):
6191 Proper initialize extended SSE registers.
6193 2013-09-06 Jan Hubicka <jh@suse.cz>
6195 PR tree-optimization/58311
6196 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6198 2013-09-06 Jan Hubicka <jh@suse.cz>
6200 * Makefile.in (tree-sra.o): Update dependencies.
6201 * tree-sra.c: Include ipa-utils.h
6202 (scan_function): Use recursive_call_p.
6203 (has_caller_p): New function.
6204 (cgraph_for_node_and_aliases): Count also callers of aliases.
6206 2013-09-06 Jan Hubicka <jh@suse.cz>
6209 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6210 * tree-tailcall.c: Include ipa-utils.h.
6211 (find_tail_calls): Use it.
6212 * ipa-pure-const.c (check_call): Likewise.
6213 * ipa-utils.c (recursive_call_p): New function.
6214 * ipa-utils.h (recursive_call_p): Dclare.
6215 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6216 (symtab_semantically_equivalent_p): New function.
6217 * Makefile.in (tree-tailcall.o): Update dependencies.
6219 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6221 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6224 2013-09-06 Richard Biener <rguenther@suse.de>
6226 * lto-streamer.h (lto_global_var_decls): Remove.
6227 * Makefile.in (OBJS): Remove lto-symtab.o.
6228 (lto-symtab.o): Remove.
6229 (GTFILES): Remove lto-symtab.c
6230 * lto-symtab.c: Move to lto/
6232 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6234 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6235 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6236 (UNSPEC_FPINT_RINT): New constant definitions.
6237 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6238 definition with 2 attributes.
6239 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6240 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6243 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6245 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6246 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6247 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6248 attribute to "z196_alone".
6249 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6252 2013-09-06 Richard Biener <rguenther@suse.de>
6254 * basic-block.h (class control_dependences): New.
6255 * tree-ssa-dce.c (control_dependence_map): Remove.
6257 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6258 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6259 find_pdom, find_control_dependence, find_all_control_dependences):
6261 (mark_control_dependent_edges_necessary,
6262 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6263 tree_dce_done, perform_tree_ssa_dce): Adjust.
6264 * cfganal.c (set_control_dependence_map_bit,
6265 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6266 find_all_control_dependences): Move from tree-ssa-dce.c and
6267 implement as methods of control_dependences class.
6268 (control_dependences::control_dependences): New.
6269 (control_dependences::~control_dependences): Likewise.
6270 (control_dependences::get_edges_dependent_on): Likewise.
6271 (control_dependences::get_edge): Likewise.
6273 2013-09-04 Jan Hubicka <jh@suse.cz>
6275 * tree.c (types_same_for_odr): Drop overactive check.
6276 * ipa-devirt.c (hash_type_name): Likewise.
6278 2013-09-04 Jan Hubicka <jh@suse.cz>
6280 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6281 (analyze_functions): ... here.
6283 2013-09-04 Jan Hubicka <jh@suse.cz>
6286 * cgraphunit.c (analyze_functions): Clear AUX fields
6287 after processing; initialize assembler name has.
6289 2013-09-05 Jeff Law <law@redhat.com>
6291 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6292 from thread_around_empty_block. Record threading path into PATH.
6293 Recurse if threading through the initial block is successful.
6294 (thread_across_edge): Corresponding changes to slightly simplify.
6296 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6298 * config/aarch64/aarch64.md
6299 (type): Remove frecpe, frecps, frecpx.
6300 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6301 fix to be a TARGET_SIMD instruction.
6302 (aarch64_frecps): Remove.
6303 * config/aarch64/aarch64-simd.md
6304 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6305 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6307 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6308 Sofiane Naci <sofiane.naci@arm.com>
6310 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6311 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6312 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6313 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6314 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6315 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6316 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6317 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6318 into "clz, "rbit". Rename "shift" to "shift_imm".
6319 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6320 changes. Update for attribute changes all occurrences of arlo_* and
6322 * config/arm/arm-fixed.md: Update for attribute changes
6323 all occurrences of arlo_* types.
6324 * config/arm/thumb2.md: Update for attribute changes all occurrences
6326 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6327 (cortexa7_older_only): Likewise.
6328 (cortexa7_younger): Likewise.
6329 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6330 (1020alu_shift_op): Likewise.
6331 (1020alu_shift_reg_op): Likewise.
6332 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6333 (alu_shift_op): Likewise.
6334 (alu_shift_reg_op): Likewise.
6335 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6336 (11_alu_shift_op): Likewise.
6337 (11_alu_shift_reg_op): Likewise.
6338 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6339 (9_alu_shift_reg_op): Likewise.
6340 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6342 (cortex_a15_alu_shift): Likewise.
6343 (cortex_a15_alu_shift_reg): Likewise.
6344 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6346 (cortex_a5_alu_shift): Likewise.
6347 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6349 (cortex_a53_alu_shift): Likewise.
6350 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6352 (cortex_a7_alu_reg): Likewise.
6353 (cortex_a7_alu_shift): Likewise.
6354 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6356 (cortex_a8_alu_shift): Likewise.
6357 (cortex_a8_alu_shift_reg): Likewise.
6358 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6359 (cortex_a9_dp_shift): Likewise.
6360 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6362 * config/arm/cortex-r4.md
6363 (cortex_r4_alu): Update for attribute changes.
6364 (cortex_r4_mov): Likewise.
6365 (cortex_r4_alu_shift_reg): Likewise.
6366 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6367 (526_alu_shift_op): Likewise.
6368 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6369 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6370 (626te_alu_shift_op): Likewise.
6371 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6372 (726te_alu_shift_op): Likewise.
6373 (726te_alu_shift_reg_op): Likewise.
6374 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6375 (mp626_alu_shift_op): Likewise.
6376 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6377 (pj4_alu_conds): Likewise.
6378 (pj4_shift): Likewise.
6379 (pj4_shift_conds): Likewise.
6380 (pj4_alu_shift): Likewise.
6381 (pj4_alu_shift_conds): Likewise.
6382 * config/aarch64/aarch64.md: Update for attribute change
6383 all occurrences of arlo_* and shift* types.
6385 2013-09-05 Mike Stump <mikestump@comcast.net>
6387 * tree.h: Move documentation for tree_function_decl to tree-core.h
6388 with the declaration.
6390 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6393 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6394 looking for widest mode.
6396 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6398 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6400 2013-09-05 Richard Biener <rguenther@suse.de>
6402 PR tree-optimization/58137
6403 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6404 Do not create vectors of pointers.
6405 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6406 types for the components of the vector initializer.
6407 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6408 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6410 2013-09-05 Martin Jambor <mjambor@suse.cz>
6412 * ipa-prop.c (remove_described_reference): Accept missing references,
6413 return false if that hppens, otherwise return true.
6414 (cgraph_node_for_jfunc): New function.
6415 (try_decrement_rdesc_refcount): Likewise.
6416 (try_make_edge_direct_simple_call): Use them.
6417 (ipa_edge_removal_hook): Remove references from rdescs.
6418 (ipa_edge_duplication_hook): Clone rdescs and their references
6419 when the new edge has the same caller as the old one.
6420 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6421 reference before removing any edges.
6423 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6425 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6427 * thumb2.md (thumb2_storewb_parisi): New pattern.
6429 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6431 * config/aarch64/aarch64-option-extensions.def: Add
6432 AARCH64_OPT_EXTENSION of 'crc'.
6433 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6434 (AARCH64_ISA_CRC): Ditto.
6435 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6436 description of the CRC extension.
6438 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6440 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6441 * config/rs6000/linux.h: Ditto.
6442 * alpha/linux.h: Ditto.
6443 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6444 no_c99_libc_has_function.
6445 * config/c6x/uclinux-elf.h: Ditto.
6446 * config/lm32/uclinux-elf.h: Ditto.
6447 * config/m68k/uclinux.h: Ditto.
6448 * config/moxie/uclinux.h: Ditto.
6449 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6450 (crisv32-*-linux*, cris-*-linux*): Ditto.
6451 * config/bfin/bfin.c: Include "tm_p.h".
6453 2013-09-05 Richard Biener <rguenther@suse.de>
6455 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6456 check for a definition without a basic-block.
6458 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6459 Sofiane Naci <sofiane.naci@arm.com>
6461 * config/aarch64/aarch64.md
6462 (*movti_aarch64): Rename r_2_f and f_2_r.
6463 (*movsf_aarch64): Likewise.
6464 (*movdf_aarch64): Likewise.
6465 (*movtf_aarch64): Likewise.
6466 (aarch64_movdi_<mode>low): Likewise.
6467 (aarch64_movdi_<mode>high): Likewise.
6468 (aarch64_mov<mode>high_di): Likewise.
6469 (aarch64_mov<mode>low_di): Likewise.
6470 (aarch64_movtilow_tilow): Likewise.
6471 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6472 values to config/arm/types.md
6473 (attribute "conds"): Update for attribute change.
6474 (anddi3_insn): Likewise.
6475 (iordi3_insn): Likewise.
6476 (xordi3_insn): Likewise.
6477 (one_cmpldi2): Likewise.
6478 * config/arm/types.md (type): Add Neon types.
6479 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6480 use "type" attribute.
6481 (movmisalign<mode>_neon_store): Likewise.
6482 (movmisalign<mode>_neon_load): Likewise.
6483 (vec_set<mode>_internal): Likewise.
6484 (vec_setv2di_internal): Likewise.
6485 (vec_extract<mode>): Likewise.
6486 (vec_extractv2di): Likewise.
6487 (add<mode>3_neon): Likewise.
6488 (adddi3_neon): Likewise.
6489 (sub<mode>3_neon): Likewise.
6490 (subdi3_neon): Likewise.
6491 (mul<mode>3_neon): Likewise.
6492 (mul<mode>3add<mode>_neon): Likewise.
6493 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6494 (fma<VCVTF:mode>4)): Likewise.
6495 (fma<VCVTF:mode>4_intrinsic): Likewise.
6496 (fmsub<VCVTF:mode>4)): Likewise.
6497 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6498 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6499 (ior<mode>3): Likewise.
6500 (and<mode>3): Likewise.
6501 (anddi3_neon): Likewise.
6502 (orn<mode>3_neon): Likewise.
6503 (orndi3_neon): Likewise.
6504 (bic<mode>3_neon): Likewise.
6505 (bicdi3_neon): Likewise.
6506 (xor<mode>3): Likewise.
6507 (one_cmpl<mode>2): Likewise.
6508 (abs<mode>2): Likewise.
6509 (neg<mode>2): Likewise.
6510 (umin<mode>3_neon): Likewise.
6511 (umax<mode>3_neon): Likewise.
6512 (smin<mode>3_neon): Likewise.
6513 (smax<mode>3_neon): Likewise.
6514 (vashl<mode>3): Likewise.
6515 (vashr<mode>3_imm): Likewise.
6516 (vlshr<mode>3_imm): Likewise.
6517 (ashl<mode>3_signed): Likewise.
6518 (ashl<mode>3_unsigned): Likewise.
6519 (neon_load_count): Likewise.
6520 (ashldi3_neon_noclobber): Likewise.
6521 (signed_shift_di3_neon): Likewise.
6522 (unsigned_shift_di3_neon): Likewise.
6523 (ashrdi3_neon_imm_noclobber): Likewise.
6524 (lshrdi3_neon_imm_noclobber): Likewise.
6525 (widen_ssum<mode>3): Likewise.
6526 (widen_usum<mode>3): Likewise.
6527 (quad_halves_<code>v4si): Likewise.
6528 (quad_halves_<code>v4sf): Likewise.
6529 (quad_halves_<code>v8hi): Likewise.
6530 (quad_halves_<code>v16qi): Likewise.
6531 (reduc_splus_v2di): Likewise.
6532 (neon_vpadd_internal<mode>): Likewise.
6533 (neon_vpsmin<mode>): Likewise.
6534 (neon_vpsmax<mode>): Likewise.
6535 (neon_vpumin<mode>): Likewise.
6536 (neon_vpumax<mode>): Likewise.
6537 (ss_add<mode>_neon): Likewise.
6538 (us_add<mode>_neon): Likewise.
6539 (ss_sub<mode>_neon): Likewise.
6540 (us_sub<mode>_neon): Likewise.
6541 (neon_vadd<mode>_unspec): Likewise.
6542 (neon_vaddl<mode>): Likewise.
6543 (neon_vaddw<mode>): Likewise.
6544 (neon_vhadd<mode>): Likewise.
6545 (neon_vqadd<mode>): Likewise.
6546 (neon_vaddhn<mode>): Likewise.
6547 (neon_vmul<mode>): Likewise.
6548 (neon_vmla<mode>): Likewise.
6549 (neon_vmlal<mode>): Likewise.
6550 (neon_vmls<mode>): Likewise.
6551 (neon_vmlsl<mode>): Likewise.
6552 (neon_vqdmulh<mode>): Likewise.
6553 (neon_vqdmlal<mode>): Likewise.
6554 (neon_vqdmlsl<mode>): Likewise.
6555 (neon_vmull<mode>): Likewise.
6556 (neon_vqdmull<mode>): Likewise.
6557 (neon_vsub<mode>_unspec): Likewise.
6558 (neon_vsubl<mode>): Likewise.
6559 (neon_vsubw<mode>): Likewise.
6560 (neon_vqsub<mode>): Likewise.
6561 (neon_vhsub<mode>): Likewise.
6562 (neon_vsubhn<mode>): Likewise.
6563 (neon_vceq<mode>): Likewise.
6564 (neon_vcge<mode>): Likewise.
6565 (neon_vcgeu<mode>): Likewise.
6566 (neon_vcgt<mode>): Likewise.
6567 (neon_vcgtu<mode>): Likewise.
6568 (neon_vcle<mode>): Likewise.
6569 (neon_vclt<mode>): Likewise.
6570 (neon_vcage<mode>): Likewise.
6571 (neon_vcagt<mode>): Likewise.
6572 (neon_vtst<mode>): Likewise.
6573 (neon_vabd<mode>): Likewise.
6574 (neon_vabdl<mode>): Likewise.
6575 (neon_vaba<mode>): Likewise.
6576 (neon_vabal<mode>): Likewise.
6577 (neon_vmax<mode>): Likewise.
6578 (neon_vmin<mode>): Likewise.
6579 (neon_vpaddl<mode>): Likewise.
6580 (neon_vpadal<mode>): Likewise.
6581 (neon_vpmax<mode>): Likewise.
6582 (neon_vpmin<mode>): Likewise.
6583 (neon_vrecps<mode>): Likewise.
6584 (neon_vrsqrts<mode>): Likewise.
6585 (neon_vqabs<mode>): Likewise.
6586 (neon_vqneg<mode>): Likewise.
6587 (neon_vcls<mode>): Likewise.
6588 (clz<mode>2): Likewise.
6589 (popcount<mode>2): Likewise.
6590 (neon_vrecpe): Likewise.
6591 (neon_vrsqrte): Likewise.
6592 (neon_vget_lane<mode>_sext_internal): Likewise.
6593 (neon_vget_lane<mode>_zext_internal): Likewise.
6594 (neon_vdup_n<mode>): Likewise.
6595 (neon_vdup_nv2di): Likewise.
6596 (neon_vdpu_lane<mode>_internal): Likewise.
6597 (neon_vswp<mode>): Likewise.
6598 (float<mode><V_cvtto>2): Likewise.
6599 (floatuns<mode><V_cvtto>2): Likewise.
6600 (fix_trunc<mode><V_cvtto>)2): Likewise
6601 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6602 (neon_vcvt<mode>): Likewise.
6603 (neon_vcvtv4sfv4hf): Likewise.
6604 (neon_vcvtv4hfv4sf): Likewise.
6605 (neon_vcvt_n<mode>): Likewise.
6606 (neon_vmovn<mode>): Likewise.
6607 (neon_vqmovn<mode>): Likewise.
6608 (neon_vqmovun<mode>): Likewise.
6609 (neon_vmovl<mode>): Likewise.
6610 (neon_vmul_lane<mode>): Likewise.
6611 (neon_vmull_lane<mode>): Likewise.
6612 (neon_vqdmull_lane<mode>): Likewise.
6613 (neon_vqdmulh_lane<mode>): Likewise.
6614 (neon_vmla_lane<mode>): Likewise.
6615 (neon_vmlal_lane<mode>): Likewise.
6616 (neon_vqdmlal_lane<mode>): Likewise.
6617 (neon_vmls_lane<mode>): Likewise.
6618 (neon_vmlsl_lane<mode>): Likewise.
6619 (neon_vqdmlsl_lane<mode>): Likewise.
6620 (neon_vext<mode>): Likewise.
6621 (neon_vrev64<mode>): Likewise.
6622 (neon_vrev32<mode>): Likewise.
6623 (neon_vrev16<mode>): Likewise.
6624 (neon_vbsl<mode>_internal): Likewise.
6625 (neon_vshl<mode>): Likewise.
6626 (neon_vqshl<mode>): Likewise.
6627 (neon_vshr_n<mode>): Likewise.
6628 (neon_vshrn_n<mode>): Likewise.
6629 (neon_vqshrn_n<mode>): Likewise.
6630 (neon_vqshrun_n<mode>): Likewise.
6631 (neon_vshl_n<mode>): Likewise.
6632 (neon_vqshl_n<mode>): Likewise.
6633 (neon_vqshlu_n<mode>): Likewise.
6634 (neon_vshll_n<mode>): Likewise.
6635 (neon_vsra_n<mode>): Likewise.
6636 (neon_vsri_n<mode>): Likewise.
6637 (neon_vsli_n<mode>): Likewise.
6638 (neon_vtbl1v8qi): Likewise.
6639 (neon_vtbl2v8qi): Likewise.
6640 (neon_vtbl3v8qi): Likewise.
6641 (neon_vtbl4v8qi): Likewise.
6642 (neon_vtbx1v8qi): Likewise.
6643 (neon_vtbx2v8qi): Likewise.
6644 (neon_vtbx3v8qi): Likewise.
6645 (neon_vtbx4v8qi): Likewise.
6646 (neon_vtrn<mode>_internal): Likewise.
6647 (neon_vzip<mode>_internal): Likewise.
6648 (neon_vuzp<mode>_internal): Likewise.
6649 (neon_vld1<mode>): Likewise.
6650 (neon_vld1_lane<mode>): Likewise.
6651 (neon_vld1_dup<mode>): Likewise.
6652 (neon_vld1_dupv2di): Likewise.
6653 (neon_vst1<mode>): Likewise.
6654 (neon_vst1_lane<mode>): Likewise.
6655 (neon_vld2<mode>): Likewise.
6656 (neon_vld2_lane<mode>): Likewise.
6657 (neon_vld2_dup<mode>): Likewise.
6658 (neon_vst2<mode>): Likewise.
6659 (neon_vst2_lane<mode>): Likewise.
6660 (neon_vld3<mode>): Likewise.
6661 (neon_vld3qa<mode>): Likewise.
6662 (neon_vld3qb<mode>): Likewise.
6663 (neon_vld3_lane<mode>): Likewise.
6664 (neon_vld3_dup<mode>): Likewise.
6665 (neon_vst3<mode>): Likewise.
6666 (neon_vst3qa<mode>): Likewise.
6667 (neon_vst3qb<mode>): Likewise.
6668 (neon_vst3_lane<mode>): Likewise.
6669 (neon_vld4<mode>): Likewise.
6670 (neon_vld4qa<mode>): Likewise.
6671 (neon_vld4qb<mode>): Likewise.
6672 (neon_vld4_lane<mode>): Likewise.
6673 (neon_vld4_dup<mode>): Likewise.
6674 (neon_vst4<mode>): Likewise.
6675 (neon_vst4qa<mode>): Likewise.
6676 (neon_vst4qb<mode>): Likewise.
6677 (neon_vst4_lane<mode>): Likewise.
6678 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6679 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6680 (neon_vec_<US>mult_lo_<mode>): Likewise.
6681 (neon_vec_<US>mult_hi_<mode>): Likewise.
6682 (neon_vec_<US>shiftl_<mode>): Likewise.
6683 (neon_unpack<US>_<mode>): Likewise.
6684 (neon_vec_<US>mult_<mode>): Likewise.
6685 (vec_pack_trunc_<mode>): Likewise.
6686 (neon_vec_pack_trunk_<mode>): Likewise.
6687 (neon_vabd<mode>_2): Likewise.
6688 (neon_vabd<mode>_3): Likewise.
6689 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6690 (thumb2_movsi_vfp): Likewise.
6691 (movdi_vfp): Likewise.
6692 (movdi_vfp_cortexa8): Likewise.
6693 (movhf_vfp_neon): Likewise.
6694 (movhf_vfp): Likewiwse.
6695 (movsf_vfp): Likewiwse.
6696 (thumb2_movsf_vfp): Likewiwse.
6697 (movdf_vfp): Likewise.
6698 (thumb2_movdf_vfp): Likewise.
6699 (movsfcc_vfp): Likewise.
6700 (thumb2_movsfcc_vfp): Likewise.
6701 (movdfcc_vfp): Likewise.
6702 (thumb2_movdfcc_vfp): Likewise.
6703 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6704 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6705 (v10_v2c): Likewise.
6706 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6708 (cortex_a15_neon_int_2): Likewise.
6709 (cortex_a15_neon_int_3): Likewise.
6710 (cortex_a15_neon_int_4): Likewise.
6711 (cortex_a15_neon_int_5): Likewise.
6712 (cortex_a15_neon_vqneg_vqabs): Likewise.
6713 (cortex_a15_neon_vmov): Likewise.
6714 (cortex_a15_neon_vaba): Likewise.
6715 (cortex_a15_neon_vaba_qqq): Likewise.
6716 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6717 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6718 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6719 scalar_64_32_long_scalar): Likewise.
6720 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6721 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6722 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6723 lotype_qdd_64_32_long): Likewise.
6724 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6725 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6726 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6727 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6728 (cortex_a15_neon_shift_1): Likewise.
6729 (cortex_a15_neon_shift_2): Likewise.
6730 (cortex_a15_neon_shift_3): Likewise.
6731 (cortex_a15_neon_vshl_ddd): Likewise.
6732 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6733 (cortex_a15_neon_vsra_vrsra): Likewise.
6734 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6735 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6736 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6737 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6738 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6739 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6740 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6741 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6742 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6743 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6744 (cortex_a15_neon_bp_simple): Likewise.
6745 (cortex_a15_neon_bp_2cycle): Likewise.
6746 (cortex_a15_neon_bp_3cycle): Likewise.
6747 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6748 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6749 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6750 (cortex_a15_neon_vld2_4_regs): Likewise.
6751 (cortex_a15_neon_vld3_vld4): Likewise.
6752 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6753 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6754 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6755 (cortex_a15_neon_vst3_vst4): Likewise.
6756 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6757 (cortex_a15_neon_vld3_vld4_lane" 10
6758 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6759 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6760 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6761 (cortex_a15_neon_ldm_2): Likewise.0
6762 (cortex_a15_neon_stm_2): Likewise.
6763 (cortex_a15_neon_mcr): Likewise.
6764 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6765 (cortex_a15_neon_mrc): Likewise.
6766 (cortex_a15_neon_mrrc): Likewise.
6767 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6769 (cortex_a15_alu_shift): Likewise.
6770 (cortex_a15_alu_shift_reg): Likewise.
6771 (cortex_a15_mult32): Likewise.
6772 (cortex_a15_mult64): Likewise.
6773 (cortex_a15_block): Likewise.
6774 (cortex_a15_branch): Likewise.
6775 (cortex_a15_load1): Likewise.
6776 (cortex_a15_load3): Likewise.
6777 (cortex_a15_store1): Likewise.
6778 (cortex_a15_store3): Likewise.
6779 (cortex_a15_call): Likewise.
6780 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6781 (cortex_a5_f2r): Likewise.
6782 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6784 (cortex_a53_f2r): Likewise.
6785 * config/arm/cortex-a7.md
6786 (cortex_a7_branch): Update for attribute change.
6787 (cortex_a7_call): Likewise.
6788 (cortex_a7_alu_imm): Likewise.
6789 (cortex_a7_alu_reg): Likewise.
6790 (cortex_a7_alu_shift): Likewise.
6791 (cortex_a7_mul): Likewise.
6792 (cortex_a7_load1): Likewise.
6793 (cortex_a7_store1): Likewise.
6794 (cortex_a7_load2): Likewise.
6795 (cortex_a7_store2): Likewise.
6796 (cortex_a7_load3): Likewise.
6797 (cortex_a7_store3): Likewise.
6798 (cortex_a7_load4): Likewise.
6799 (cortex_a7_store4): Likewise.
6800 (cortex_a7_fpalu): Likewise.
6801 (cortex_a7_fconst): Likewise.
6802 (cortex_a7_fpmuls): Likewise.
6803 (cortex_a7_neon_mul): Likewise.
6804 (cortex_a7_fpmacs): Likewise.
6805 (cortex_a7_neon_mla: Likewise.
6806 (cortex_a7_fpmuld: Likewise.
6807 (cortex_a7_fpmacd: Likewise.
6808 (cortex_a7_fpfmad: Likewise.
6809 (cortex_a7_fdivs: Likewise.
6810 (cortex_a7_fdivd: Likewise.
6811 (cortex_a7_r2f: Likewise.
6812 (cortex_a7_f2r: Likewise.
6813 (cortex_a7_f_flags: Likewise.
6814 (cortex_a7_f_loads: Likewise.
6815 (cortex_a7_f_loadd: Likewise.
6816 (cortex_a7_f_stores: Likewise.
6817 (cortex_a7_f_stored: Likewise.
6818 (cortex_a7_neon): Likewise.
6819 * config/arm/cortex-a8-neon.md
6820 (cortex_a8_neon_mrc): Update for attribute change.
6821 (cortex_a8_neon_mrrc): Likewise.
6822 (cortex_a8_neon_int_1): Likewise.
6823 (cortex_a8_neon_int_2): Likewise.
6824 (cortex_a8_neon_int_3): Likewise.
6825 (cortex_a8_neon_int_4): Likewise.
6826 (cortex_a8_neon_int_5): Likewise.
6827 (cortex_a8_neon_vqneg_vqabs): Likewise.
6828 (cortex_a8_neon_vmov): Likewise.
6829 (cortex_a8_neon_vaba): Likewise.
6830 (cortex_a8_neon_vaba_qqq): Likewise.
6831 (cortex_a8_neon_vsma): Likewise.
6832 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6833 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6834 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6835 long_scalar): Likewise.
6836 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6837 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6838 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6839 long_scalar_qdd_64_32_long): Likewise.
6840 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6841 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6842 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6843 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6844 (cortex_a8_neon_shift_1): Likewise.
6845 (cortex_a8_neon_shift_2): Likewise.
6846 (cortex_a8_neon_shift_3): Likewise.
6847 (cortex_a8_neon_vshl_ddd): Likewise.
6848 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6849 (cortex_a8_neon_vsra_vrsra): Likewise.
6850 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6851 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6852 (cortex_a8_neon_fp_vsum): Likewise.
6853 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6854 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6855 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6856 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6857 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6858 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6859 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6860 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6861 (cortex_a8_neon_bp_simple): Likewise.
6862 (cortex_a8_neon_bp_2cycle): Likewise.
6863 (cortex_a8_neon_bp_3cycle): Likewise.
6864 (cortex_a8_neon_ldr): Likewise.
6865 (cortex_a8_neon_str): Likewise.
6866 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6867 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6868 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6869 (cortex_a8_neon_vld2_4_regs): Likewise.
6870 (cortex_a8_neon_vld3_vld4): Likewise.
6871 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6872 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6873 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6874 (cortex_a8_neon_vst3_vst4): Likewise.
6875 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6876 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6877 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6878 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6879 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6880 (cortex_a8_neon_mcr): Likewise.
6881 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6882 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6883 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6885 (ca9_neon_mrrc): Likewise.
6886 (cortex_a9_neon_int_1): Likewise.
6887 (cortex_a9_neon_int_2): Likewise.
6888 (cortex_a9_neon_int_3): Likewise.
6889 (cortex_a9_neon_int_4): Likewise.
6890 (cortex_a9_neon_int_5): Likewise.
6891 (cortex_a9_neon_vqneg_vqabs): Likewise.
6892 (cortex_a9_neon_vmov): Likewise.
6893 (cortex_a9_neon_vaba): Likewise.
6894 (cortex_a9_neon_vaba_qqq): Likewise.
6895 (cortex_a9_neon_vsma): Likewise.
6896 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6897 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6898 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6899 long_scalar): Likewise.
6900 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6901 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6902 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6903 long_scalar_qdd_64_32_long): Likewise.
6904 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6905 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6906 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6907 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6908 (cortex_a9_neon_shift_1): Likewise.
6909 (cortex_a9_neon_shift_2): Likewise.
6910 (cortex_a9_neon_shift_3): Likewise.
6911 (cortex_a9_neon_vshl_ddd): Likewise.
6912 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6913 (cortex_a9_neon_vsra_vrsra): Likewise.
6914 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6915 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6916 (cortex_a9_neon_fp_vsum): Likewise.
6917 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6918 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6919 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6920 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6921 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6922 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6923 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6924 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6925 (cortex_a9_neon_bp_simple): Likewise.
6926 (cortex_a9_neon_bp_2cycle): Likewise.
6927 (cortex_a9_neon_bp_3cycle): Likewise.
6928 (cortex_a9_neon_ldr): Likewise.
6929 (cortex_a9_neon_str): Likewise.
6930 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6931 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6932 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6933 (cortex_a9_neon_vld2_4_regs): Likewise.
6934 (cortex_a9_neon_vld3_vld4): Likewise.
6935 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6936 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6937 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6938 (cortex_a9_neon_vst3_vst4): Likewise.
6939 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6940 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6941 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6942 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6943 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6944 (cortex_a9_neon_mcr): Likewise.
6945 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6946 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6947 (cortex_a9_fps): Likewise.
6948 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6950 (cortex_m4_fmuls): Likewise.
6951 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6953 (cortex_r4_mrc): Likewise.
6954 * config/arm/iterators.md: Update comment referring to neon_type.
6955 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
6956 (iwmmxt_movsi_insn): Likewise.
6957 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6959 (pj4_core_to_vfp): Likewise.
6960 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6962 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6963 (vfp_fstore): Likewise.
6964 * doc/md.texi: Change references to neon_type to refer to type.
6966 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6968 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6970 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
6972 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6975 2013-09-04 Easwaran Raman <eraman@google.com>
6978 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6979 (build_and_add_sum): Use it.
6980 (appears_later_in_bb): Simplify code.
6982 2013-09-04 Teresa Johnson <tejohnson@google.com>
6984 * dumpfile.c (dump_finish): Don't close stderr/stdout.
6986 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
6988 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
6990 2013-09-04 Jan Hubicka <jh@suse.cz>
6992 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
6993 * ipa-devirt.c: Include diganostic.h
6994 (odr_type_d): Add types and types_set.
6995 (hash_type_name): Work for types with vtables during LTO.
6996 (odr_hasher::remove): Fix comment; destroy types_set.
6997 (add_type_duplicate): New function,
6998 (get_odr_type): Use it.
6999 (dump_type_inheritance_graph): Dump type duplicates.
7000 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7002 * tree.c (types_same_for_odr): Give exact answers on types with
7005 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7007 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7008 explaining their differences.
7010 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7012 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7014 2013-09-03 Jeff Law <law@redhat.com>
7016 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7017 when not threading through a joiner block. Pass joiner/no joiner
7018 state to register_jump_thread.
7019 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7020 state from argument rather than implying on path length.
7021 Dump the entire jump thread path into debugging dump.
7022 * tree-flow.h (register_jump_thread): Update prototype.
7024 2013-08-29 Xinliang David Li <davidxl@google.com>
7026 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7027 Remove a trivial gcc_assert.
7029 2013-08-29 Xinliang David Li <davidxl@google.com>
7031 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7032 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7033 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7034 Delay base decl alignment adjustment.
7035 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7036 * tree-vectorizer.h: New data structure.
7037 (set_dr_misalignment): New function.
7038 (dr_misalignment): Ditto.
7039 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7040 (vectorizable_load): Ditto.
7041 (ensure_base_align): New function.
7042 (vectorize_loops): Add dbg_cnt support.
7043 (execute_vect_slp): Ditto.
7044 * dbgcnt.def: New debug counter.
7045 * Makefile: New dependency.
7047 2013-09-03 Meador Inge <meadori@codesourcery.com>
7051 2013-08-30 Meador Inge <meadori@codesourcery.com>
7053 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7055 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7057 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7058 function descriptor.
7060 2013-09-03 Richard Biener <rguenther@suse.de>
7062 * tree-affine.c (add_elt_to_tree): Fix association issue,
7063 avoid useless converts and make sure to always return a
7064 properly typed result.
7066 2013-09-03 Richard Biener <rguenther@suse.de>
7069 * fold-const.c (negate_expr_p): Fix division case.
7070 (negate_expr): Likewise.
7072 2013-09-03 Richard Biener <rguenther@suse.de>
7075 * tree-streamer-out.c: Include tm.h.
7076 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7078 2013-09-03 Jan Hubicka <jh@suse.cz>
7080 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7082 2013-09-03 Alan Modra <amodra@gmail.com>
7084 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7085 multilibs to big-endian target and vice versa.
7086 * config/rs6000/t-linux64: Use := assignment on all vars.
7087 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7088 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7089 * config/rs6000/t-linux64le: New file.
7090 * config/rs6000/t-linux64bele: New file.
7091 * config/rs6000/t-linux64lebe: New file.
7093 2013-09-02 Jan Hubicka <jh@suse.cz>
7095 * ipa-inline-transform.c (inline_transform): Do not
7096 optimize_inline_calls when not optimizing.
7098 2013-09-02 Jan Hubicka <jh@suse.cz>
7100 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7101 duplicated nodes for assembler names.
7102 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7103 hard registers from assembler name hash.
7105 2013-09-02 Jan Hubicka <jh@suse.cz>
7107 * ipa-split.c (execute_split_functions): Split externally visible
7108 functions called once.
7110 2013-09-02 Martin Jambor <mjambor@suse.cz>
7113 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7114 linked list. When finding the correct duplicate, also consider also
7115 the caller in additon to its inlined_to node.
7117 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7119 * config/aarch64/aarch64-simd-builtins.def
7120 (dup_lane_scalar): Remove.
7121 * config/aarch64/aarch64-simd.md
7122 (aarch64_simd_dup): Add 'w->w' alternative.
7123 (aarch64_dup_lane<mode>): Allow for VALL.
7124 (aarch64_dup_lane_scalar<mode>): Remove.
7125 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7126 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7127 (aarch64_get_lane_unsigned<mode>): Likewise.
7128 (aarch64_get_lane<mode>): Likewise.
7129 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7130 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7131 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7132 (VCON): Change container of V2SF.
7133 (vswap_width_name): Likewise.
7134 * config/aarch64/arm_neon.h
7135 (__aarch64_vdup_lane_any): New.
7136 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7137 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7138 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7140 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7143 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7144 the source or the destination is a single hard register.
7146 2013-09-02 Richard Biener <rguenther@suse.de>
7149 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7152 2013-09-02 Richard Biener <rguenther@suse.de>
7154 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7155 arithmetic to sizetype.
7157 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7159 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7160 Find auto-increment use both before and after candidate.
7162 2013-09-02 Marek Polacek <polacek@redhat.com>
7164 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7166 2013-09-01 Jan Hubicka <jh@suse.cz>
7168 * Makefile.in: Add ipa-profile.o
7169 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7170 * cgraph.c (struct cgraph_propagate_frequency_data,
7171 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7172 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7173 * cgraph.h (cgraph_propagate_frequency): Remove.
7174 * ipa-inline-analysis.c: Include ipa-utils.h;
7175 drop duplicated cfgloop.h.
7176 (inline_update_callee_summaries): Update.
7177 * ipa-profile.c: New file.
7178 * ipa-utils.h (ipa_propagate_frequency): Declare.
7179 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7180 data-streamer.h, value-prof.h.
7181 (symtab_remove_unreachable_nodes): Update profile.
7182 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7183 account_time_size, cmp_counts, dump_histogram,
7184 ipa_profile_generate_summary, ipa_profile_write_summary,
7185 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7186 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7187 Move to ipa-profile.c.
7189 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7191 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7193 2013-09-01 Jan Hubicka <jh@suse.cz>
7195 * common.opt (fdevirtualize-speculatively): New function.
7196 * invoke.texi (fdevirtualize-speculatively): Document.
7197 * ipa-devirt.c: Include ipa-inline.h
7198 (likely_target_p): New function.
7199 (ipa_devirt): New function.
7200 (gate_ipa_devirt): New function.
7201 (pass_data_ipa_devirt): New static var.
7202 (pass_ipa_devirt): Likewise.
7203 (make_pass_ipa_devirt): New function.
7204 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7205 (common_handle_option): Disable devirtualization when
7206 value range profiling is available.
7207 * passes.def (pass_ipa_devirt): Add.
7208 * timever.def (TV_IPA_DEVIRT): New timevar.
7209 * tree-pass.h (make_pass_ipa_devirt):
7211 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7213 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7214 include sanitize(undefined).
7216 2013-08-31 Diego Novillo <dnovillo@google.com>
7218 * Makefile.in (TREE_CORE_H): Define.
7220 (GTFILES): Add tree-core.h.
7221 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7223 * tree-core.h: New file.
7224 Move all data structures, enum, typedefs, global
7225 declarations and constants from ...
7228 2013-08-31 Jan Hubicka <jh@suse.cz>
7230 * bulitins.c (expand_builtin): Do not early exit for gcov
7231 instrumented functions.
7233 2013-08-31 Marek Polacek <polacek@redhat.com>
7235 * ubsan.c: Include tm_p.h.
7237 2013-08-31 Jan Hubicka <jh@suse.cz>
7239 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7242 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7243 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7244 * tree-cfg.c (verify_gimple_label): ... here.
7245 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7246 (ipa_merge_profiles): New function.
7247 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7248 (lto_input_function_body): Likewise.
7249 * ipa-utils.h (ipa_merge_profiles): Declare.
7250 * lto-streamer.h (lto_input_function_body): Update prototype.
7251 (emit_label_in_global_context_p): Remove.
7252 * lto-symtab.c: Include ipa-utils.h
7253 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7255 2013-08-31 Jan Hubicka <jh@suse.cz>
7257 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7259 2013-08-31 Jan Hubicka <jh@suse.cz>
7261 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7263 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7265 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7266 "cmp" RTX before signed_comparison_operator check to account
7269 2013-08-30 Jan Hubicka <jh@suse.cz>
7271 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7272 (detect_type_change_1): Rename to ...
7273 (detect_type_change): ... this one; early return on non-polymorphic
7275 (detect_type_change_ssa): Add comp_type parameter; update
7276 use of detect_type_change.
7277 (compute_complex_assign_jump_func): Add param_type parameter;
7278 update use of detect_type_change_ssa.
7279 (compute_complex_ancestor_jump_func): Likewise.
7280 (ipa_get_callee_param_type): New function.
7281 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7282 update calls to the jump function computation functions.
7284 2013-08-30 Teresa Johnson <tejohnson@google.com>
7285 Steven Bosscher <steven@gcc.gnu.org>
7287 * cfgrtl.c (fixup_new_cold_bb): New routine.
7288 (commit_edge_insertions): Invoke fixup_partitions.
7289 (find_partition_fixes): New routine.
7290 (fixup_partitions): Ditto.
7291 (verify_hot_cold_block_grouping): Update comments.
7292 (rtl_verify_edges): Invoke find_partition_fixes.
7293 (rtl_verify_bb_pointers): Update comments.
7294 (rtl_verify_bb_layout): Ditto.
7295 * basic-block.h (probably_never_executed_edge_p): Declare.
7296 (fixup_partitions): Ditto.
7297 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7298 * bb-reorder.c (sanitize_hot_paths): New function.
7299 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7301 * predict.c (probably_never_executed_edge_p): New routine.
7302 * cfg.c (check_bb_profile): Add partition insanity warnings.
7304 2013-08-30 Meador Inge <meadori@codesourcery.com>
7306 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7308 2013-08-30 Marek Polacek <polacek@redhat.com>
7310 * Makefile.in (ubsan.o): Add.
7311 (c-family/c-ubsan.o): Add.
7312 (builtins.o): Add ubsan.h dependency.
7313 * ubsan.h: New file.
7314 * ubsan.c: New file.
7315 * common.opt: Add -fsanitize=undefined option.
7316 (flag_sanitize): Add variable.
7317 (fsanitize=): Add option. Add Driver.
7318 (fsanitize=thread): Remove option.
7319 (fsanitize=address): Likewise.
7320 (static-libubsan): New option.
7321 * doc/invoke.texi: Document the new flag and -static-libubsan.
7322 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7323 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7324 * builtin-attrs.def (ATTR_COLD): Define.
7325 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7326 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7327 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7328 * flag-types.h (sanitize_code): New enum.
7329 * opts.c (common_handle_option): Parse command line arguments
7330 of -fsanitize=. Add -fsanitize=unreachable option.
7331 * varasm.c (get_variable_section): Adjust.
7332 (assemble_noswitch_variable): Likewise.
7333 (assemble_variable): Likewise.
7334 (output_constant_def_contents): Likewise.
7335 (categorize_decl_for_section): Likewise.
7336 (place_block_symbol): Likewise.
7337 (output_object_block): Likewise.
7338 * builtins.def: Likewise.
7339 * toplev.c (compile_file): Likewise.
7340 (process_options): Likewise.
7341 * cppbuiltin.c: Likewise.
7342 * tsan.c (tsan_pass): Likewise.
7343 (tsan_gate): Likewise.
7344 (tsan_gate_O0): Likewise.
7345 * cfgexpand.c (partition_stack_vars): Likewise.
7346 (expand_stack_vars): Likewise.
7347 (defer_stack_allocation): Likewise.
7348 (expand_used_vars): Likewise.
7349 * cfgcleanup.c (old_insns_match_p): Likewise.
7350 * asan.c (asan_finish_file): Likewise.
7351 (asan_instrument): Likewise.
7352 (gate_asan): Likewise.
7353 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7354 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7355 (asan_global_struct): Use pointer_sized_int_node instead
7356 calling build_nonstandard_integer_type.
7357 (initialize_sanitizer_builtins): Likewise.
7358 (asan_finish_file): Likewise.
7359 * gcc.c: Document %{%:function(args):X}.
7360 (static_spec_functions): Add sanitize.
7361 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7362 store funcval != NULL there.
7363 (do_spec_1): Adjust handle_spec_function caller.
7364 (handle_braces): Allow %:function(args) as condition.
7365 (sanitize_spec_function): New function.
7366 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7367 (LIBUBSAN_SPEC): Likewise.
7368 (LIBUBSAN_EARLY_SPEC): Likewise.
7369 (SANITIZER_SPEC): Handle libubsan.
7370 (SANITIZER_EARLY_SPEC): Likewise.
7371 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7372 instead of fsanitize=address.
7373 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7374 instead of fsanitize=address*.
7375 * builtins.c: Include ubsan.h.
7376 (fold_builtin_0): Instrument __builtin_unreachable.
7377 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7378 instead of flag_asan.
7379 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7380 (pointer_sized_int_node): Define.
7381 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7383 2013-08-30 Mike Stump <mikestump@comcast.net>
7385 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7388 2013-08-29 Jan Hubicka <jh@suse.cz>
7390 * cgraph.c (cgraph_function_body_availability): Handle weakref
7392 * passes.def: Remove pass_fixup_cfg.
7393 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7394 track when we need to remove functions.
7395 (gate_ipa_inline): Execute inlining always; add comment why.
7396 (pass_data_ipa_inline): Remove TODO_remove_functions.
7397 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7398 do not produce summaries.
7399 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7400 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7401 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7402 constant pool and vtable.
7404 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7406 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7407 New arm_neon.h's internal macros to specify 64-bit constants.
7408 Avoid using stdint.h's macros.
7410 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7412 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7413 before applying REG_P.
7415 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7417 PR tree-optimization/58277
7418 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7419 after seeing too many stmts with vdef in between dombb and current
7420 bb, invalidate everything.
7422 2013-08-30 Richard Biener <rguenther@suse.de>
7424 * fold-const.c (fold_single_bit_test): Fix overflow test.
7426 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7428 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7429 and which can live in a register, always retrieve the value on entry.
7430 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7431 passed by invisible reference specially.
7432 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7433 (vt_add_function_parameter): Correctly deal with a parameter passed by
7434 invisible reference.
7436 2013-08-30 Jan Hubicka <jh@suse.cz>
7438 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7440 2013-08-30 Richard Biener <rguenther@suse.de>
7442 PR tree-optimization/58228
7443 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7444 allow invariant loads in nested loop vectorization.
7446 2013-08-30 Richard Biener <rguenther@suse.de>
7448 PR tree-optimization/58223
7449 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7450 (has_anti_or_output_dependence): ... this and adjust to also
7451 look for output dependences.
7452 (mark_nodes_having_upstream_mem_writes): Adjust.
7453 (rdg_flag_uses): Likewise.
7455 2013-08-30 Richard Biener <rguenther@suse.de>
7457 PR tree-optimization/58010
7458 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7459 assert that we have a loop-closed PHI.
7461 2013-08-29 Jan Hubicka <jh@suse.cz>
7463 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7464 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7465 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7466 (lto_free_function_in_decl_state_for_node): New function.
7468 2013-08-29 Xinliang David Li <davidxl@google.com>
7470 * loop-unroll.c (report_unroll_peel): Minor message change.
7471 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7472 Emit alignment peeling message with default -fopt-info.
7473 (vect_loop_versioning): Emit loop version info message.
7474 * tree-vectorizer.c (vectorize_loops): Minor message change.
7475 (execute_vect_slp): Ditto.
7477 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7479 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7480 of the clone from the DECL_NAME of the original function.
7482 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7484 * passes.c (register_pass): Add overload.
7485 * tree-pass.h (register_pass): Forward declare it. Add comment.
7487 2013-08-29 Jan Hubicka <jh@suse.cz>
7489 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7490 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7491 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7492 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7493 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7494 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7495 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7496 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7498 2013-08-29 Teresa Johnson <tejohnson@google.com>
7500 * dumpfile.c (dump_loc): Output column number.
7501 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7502 * doc/invoke.texi: Document optall -fopt-info flag.
7503 * profile.c (read_profile_edge_counts): Use new dump framework.
7504 (compute_branch_probabilities): Ditto.
7505 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7506 when pass not in any opt group.
7507 * pass_manager.h (pass_manager::get_pass_profile): New method.
7508 * value-prof.c (check_counter): Use new dump framework.
7509 (check_ic_target): Ditto.
7510 * coverage.c (get_coverage_counts): Ditto.
7511 (coverage_init): Setup new dump framework.
7513 2013-08-29 Richard Biener <rguenther@suse.de>
7515 PR tree-optimization/58246
7516 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7517 handle the dominance check inside a basic-block.
7519 2013-08-29 Richard Biener <rguenther@suse.de>
7522 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7523 of default defs that appear in abnormal PHI nodes.
7525 2013-08-29 Richard Biener <rguenther@suse.de>
7527 PR tree-optimization/57685
7528 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7529 single-use operands to avoid exponential complexity.
7531 2013-08-28 Dehao Chen <dehao@google.com>
7533 * ipa-inline.c (edge_badness): Fix integer underflow.
7535 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7537 * gtm-builtins.def (_ITM_free): Declare leaf.
7539 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7542 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7543 (*tls_local_dynamic_base_64_largepic): Likewise.
7544 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7545 Remove predicate from call operand.
7546 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7547 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7549 2013-08-28 Jeff Law <law@redhat.com>
7551 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7552 checks for the number of predecessors and successors allowed.
7553 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7554 which require copying a joiner block if there is a request which
7555 is a subpath that requires no joiner block copying.
7557 2013-08-28 Jan Hubicka <jh@suse.cz>
7559 * lto-streamer-out.c (DFS_write_tree_body): Drop
7560 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7561 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7562 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7563 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7564 Do not read DECL_ERROR_ISSUED.
7565 (unpack_ts_decl_with_vis_value_fields): Do not read
7567 (lto_input_ts_binfo_tree_pointers): Do not read
7568 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7569 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7570 write DECL_ERROR_ISSUED..
7571 (pack_ts_decl_with_vis_value_fields): Do not write
7573 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7574 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7575 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7576 * tree.h (tree_decl_common): Update comment.
7577 (DECL_ERROR_ISSUED): Remove.
7579 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7582 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7584 2013-08-28 Jan Hubicka <jh@suse.cz>
7586 * builtins.def (free): Declare leaf.
7588 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7591 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7592 * configure: Regenerate.
7594 2013-08-27 Martin Jambor <mjambor@suse.cz>
7596 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7597 (ipa_ancestor_jf_data): Likewise.
7598 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7599 (ipa_get_jf_pass_through_type_preserved): New function.
7600 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7601 (ipa_get_jf_ancestor_type_preserved): New function.
7602 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7603 (ipa_get_jf_ancestor_result): Likewise.
7604 (propagate_vals_accross_pass_through): Use
7605 ipa_get_jf_pass_through_result to do all the value mappings.
7606 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7607 type_preserved flag.
7608 (ipa_set_jf_cst_copy): New function.
7609 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7610 (ipa_set_jf_arith_pass_through): Likewise.
7611 (ipa_set_ancestor_jf): Likewise.
7612 (compute_complex_assign_jump_func): Set type_preserved instead of
7614 (ipa_compute_jump_functions_for_edge): Likewise.
7615 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7616 (update_jump_functions_after_inlining): Update type_preserved.
7617 Explicitely create jump functions when combining one with pass_through.
7618 (ipa_write_jump_function): Stream the type_preserved flags.
7619 (ipa_read_jump_function): Likewise.
7621 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7622 Aldy Hernandez <aldyh@redhat.com>
7624 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7625 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7626 * function.h (struct function): Add has_force_vect_loops and
7628 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7629 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7631 * gimple.def: Update CLAUSES comments.
7632 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7633 (gimple_build_omp_for): Add argument to prototype.
7634 (gimple_omp_for_kind): New.
7635 (gimple_omp_for_set_kind): New.
7636 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7637 GOVD_DATA_SHARE_CLASS.
7638 (enum omp_region_type): Add ORT_SIMD.
7639 (gimple_add_tmp_var): Handle ORT_SIMD.
7640 (gimplify_var_or_parm_decl): Same.
7641 (is_gimple_stmt): Same.
7642 (omp_firstprivatize_variable): Same.
7643 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7644 (omp_notice_variable): Handle ORT_SIMD.
7645 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7646 (omp_check_private): Handle ORT_SIMD.
7647 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7649 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7650 Handle OMP_CLAUSE_LASTPRIVATE.
7651 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7653 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7654 (gimplify_expr): Handle OMP_SIMD.
7655 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7656 (expand_GOMP_SIMD_VF): New.
7657 (expand_GOMP_SIMD_LAST_LANE): New.
7658 * internal-fn.def (GOMP_SIMD_LANE): New.
7659 (GOMP_SIMD_VF): New.
7660 (GOMP_SIMD_LAST_LANE): New.
7661 * omp-low.c: Include target.h.
7662 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7664 (check_omp_nesting_restrictions): Same.
7666 (lower_rec_simd_input_clauses): New.
7667 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7669 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7670 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7671 (expand_omp_build_assign): New.
7672 (expand_omp_for_init_counts): New.
7673 (expand_omp_for_init_vars): New.
7674 (extract_omp_for_update_vars): New.
7675 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7676 and rewrite accordingly.
7677 (expand_omp_simd): New.
7678 (expand_omp_for): Use expand_omp_simd.
7679 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7680 (lower_omp_for): Do not lower the body.
7681 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7683 * tree-flow.h (find_omp_clause): Remove prototype.
7684 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7685 forcing vectorization of the loop, or if flag_tree_vectorize.
7686 (gate_tree_if_conversion): Similarly.
7687 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7688 gimple_build_omp_for.
7689 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7690 * tree-parloops (create_parallel_loop): Pass kind argument to
7691 gimple_build_omp_for.
7692 * tree-pretty-print.c (dump_omp_clause): Add cases for
7693 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7694 OMP_CLAUSE__SIMDUID_.
7695 (dump_generic_node): Handle OMP_SIMD.
7696 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7697 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7698 unroll OMP_SIMD loops here.
7699 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7700 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7702 (vect_analyze_data_refs): Handle simd loops.
7703 * tree-vect-loop.c (vectorizable_live_operation): Handle
7705 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7706 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7707 (vectorizable_load): Same.
7708 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7709 (struct simduid_to_vf): New.
7710 (simduid_to_vf::hash): New.
7711 (simduid_to-vf::equal): New.
7712 (struct simd_array_to_simduid): New.
7713 (simd_array_to_simduid::hash): New.
7714 (simd_array_to_simduid::equal): New.
7715 (adjust_simduid_builtins): New.
7716 (struct note_simd_array_uses_struct): New.
7717 (note_simd_array_uses_cb): New.
7718 (note_simd_array_uses): New.
7719 (vectorize_loops): Handle simd hints and adjust simd builtins
7721 * tree-vectorizer.h (struct _stmt_vec_info): Add
7722 simd_lane_access_p field.
7723 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7724 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7725 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7726 (omp_clause_code_name): Same.
7727 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7728 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7729 * tree.def (OMP_SIMD): New entry.
7730 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7731 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7732 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7733 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7734 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7735 (OMP_CLAUSE_LINEAR_STEP): New.
7736 (OMP_CLAUSE_SAFELEN_EXPR): New.
7737 (OMP_CLAUSE__SIMDUID__DECL): New.
7738 (find_omp_clause): New prototype.
7740 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7742 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7743 Haswell processor detection.
7745 2013-08-27 Christian Widmer <shadow@umbrox.de>
7748 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7749 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7750 AVX2 capable processors.
7752 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7754 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7755 of vget_low_* with implementations in terms of other intrinsics.
7757 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7760 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7761 values to -1, 0 and 1.
7763 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7765 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7766 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7767 (<optab><mode>3_insn): Remove.
7768 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7769 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7770 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7771 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7772 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7773 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7774 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7775 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7776 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7777 (ror<mode>3_insn): Likewise.
7778 * config/aarch64/predicates.md (aarch64_simd_register): New.
7780 2013-08-27 Richard Biener <rguenther@suse.de>
7782 PR tree-optimization/57521
7783 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7784 one edge is non-critical.
7785 (find_phi_replacement_condition): Make sure to use a non-critical
7786 edge. Cleanup and remove old bug workarounds.
7787 (bb_postdominates_preds): Remove.
7788 (if_convertible_loop_p_1): Do not compute post-dominators.
7789 (combine_blocks): Do not free post-dominators.
7790 (main_tree_if_conversion): Likewise.
7791 (pass_data_if_conversion): Add TODO_verify_ssa.
7793 2013-08-27 DJ Delorie <dj@redhat.com>
7795 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7797 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7799 * function.c (assign_parm_find_data_types): Set passed_mode and
7800 nominal_mode to the TYPE_MODE of nominal_type for the built
7801 pointer type in case of the struct-pass-by-reference.
7803 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7805 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7806 (UINT16_TYPE): Change default to "unsigned int".
7808 * config/avr/avr.opt (mfract-convert-truncate): New option.
7809 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7810 is set, round negative fractional integers according to n1169
7811 when converting to integer types.
7813 2013-08-26 Jan Hubicka <jh@suse.cz>
7815 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7816 methods can not be called indirectly when their address is not taken.
7818 2013-08-26 Jan Hubicka <jh@suse.cz>
7820 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7823 2013-08-26 Jan Hubicka <jh@suse.cz>
7825 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7828 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7830 * reload.c (find_valid_class): Allow classes that do not include
7831 FIRST_PSEUDO_REGISTER - 1.
7833 2013-08-26 Jan Hubicka <jh@suse.cz>
7835 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7836 fix edge count/frequency when speculation failed; fix type check
7837 for the direct call.
7839 2013-08-26 Jan Hubicka <jh@suse.cz>
7841 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7843 2013-08-26 Jan Hubicka <jh@suse.cz>
7845 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7846 to be changed by edge redirection.
7848 2013-08-26 Jan Hubicka <jh@suse.cz>
7850 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7851 formating; add sanity check.
7852 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7853 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7854 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7855 (resolve_noninline_speculation): Update callee keys, too.
7857 2013-08-26 Jan Hubicka <jh@suse.cz>
7859 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7860 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7862 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7864 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7867 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
7869 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7870 (x86_64_elf_select_section): Ditto.
7871 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7873 (ix86_pass_by_reference): Ditto.
7874 (output_set_got): Ditto.
7875 (ix86_unary_operator_ok): Ditto.
7876 (ix86_expand_builtin): Ditto.
7878 2013-08-23 Jan Hubicka <jh@suse.cz>
7880 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
7882 2013-08-23 Jan Hubicka <jh@suse.cz>
7884 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7885 (tree_decl_with_vis): Add FINAL field.
7887 2013-08-23 Jeff Law <law@redhat.com>
7889 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7890 the debugging dump when the expression is fully redundant.
7892 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7894 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7895 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7896 * pretty-print.c (pp_formatted_text_data): Likewise.
7897 (pp_write_text_to_stream): Likewise.
7898 (pp_write_text_as_dot_label_to_stream): Likewise.
7899 (pp_append_r): Likewise.
7900 (pp_format): Likewise.
7901 (pp_flush): Likewise.
7902 (pp_clear_output_area): Likewise.
7903 (pp_append_text): Likewise.
7904 (pp_formatted_text): Likewise.
7905 (pp_remaining_character_count_for_line): Likewise.
7906 (pp_newline): Likewise.
7907 (pp_character): Likewise.
7908 (output_buffer::~output_buffer): Define.
7909 (pretty_printer::~pretty_printer): Destruct output buffer.
7910 * pretty-print.h (output_buffer::~output_buffer): Declare.
7911 (pretty_printer::~pretty_printer): Declare virtual.
7913 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
7916 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7917 HOST_WIDE_INT_M1U): New macros.
7918 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7919 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7920 unsigned -1 for lshift.
7921 * cse.c (cse_insn): Likewise.
7922 * double-int.c (rshift_double, lshift_double): Likewise.
7923 * builtins.c (fold_builtin_bitop): Likewise.
7924 * combine.c (force_to_mode): Likewise.
7925 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7926 * simplify-rtx.c (simplify_const_unary_operation,
7927 simplify_const_binary_operation): Likewise.
7928 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7929 check_va_list_escapes): Likewise.
7930 * rtlanal.c (nonzero_bits1): Likewise.
7931 * expmed.c (expand_smod_pow2): Likewise.
7932 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7934 2013-08-23 Jan Hubicka <jh@suse.cz>
7936 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7937 as having address taken.
7939 2013-08-23 Jan Hubicka <jh@suse.cz>
7941 * ipa-utils.h (method_class_type): Declare.
7942 * ipa-devirt.c (method_class_type): Export.
7944 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7945 do not walk base classes of anonymous types.
7947 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7949 PR rtl-optimization/58220
7951 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7952 handle SEQUENCE insns properly.
7954 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7956 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7958 (pp_newline_and_indent): Likewise.
7959 (pp_separate_with): Likewise.
7960 * pretty-print.c (pp_newline_and_flush): Define.
7961 (pp_newline_and_indent): Likewise.
7962 (pp_separate_with): Likewise.
7964 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7967 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7968 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7970 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7972 * config/i386/predicates.md (ext_sse_reg_operand): New.
7973 * config/i386/i386.md (*movti_internal): Use
7974 predicate to determine if EVEX is needed.
7975 (*movsi_internal): Ditto.
7976 (*movdf_internal): Ditto.
7977 (*movsf_internal): Ditto.
7978 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
7980 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7982 PR tree-optimization/58209
7983 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
7984 (find_tail_calls): Give up for pointer result types if m is non-NULL.
7985 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
7986 emit POINTER_PLUS_EXPR.
7987 (create_tailcall_accumulator): For pointer result type accumulate in
7990 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
7992 * configure.ac: Add backslashes missing from the last change.
7993 * configure: Regenerate.
7995 2013-08-22 Jan Hubicka <jh@susue.cz>
7997 * ipa.c (function_and_variable_visibility): First remember function
7998 was global and then make it local.
8000 2013-08-22 Julian Brown <julian@codesourcery.com>
8002 * configure.ac: Add aarch64 to list of arches which use "nop" in
8004 * configure: Regenerate.
8006 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8008 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8009 gnu_libc_has_function.
8010 * config/s390/tpf.h: Likewise.
8012 2013-08-22 Jan Hubicka <jh@susue.cz>
8014 * timevar.c (validate_phases): Add cast.
8016 2013-08-22 Jan Hubicka <jh@susue.cz>
8018 * timevar.c (validate_phases): Use size_t for memory.
8019 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8021 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8023 * pretty-print.h (output_buffer::output_buffer): Declare.
8024 (pretty_printer::pretty_printer): Likewise.
8025 (pp_construct): Remove.
8026 * pretty-print.c (output_buffer::output_buffer): Define.
8027 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8028 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8029 (print_gimple_expr): Likewise.
8030 (print_gimple_seq): Likewise.
8031 (gimple_dump_bb): Likewise.
8032 * sched-vis.c (dump_value_slim): Likewise.
8033 (dump_insn_slim): Likewise.
8034 (dump_rtl_slim): Likewise.
8035 (str_pattern_slim): Likewise.
8036 * tree-mudflap.c (mf_varname_tree): Likewise.
8037 * graph.c (print_graph_cfg): Likewise.
8038 (start_graph_dump): Likewise.
8039 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8041 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8042 pretty printer initialization.
8043 * coretypes.h (diagnostic_context): Remove superflous type alias
8045 (pretty_printer): Likewise. Declare directly as a class.
8046 (pretty_print_info): Remove declaration as class.
8047 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8048 and pp_clear_output_area.
8049 (asan_add_global): Likewise.
8051 2013-08-22 Jan Hubicka <jh@suse.cz>
8053 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8054 * ipa-utils.h (update_type_inheritance_graph): Declare.
8055 (possible_polymorphic_call_target_p): Declare.
8056 (possible_polymorphic_call_target_p): New.
8057 * ipa-devirt.c: Update toplevel comments.
8058 (cached_polymorphic_call_targets): Move up.
8059 (odr_type_d): Move ID down.
8060 (polymorphic_type_binfo_p): Update comment.
8061 (odr_hasher::remove): Likewise;
8062 (get_odr_type): Set anonymous_namespace.
8063 (dump_odr_type): Dump it.
8064 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8065 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8066 (record_binfo): Add comment.
8067 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8069 (devirt_node_removal_hook): Do not iCE when cache is freed.
8070 (possible_polymorphic_call_target_p): New predicate.
8071 (update_type_inheritance_graph): New function.
8073 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8074 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8075 Sergey Lega <sergey.s.lega@intel.com>
8076 Anna Tikhonova <anna.tikhonova@intel.com>
8077 Ilya Tocar <ilya.tocar@intel.com>
8078 Andrey Turetskiy <andrey.turetskiy@intel.com>
8079 Ilya Verbin <ilya.verbin@intel.com>
8080 Kirill Yukhin <kirill.yukhin@intel.com>
8081 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8083 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8084 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8085 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8086 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8087 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8088 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8089 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8090 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8091 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8092 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8093 OPT_mavx512pf, OPT_mavx512er cases.
8094 * config/i386/constraints.md (v): New constraint.
8095 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8096 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8097 (bit_AVX512CD): New.
8098 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8099 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8100 * config/i386/i386-c.c (ix86_target_macros_internal):
8101 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8103 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8104 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8105 * config/i386/i386.c (regclass_map, dbx_register_map)
8106 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8107 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8108 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8110 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8111 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8112 -mavx512pf options. Fix formatting.
8113 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8114 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8115 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8116 -mavx512cd, -mavx512pf options.
8117 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8118 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8119 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8121 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8122 (ix86_set_reg_reg_cost): Ditto.
8123 (x86_order_regs_for_local_alloc): Ditto.
8124 (MAX_VECT_LEN): Extend to 64-byte.
8125 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8126 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8127 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8128 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8129 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8130 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8131 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8132 (SSE_REG_MODE_P): Support new modes.
8133 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8134 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8135 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8136 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8137 (REG_CLASS_CONTENTS): Add new registers.
8138 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8139 (EXT_REX_SSE_REGNO_P): New.
8140 (HI_REGISTER_NAMES): Add new registers.
8141 * config/i386/i386.md: Define constants for new registers.
8142 (mode): Add new 512-bit modes.
8143 (prefix): Support evex prefix.
8144 (isa): Support avx512f, noavx512f, fma_avx512f.
8145 (ssemodesuffix): Add new 512-bit modes.
8147 (*movxi_internal_avx512f): Ditto.
8148 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8150 (*movsi_internal): Likewise.
8151 (*movdf_internal): Likewise.
8152 (*movsf_internal): Likewise.
8153 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8154 (<code><mode>3): Likewise.
8155 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8157 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8158 with the new constraint "v".
8159 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8161 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8162 with the new constraint "v".
8163 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8164 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8165 (*<plusminus_insn><mode>3): Likewise.
8166 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8167 (*mul<mode>3): Likewise.
8168 (<sse>_vmmul<mode>3): Likewise.
8169 (<sse>_div<mode>3): Likewise.
8170 (<sse>_vmdiv<mode>3): Likewise.
8171 (<sse>_sqrt<mode>2): Likewise.
8172 (<sse>_vmsqrt<mode>2): Likewise.
8173 (*<code><mode>3_finite): Likewise.
8174 (*<code><mode>3) <smaxmin>: Likewise.
8175 (<sse>_vm<code><mode>3): Likewise.
8176 (*<code><mode>3) <any_logic>: Likewise.
8177 (*fma_fmadd_<mode>): Likewise.
8178 (*fma_fmsub_<mode>): Likewise.
8179 (*fma_fnmadd_<mode>): Likewise.
8180 (*fma_fnmsub_<mode>): Likewise.
8181 (*fma_fmaddsub_<mode>): Likewise.
8182 (*fma_fmsubadd_<mode>): Likewise.
8183 (*fmai_fmadd_<mode>): Likewise.
8184 (*fmai_fmsub_<mode>): Likewise.
8185 (*fmai_fnmadd_<mode>): Likewise.
8186 (*fmai_fnmsub_<mode>): Likewise.
8187 (sse_cvtsi2ss): Likewise.
8188 (sse_cvtsi2ssq): Likewise.
8189 (sse_cvtss2si): Likewise.
8190 (sse_cvtss2si_2): Likewise.
8191 (sse_cvtss2siq): Likewise.
8192 (sse_cvtss2siq_2): Likewise.
8193 (sse_cvttss2si): Likewise.
8194 (sse_cvtss2siq_2): Likewise.
8195 (float<sseintvecmodelower><mode>2): Likewise.
8196 (sse2_cvtsd2si_2): Likewise.
8197 (sse2_cvtsd2siq_2): Likewise.
8198 (*<plusminus_insn><mode>3): Likewise.
8199 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8200 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8201 (ashr<mode>3): Likewise.
8202 (<shift_insn><mode>3): Likewise.
8203 (avx2_<code><mode>3): Likewise.
8204 (*avx2_<code><mode>3): Likewise.
8205 (*andnot<mode>3): Likewise.
8206 (*<code><mode>3) <any_logic>: Likewise.
8207 (abs<mode>2): Likewise.
8208 (avx2_permvar<mode>): Likewise.
8209 (avx2_perm<mode>_1): Likewise.
8210 (*avx_vpermilp<mode>): Likewise.
8211 (avx_vpermilvar<mode>3): Likewise.
8212 (avx2_ashrv<mode>): Likewise.
8213 (avx2_<shift_insn>v<mode>): Likewise.
8214 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8216 * doc/rtl.texi: Document XImode.
8218 2013-08-21 Jeff Law <law@redhat.com>
8220 * tree-flow.h (register_jump_thread): Pass vector of edges
8221 instead of each important edge.
8222 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8223 thread path into a vector and pass that to register_jump_thread.
8224 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8225 passed in edge vector to the current 3-edge form.
8228 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8230 * dce.c (fini_dce): Call df_analyze again just in case
8231 delete_unmarked_insns removed anything.
8233 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8235 * reload.h (struct reg_equivs): Rename to ..
8236 (struct reg_equivs_s): .. this.
8238 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8240 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8242 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8244 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8246 2013-08-21 Jeff Law <law@redhat.com>
8248 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8249 simplify assignments too. If the RHS collapses to a singleton
8250 range, then return the value for the range.
8252 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8254 * config/i386/sse.md (V16): Rename to...
8256 (mov<mode>): Update iterator name.
8257 (*mov<mode>_internal): Ditto.
8258 (push<mode>1): Ditto.
8259 (movmisalign<mode>): Ditto.
8261 2013-08-20 Jan Hubicka <jh@suse.cz>
8264 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8265 entry for direct edges.
8266 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8268 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8270 Revert my last two changes, r201865 and r201864:
8273 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8275 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8276 instances can own GC refs.
8278 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8279 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8280 (gcc::context::gt_pch_nx): Likewise.
8281 (gcc::context::gt_pch_nx): Likewise.
8282 * ggc.h (gt_ggc_mx <T>): New.
8283 (gt_pch_nx_with_op <T>): New.
8284 (gt_pch_nx <T>): New.
8285 * passes.c (opt_pass::gt_ggc_mx): New.
8286 (opt_pass::gt_pch_nx): New.
8287 (opt_pass::gt_pch_nx_with_op): New.
8288 (pass_manager::gt_ggc_mx): New.
8289 (pass_manager::gt_pch_nx): New.
8290 (pass_manager::gt_pch_nx_with_op): New.
8291 (pass_manager::operator new): Use
8292 ggc_internal_cleared_alloc_stat rather than xcalloc.
8293 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8294 (pass_manager::gt_ggc_mx): New.
8295 (pass_manager::gt_pch_nx): New.
8296 (pass_manager::gt_pch_nx_with_op): New.
8297 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8298 (opt_pass::operator new): New.
8299 (opt_pass::gt_ggc_mx): New.
8300 (opt_pass::gt_pch_nx): New.
8301 (opt_pass::gt_pch_nx_with_op): New.
8304 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8306 * Makefile.in (GTFILES): Add context.h.
8307 * context.c (gcc::context::operator new): New.
8308 (gcc::context::gt_ggc_mx): New.
8309 (gcc::context::gt_pch_nx): New.
8310 (gcc::context::gt_pch_nx): New.
8311 * context.h (gcc::context): Add GTY((user)) marking.
8312 (gcc::context::operator new): New.
8313 (gcc::context::gt_ggc_mx): New.
8314 (gcc::context::gt_pch_nx): New.
8315 (gcc::context::gt_pch_nx): New.
8316 (g): Add GTY marking.
8317 (gt_ggc_mx (gcc::context *)): New.
8318 (gt_pch_nx (gcc::context *)): New.
8319 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8320 void *cookie)): New.
8321 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8323 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8325 * dce.c (fini_dce): Call df_analyze again just in case
8326 delete_unmarked_insns removed anything.
8328 2013-08-20 Teresa Johnson <tejohnson@google.com>
8330 PR rtl-optimizations/57451
8331 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8332 from crossing split section boundaries.
8334 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8336 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8337 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8338 with MULTLIB_DEFAULTS.
8340 2013-08-20 Nick Clifton <nickc@redhat.com>
8342 * target.def (narrow_volatile_bitfield): Note that the default
8343 value is false, not !TARGET_STRICT_ALIGN.
8344 * doc/tm.texi: Regenerate.
8346 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8348 Fix LIB_SPEC for systems without libpthread.
8350 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8351 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8353 * config/i386/linux-common.h: Likewise.
8354 * config/mips/linux-common.h: Likewise.
8356 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8358 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8361 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8363 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8364 instances can own GC refs.
8366 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8367 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8368 (gcc::context::gt_pch_nx): Likewise.
8369 (gcc::context::gt_pch_nx): Likewise.
8370 * ggc.h (gt_ggc_mx <T>): New.
8371 (gt_pch_nx_with_op <T>): New.
8372 (gt_pch_nx <T>): New.
8373 * passes.c (opt_pass::gt_ggc_mx): New.
8374 (opt_pass::gt_pch_nx): New.
8375 (opt_pass::gt_pch_nx_with_op): New.
8376 (pass_manager::gt_ggc_mx): New.
8377 (pass_manager::gt_pch_nx): New.
8378 (pass_manager::gt_pch_nx_with_op): New.
8379 (pass_manager::operator new): Use
8380 ggc_internal_cleared_alloc_stat rather than xcalloc.
8381 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8382 (pass_manager::gt_ggc_mx): New.
8383 (pass_manager::gt_pch_nx): New.
8384 (pass_manager::gt_pch_nx_with_op): New.
8385 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8386 (opt_pass::operator new): New.
8387 (opt_pass::gt_ggc_mx): New.
8388 (opt_pass::gt_pch_nx): New.
8389 (opt_pass::gt_pch_nx_with_op): New.
8391 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8393 * Makefile.in (GTFILES): Add context.h.
8394 * context.c (gcc::context::operator new): New.
8395 (gcc::context::gt_ggc_mx): New.
8396 (gcc::context::gt_pch_nx): New.
8397 (gcc::context::gt_pch_nx): New.
8398 * context.h (gcc::context): Add GTY((user)) marking.
8399 (gcc::context::operator new): New.
8400 (gcc::context::gt_ggc_mx): New.
8401 (gcc::context::gt_pch_nx): New.
8402 (gcc::context::gt_pch_nx): New.
8403 (g): Add GTY marking.
8404 (gt_ggc_mx (gcc::context *)): New.
8405 (gt_pch_nx (gcc::context *)): New.
8406 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8407 void *cookie)): New.
8408 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8410 2013-08-20 Alan Modra <amodra@gmail.com>
8413 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8414 (rs6000_emit_epilogue): Likewise.
8416 2013-08-19 Dehao Chen <dehao@google.com>
8418 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8420 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8421 Jakub Jelinek <jakub@redhat.com>
8423 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8424 (BUILT_IN_FABSD64): Likewise.
8425 (BUILT_IN_FABSD128): Likewise.
8426 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8427 (fold_builtin_1): Likewise.
8428 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8429 destination and source operands.
8430 (*abstd2_fpr): Likewise.
8431 (*nabstd2_fpr): Likewise.
8433 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8435 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8438 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8440 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8443 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8445 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8446 * builtins.c (default_libc_has_function): New.
8447 (gnu_libc_has_function): Ditto.
8448 (no_c99_libc_has_function): Ditto.
8449 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8450 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8451 (fold_builtin_sincos): Likewise.
8452 (fold_builtin_cexp): Likewise.
8453 * builtins.def (DEF_C94_BUILTIN): Likewise.
8454 (DEF_C99_BUILTIN): Likewise.
8455 (DEF_C99_C90RES_BUILTIN): Likewise.
8456 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8457 definitions to using this define.
8458 * config/darwin-protos.h (darwin_libc_has_function): New.
8459 * config/darwin.c (darwin_libc_has_function): Ditto.
8460 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8461 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8462 * config/darwin.h: Ditto.
8463 * config/elfos.h: Ditto.
8464 * config/freebsd.h: Ditto.
8465 * config/i386/cygming.h: Ditto.
8466 * config/i386/djgpp.h: Ditto.
8467 * config/i386/i386-interix.h: Ditto.
8468 * config/microblaze/microblaze.h: Ditto.
8469 * config/mmix/mmix.h: Ditto.
8470 * config/gnu-user.h: Ditto.
8471 * config/ia64/hpux.h: Ditto.
8472 * config/pa/pa-hpux.h: Ditto.
8473 * config/pdp11/pdp11.h: Ditto.
8474 * config/picochip/picochip.h: Ditto.
8475 * config/linux.h: Ditto.
8476 * config/netbsd.h: Ditto.
8477 * config/openbsd.h: Ditto.
8478 * config/rs6000/aix43.h: Ditto.
8479 * config/rs6000/aix51.h: Ditto.
8480 * config/rs6000/aix52.h: Ditto.
8481 * config/rs6000/aix53.h: Ditto.
8482 * config/rs6000/aix61.h: Ditto.
8483 * config/rs6000/darwin.h: Ditto.
8484 * config/rs6000/linux.h: Ditto.
8485 * config/rs6000/linux64.h: Ditto.
8486 * config/s390/tpf.h: Ditto.
8487 * config/sol2-10.h: Ditto.
8488 * config/sol2.h: Ditto.
8489 * config/vms/vms.h: Ditto.
8490 * config/vxworks.h: Ditto.
8491 * config/linux-android.c (linux_android_libc_has_function):
8492 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8493 * config/linux-protos.h (linux_android_libc_has_function):
8495 * config/i386/i386.c (ix86_libc_has_function): New.
8496 * config/i386/i386-protos.h
8497 (ix86_libc_has_function): New declaration.
8498 * config/i386/i386.md
8499 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8500 ("isinf<mode>2): Likewise.
8501 * convert.c (convert_to_integer): Using new target hook
8502 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8503 TARGET_C99_FUNCTIONS.
8504 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8505 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8506 * coretypes.h (function_class): New enum for different
8507 classes of functions.
8508 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8509 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8510 (TARGET_HAS_SINCOS): Likewise.
8511 (TARGET_LIBC_HAS_FUNCTION): New.
8512 * doc/tm.texi: Regenerated.
8513 * targhooks.h (default_libc_has_function): New declaration.
8514 (no_c99_libc_has_function): Ditto.
8515 (gnu_libc_has_function): Ditto.
8516 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8517 and TARGET_HAS_SINCOS.
8519 2013-08-18 Jan Hubicka <jh@suse.cz>
8521 * Makeifle-in (ipa-devirt.o): New.
8522 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8523 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8524 (analyze_functions): Look into possible targets of polymorphic call.
8525 * dumpfile.c (dump_files): Add type-inheritance dump.
8526 * dumpfile.h (TDI_inheritance): New.
8527 * ipa-devirt.c: New file.
8528 * ipa-utils.h (odr_type_d): Forward declare.
8529 (odr_type): New type.
8530 (build_type_inheritance_graph): Declare.
8531 (possible_polymorphic_call_targets): Declare and introduce inline
8532 variant when only edge is pased.
8533 (dump_possible_polymorphic_call_targets): Likewise.
8534 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8535 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8536 (types_same_for_odr): ... here.
8537 * tree.h (type_in_anonymous_namespace_p): Declare.
8539 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8541 PR tree-optimization/58006
8542 * tree-parloops.c (take_address_of): Don't ICE if get_name
8544 (eliminate_local_variables_stmt): Remove clobber stmts.
8546 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8548 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8549 error message is issued for an alias to undefined symbol.
8551 2013-08-18 Jan Hubicka <jh@suse.cz>
8553 * cgraph.c (cgraph_create_indirect_edge): Discover
8554 polymorphic calls and record basic info into indirect_info.
8555 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8556 devirtualization, ignore objc function calls.
8557 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8558 call with no parm index info.
8559 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8560 * tree.c (virtual_method_call_p): New function.
8561 * tree.h (virtual_method_call_p): Declare.
8563 2013-08-16 Jan Hubicka <jh@suse.cz>
8566 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8568 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8570 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8573 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8575 * gengtype.c (type_for_name): Add special-case support for
8576 locating types within the "gcc::" namespace.
8577 (open_base_files): Emit a "using namespace gcc" directive.
8579 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8582 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8583 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8585 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8586 array instead of each individual operand as a separate argument.
8587 (emit_fusion_gpr_load): Likewise.
8588 (expand_fusion_gpr_load): Add new function declaration.
8590 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8591 signature to have the operands passed as an array, instead of as
8592 separate arguments. Allow ZERO_EXTEND to be in the memory
8593 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8594 depend on the register live/dead flags when peepholes are run.
8595 (expand_fusion_gpr_load): New function to be called from the
8596 peephole2 pass, to change the register that addis sets to be the
8598 (emit_fusion_gpr_load): Change the calling signature to have the
8599 operands passed as an array, instead of as separate arguments.
8600 Allow ZERO_EXTEND to be in the memory address, and also
8601 SIGN_EXTEND if -mpower8-fusion-sign.
8603 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8605 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8606 adjust the register addis loads up in the peephole2 pass. Do not
8607 depend on the register live/dead state when the peephole pass is done.
8609 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8611 * gengtype.c (create_user_defined_type): Ensure that the kind
8612 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8613 declaration is seen before the GTY((user)) marking.
8615 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8618 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8620 2013-08-16 Jan Hubicka <jh@suse.cz>
8622 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8623 arugment expected_type.
8624 (gimple_fold_call): Use it.
8625 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8626 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8627 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8628 (try_make_edge_direct_virtual_call): Likewise.
8629 * tree.c (obj_type_ref_class): New.
8630 * tree.h (obj_type_ref_class): Use it.
8632 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8634 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8635 (rtl_slim_pp): Likewise.
8636 (init_rtl_slim_pretty_print): Likewise.
8637 (dump_value_slim): Don't call it. Use local pretty printer.
8638 (dump_insn_slim): Likewise.
8639 (dump_rtl_slim): Likewise.
8640 (str_pattern_slim): Likewise.
8641 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8644 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8646 PR tree-optimization/58164
8647 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8648 walk gimple_goto_dest of GIMPLE_GOTO.
8650 PR tree-optimization/58165
8651 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8652 bi_call must be the last stmt in a bb, don't split_block, instead
8653 use fallthru edge from it and give up if there is none.
8654 Release conds vector when returning early.
8656 2013-08-14 Xinliang David Li <davidxl@google.com>
8658 * config/i386/i386.c (ix86_option_override_internal):
8659 Remove unused variable and field.
8661 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8664 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8665 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8666 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8667 and Linux, correct BLKmode alignment when 128-bit alignment is
8668 required and compatibility flag is not set.
8669 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8670 for zero-size arguments when compatibility flag is not set.
8672 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8674 PR tree-optimization/58145
8675 * tree-sra.c (build_ref_for_offset): If prev_base has
8676 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8678 2013-08-14 Xinliang David Li <davidxl@google.com>
8680 * config/i386/i386.c (ix86_option_override_internal):
8681 Fix uninitialized variable error.
8683 2013-08-14 Xinliang David Li <davidxl@google.com>
8685 * config/i386/i386.opt: Define two new options.
8686 * config/i386/x86-tune.def: Add arch selector field in macros.
8687 * config/i386/i386.h: Adjust macro definition.
8688 * config/i386/i386.c (ix86_option_override_internal):
8690 (parse_mtune_ctrl_str): New function.
8691 (set_ix86_tune_features): New function.
8692 (ix86_function_specific_restore): Call the new helper function.
8694 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8696 PR rtl-optimization/57662
8697 * sel-sched.c (code_motion_process_successors): When the current insn
8698 is removed after the recursive traversal, break from the loop.
8699 Add comments and debug printouts.
8701 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8702 Alexandre Oliva <aoliva@redhat.com>
8705 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8706 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8707 there also UNSPEC_PLTOFF.
8709 2013-08-14 Marek Polacek <polacek@redhat.com>
8711 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8712 NUM_CONDITIONS bit positions.
8714 2013-08-13 Cary Coutant <ccoutant@google.com>
8716 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8717 (attr_checksum): Hash vector contents instead of pointer.
8718 (attr_checksum_ordered): Likewise.
8720 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8722 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8723 when Pmode != word_mode. Add length_address attribute.
8724 (sse3_monitor_<mode>): Merge from sse3_monitor and
8725 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8726 Pmode != word_mode. Update insn length attribute.
8727 * config/i386/i386.c (ix86_option_override_internal): Update
8728 ix86_gen_monitor selection for merged sse3_monitor insn.
8730 2013-08-13 Julian Brown <julian@codesourcery.com>
8732 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8733 perform invalid legitimization on greater-than-word-size modes for
8736 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8738 * ira.c (setup_class_translate_array): Use aclass instead of cl
8739 for classes not fully covered by allocno classes.
8741 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8743 PR tree-optimization/57661
8744 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8745 * tree-inline.c (tree_function_versioning): Initialize it.
8746 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8747 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8748 that is not being copied.
8751 * asan.c (instrument_strlen_call): Fix typo in comment.
8752 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8754 2013-08-13 Steve Ellcey <sellcey@mips.com>
8756 * config/mips/mips.md (prefetch): Use lw instead of ld on
8757 loongson in 32bit mode.
8759 2013-08-13 Nick Clifton <nickc@redhat.com>
8761 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8763 2013-08-13 Jan Hubicka <jh@suse.cz>
8765 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8766 introduced edge; fix typo in sanity check.
8767 (cgraph_resolve_speculation): Export; improve diagnostic.
8768 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8769 speculation at type mismatch.
8770 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8771 (cgraph_resolve_speculation): Declare.
8772 (symtab_can_be_discarded): New function.
8773 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8774 * ipa-inline-transform.c (speculation_removed): New global var.
8775 (clone_inlined_nodes): See if speculation can be removed.
8776 (inline_call): If speculations was removed, we growths may not match.
8777 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8778 (speculation_useful_p): New function.
8779 (resolve_noninline_speculation): New function.
8780 (inline_small_functions): Resolve useless speculations.
8781 * ipa-inline.h (speculation_useful_p): Declare
8782 * ipa.c (can_replace_by_local_alias): Simplify.
8783 (ipa_profile): Produce speculative calls in non-lto, too;
8784 add simple cost model; produce local aliases.
8786 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8788 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8791 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8793 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8795 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8797 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8798 iterator instead of X87MODEF.
8800 2013-08-12 Perez Read <netfirewall@gmail.com>
8803 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8804 operand 0 for intel asm alternative.
8805 (*movabs<mode>_2): Ditto for operand 1.
8807 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8809 * config/aarch64/arm_none.h
8810 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8812 2013-08-12 Nick Clifton <nickc@redhat.com>
8814 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8816 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8818 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8819 expand for QI/HImode operand to produce more effictive code for
8820 unsigned char(short) --> float(double) conversion.
8822 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8824 * doc/invoke.texi: Mention that -ftls-model does not force the final
8827 2013-08-12 Marek Polacek <polacek@redhat.com>
8828 Marc Glisse <marc.glisse@inria.fr>
8830 PR tree-optimization/57980
8831 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8832 when creating -1 constant.
8834 2013-08-10 Jan Hubicka <jh@suse.cz>
8836 Workaround binutils PR14342.
8837 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8838 (gimple_init_edge_profiler): Likewise.
8839 (gimple_gen_ic_func_profiler): Likewise.
8841 2013-08-09 Jan Hubicka <jh@suse.cz>
8843 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8845 2013-08-09 Xinliang David Li <davidxl@google.com>
8847 * config/i386/stringop.def: New file.
8848 * config/i386/stringop.opt: New file.
8849 * config/i386/i386-opts.h: Include stringopt.def.
8850 * config/i386/i386.opt: Include stringopt.opt.
8851 * config/i386/i386.c (ix86_option_override_internal):
8852 Override default size based stringop inline strategies with options.
8853 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8856 2013-08-09 Jan Hubicka <jh@suse.cz>
8858 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
8860 2013-08-09 Jan Hubicka <jh@suse.cz>
8862 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8864 (dump_cgraph_node): Dump profile-id.
8865 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8866 and common_target_probability.
8867 * lto-cgraph.c (lto_output_edge): Stream common targets.
8868 (lto_output_node): Stream profile ids.
8869 (input_node): Stream profile ids.
8870 (input_edge): Stream common targets.
8871 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8872 * ipa.c: Include value-prof.h
8873 (ipa_profile_generate_summary): Turn indirect call statement histograms
8874 into common targets.
8875 (ipa_profile): Turn common targets into speculative edges.
8877 2013-08-09 Jan Hubicka <jh@suse.cz>
8879 * cgraph.h (cgraph_node): Add profile_id.
8880 * value-prof.c (cgraph_node_map): Turn into pointer_map.
8881 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
8882 (del_node_map): Update.
8883 (find_func_by_funcdef_no): Replace by ...
8884 (find_func_by_profile_id): ... this one.
8885 (gimple_ic_transform): Do not remove useful histograms when
8886 speculation is not done; dump info when indirect call removal
8888 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8889 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8890 (__gcov_indirect_call_profiler_v2): .. this one.
8891 * profile.h (init_node_map): Update.
8892 * coverage.c (coverage_compute_profile_id): New function.
8893 * coverage.h (coverage_compute_profile_id): Declare.
8894 * tree-profile.c (init_ic_make_global_vars): Make
8895 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8896 (gimple_init_edge_profiler): Update prototype of
8897 __gcov_indirect_call_profiler.
8898 (gimple_gen_ic_func_profiler): Simplify.
8899 (tree_profiling): Use init_node_map
8901 2013-08-09 Jan Hubicka <jh@suse.cz>
8903 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8904 non-speculative refs.
8905 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8906 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8907 (cgraph_set_call_stmt): Likewise.
8908 (cgraph_create_edge_1): Fix release checking compilatoin;
8910 (cgraph_free_edge): Free indirect info.
8911 (cgraph_turn_edge_to_speculative): New function.
8912 (cgraph_speculative_call_info): New function.
8913 (cgraph_make_edge_direct): Return direct edge; handle speculation.
8914 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
8915 (dump_cgraph_node): Dump speculation.
8916 (verify_edge_count_and_frequency): Accept speculative edges.
8917 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8918 (verify_cgraph_node): Handle speculation.
8919 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8920 (cgraph_set_call_stmt): Update prototype.
8921 (cgraph_make_edge_direct): Update prototype.
8922 (cgraph_speculative_call_info): Declare.
8923 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8924 to change; update call of ipa_find_references.
8925 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8926 lto_stmt_uid and speculative flags.
8927 (ipa_dump_references): Dump speculation.
8928 (ipa_clone_references): Clone speculative flag.
8929 (ipa_clone_referring): Likewise.
8930 (ipa_clone_ref): New function.
8931 (ipa_find_reference): Look into lto_stmt_uids
8932 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8933 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8934 (ipa_find_reference): Update declaration.
8935 (ipa_clone_ref): Declare.
8936 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8937 stream speculative flag.
8938 (lto_output_ref): Stream statements uids and speculation.
8939 (input_ref): Likewise.
8940 (input_edge): Stream speuclation.
8941 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8942 (cgraph_set_call_stmt_including_clones): Handle speculation.
8943 * ipa-inline.c (heap_edge_removal_hook): New function.
8944 (inline_small_functions): Register it.
8945 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8946 also initialize refs.
8947 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8949 (try_make_edge_direct_simple_call): Likewise.
8950 (try_make_edge_direct_simple_call): Likewise.
8951 (update_indirect_edges_after_inlining): Likewise.
8952 (remove_described_reference): Look proper lto_stmt_uid.
8953 (propagate_controlled_uses): Likewise.
8954 (propagate_controlled_uses): Liekwise.
8955 * tree-inline.c (copy_bb): Copy speculative edges.
8956 (redirect_all_calls): New function.
8957 (copy_cfg_body): Do redirection after loop info is updated.
8958 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
8960 2013-08-09 Jan Hubicka <jh@suse.cz>
8962 * lto-streamer-out.c (output_function): Renumber PHIs.
8963 * lto-streamer-in.c (input_function): Likewise.
8965 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8967 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8968 (get_lane_unsigned): Likewise.
8969 (dup_lane_scalar): Likewise.
8970 (get_lane): enable for VALL.
8971 * config/aarch64/aarch64-simd.md
8972 (aarch64_dup_lane_scalar<mode>): Remove.
8973 (aarch64_get_lane_signed<mode>): Likewise.
8974 (aarch64_get_lane_unsigned<mode>): Likewise.
8975 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8976 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8977 (aarch64_get_lane<mode>): Enable for all vector modes.
8978 (aarch64_get_lanedi): Remove misleading constraints.
8979 * config/aarch64/arm_neon.h
8980 (__aarch64_vget_lane_any): Define.
8981 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
8982 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
8983 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
8984 * config/aarch64/iterators.md (VDQQH): New.
8988 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
8990 * configure.ac: Add GAS check for LEON instructions on SPARC.
8991 * configure: Regenerate.
8992 * config.in: Likewise.
8993 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
8995 * config/sparc/sparc.opt (LEON, LEON3): New masks.
8996 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
8998 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
8999 (AS_LEON_FLAG): New macro.
9000 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9001 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9002 Deal with LEON and LEON3 for the memory model.
9003 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9004 (atomic_compare_and_swap<mode>_1): Likewise.
9005 (*atomic_compare_and_swap<mode>_1): Likewise.
9007 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9009 * config/arm/neon.md (vcond): Fix floating-point vector
9010 comparisons against 0.
9012 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9014 * lra-constraints.c (emit_spill_move): Remove assert.
9015 (process_alt_operands): Add more debugging
9016 output. Increase reject for spilling into memory. Decrease
9017 reject for reloading scratch.
9018 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9020 2013-08-08 Steve Ellcey <sellcey@mips.com>
9022 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9023 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9024 micromips incompatible. Add nan2008.
9025 (MULTILIB_DIRNAMES): Add nan2008.
9026 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9027 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9028 and micromips incompatible. Add nan2008.
9029 (MULTILIB_DIRNAMES): Add nan2008.
9030 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9032 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9034 PR rtl-optimization/58079
9035 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9036 simplify_comparison has widened a comparison with an integer.
9038 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9040 * config/arm/neon.md (movmisalign<mode>): Disable when we
9041 don't allow unaligned accesses.
9042 (*movmisalign<mode>_neon_store): Likewise.
9043 (*movmisalign<mode>_neon_load): Likewise.
9044 (*movmisalign<mode>_neon_store): Likewise.
9045 (*movmisalign<mode>_neon_load): Likewise.
9047 2013-08-08 Jan Hubicka <jh@suse.cz>
9049 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9050 (make_pass_rebuild_cgraph_edges): Also clear references.
9051 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9052 * ipa-inline-transform.c (inline_transform): Remove all references
9054 * cgraphunit.c (expand_function): Remove all references after
9056 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9057 (ipa_find_reference): Rewrite to iterator.
9058 (remove_stmt_references): Likewise.
9059 (ipa_clear_stmts_in_references): New function.
9060 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9061 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9063 * ipa-split.c (split_function): Remove references in split function.
9065 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9068 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9069 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9071 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9074 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9075 suggested mode for the assignment isn't compatible with the
9078 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9081 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9083 2013-08-07 Xinliang David Li <davidxl@google.com>
9085 * config/i386/i386.opt: New option -mtune-ctrl=.
9086 * config/i386/x86-tune.def: New file.
9087 * config/i386/i386.h: include x86-tune.def.
9088 * config/i386/i386.c (ix86_option_override_internal):
9089 Parsing -mtune-ctrl= option and set tune features.
9091 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9094 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9095 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9098 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9100 * rtl.h (update_alignments): Declare.
9101 * final.c (grow_label_align): New function extracted from...
9102 (shorten_branches): ...here. Call it.
9103 (update_alignments): New function.
9104 * reorg.c (sibling_labels): New variable.
9105 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9106 the new label along with it onto the sibling_labels vector.
9107 (fill_simple_delay_slots): Adjust call to get_label_before.
9108 (fill_slots_from_thread): Likewise.
9109 (relax_delay_slots): Likewise.
9110 (make_return_insns): Likewise.
9111 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9113 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9115 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9116 document its semantics.
9117 (diagnostic_report_diagnostic): Adjust accordingly.
9119 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9121 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9122 (sparc_option_override): ...and port to new C++ pass API.
9123 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9125 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9127 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9129 2013-08-06 Caroline Tice <cmtice@google.com>
9131 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9132 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9133 * tree-pass.h: Add pass_vtable_verify.
9134 * varasm.c (assemble_variable): Add code to properly set the comdat
9135 section and name for the .vtable_map_vars section.
9136 (assemble_vtyv_preinit_initializer): New function.
9137 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9139 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9140 * vtable-verify.c: New file.
9141 * vtable-verify.h: New file.
9142 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9143 initialiation levels.
9144 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9145 * passes.def: Insert pass_vtable_verify.
9146 * aclocal.m4: Reorder includes.
9147 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9148 -fvtv-counts options.
9149 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9150 as appropriate, if -fvtable-verify=... is used.
9151 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9152 -fvtable-verify=... is used.
9153 * Makefile.in (OBJS): Add vtable-verify.o to list.
9154 (vtable-verify.o): Add new build rule.
9155 (GTFILES): Add vtable-verify.c to list.
9156 * common.opt (fvtable-verify=): New flag.
9157 (vtv_priority): Values for fvtable-verify= flag.
9158 (fvtv-counts): New flag.
9159 (fvtv-debug): New flag.
9160 * tree.h (save_vtable_map_decl): New extern function decl.
9162 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9164 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9165 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9166 (pass_data_rl78_devirt): ...new pass_data instance and...
9167 (make_pass_rl78_devirt): ...new function.
9168 (rl78_asm_file_start): Port pass registration to new C++ API.
9170 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9172 * coretypes.h (rtl_opt_pass): Add.
9173 (gcc::context): Add.
9174 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9175 (epiphany_init): Port to new C++ pass API.
9176 (epiphany_optimize_mode_switching): Likewise.
9177 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9178 (pass_manager::get_pass_mode_switching): New.
9179 (pass_manager::get_pass_peephole2): New.
9180 * mode-switching.c (pass_mode_switching): Add clone method.
9181 * recog.c (pass_peephole2): Add clone method.
9182 (pass_split_all_insns): Add clone method.
9184 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9186 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9187 (mips_option_override): ...here, porting to new C++ API for passes.
9189 2013-08-06 Jan Hubicka <jh@suse.cz>
9191 * cgraph.c (cgraph_get_body): New function based on lto.c
9193 * cgraph.h (cgraph_get_body): Declare.
9194 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9196 * cgraphunit.c (expand_function): Get body prior expanding.
9197 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9198 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9200 * passes.c (do_per_function_toporder): Get body.
9201 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9202 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9204 2013-08-06 Martin Jambor <mjambor@suse.cz>
9207 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9208 is not re-finalized. Rename second parameter to no_collect.
9210 2013-08-06 Martin Jambor <mjambor@suse.cz>
9213 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9214 MEM_REF has proper alignment information.
9216 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9219 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9221 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9222 rtx (*) (rtx, ...) with insn_gen_fn.
9223 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9224 insn_gen_fn::stored_funcptr. Add initializer braces.
9226 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9228 Rewrite how instances of passes are cloned to remove assumptions
9229 about their sizes (thus allowing pass subclasses to have
9230 additional data fields, albeit non-GC-managed ones at this point).
9232 * passes.c (make_pass_instance): Now that passes have clone
9233 methods, rewrite this function to eliminate XNEW and memcpy
9234 calls that used hardcoded sizes. Since this function no longer
9235 creates pass instances, rename it to...
9236 (add_pass_instance): ...this. Document the old way that passes were
9237 numbered and flagged, and rework this function to continue using it.
9238 (next_pass_1): Add an initial_pass argument for use by
9240 (position_pass): When adding multiple instances of a pass, use
9241 the pass's clone method, rather than relying on the XNEW/memcpy
9242 within the former make_pass_instance (now add_pass_instance).
9243 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9244 the initial instance of the current pass within the pass manager.
9246 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9248 This is the automated part of the conversion of passes from C
9249 structs to C++ classes.
9251 Patch autogenerated by refactor_passes.py from
9252 https://github.com/davidmalcolm/gcc-refactoring-scripts
9253 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9255 * asan.c (pass_asan): Convert from a global struct to a subclass of
9256 gimple_opt_pass along with...
9257 (pass_data_asan): ...new pass_data instance and...
9258 (make_pass_asan): ...new function.
9259 (pass_asan_O0): Convert from a global struct to a subclass of
9260 gimple_opt_pass along with...
9261 (pass_data_asan_O0): ...new pass_data instance and...
9262 (make_pass_asan_O0): ...new function.
9263 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9264 subclass of rtl_opt_pass along with...
9265 (pass_data_inc_dec): ...new pass_data instance and...
9266 (make_pass_inc_dec): ...new function.
9267 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9268 a subclass of rtl_opt_pass along with...
9269 (pass_data_reorder_blocks): ...new pass_data instance and...
9270 (make_pass_reorder_blocks): ...new function.
9271 (pass_duplicate_computed_gotos): Convert from a global struct to a
9272 subclass of rtl_opt_pass along with...
9273 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9274 (make_pass_duplicate_computed_gotos): ...new function.
9275 (pass_partition_blocks): Convert from a global struct to a subclass of
9276 rtl_opt_pass along with...
9277 (pass_data_partition_blocks): ...new pass_data instance and...
9278 (make_pass_partition_blocks): ...new function.
9279 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9280 struct to a subclass of rtl_opt_pass along with...
9281 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9283 (make_pass_branch_target_load_optimize1): ...new function.
9284 (pass_branch_target_load_optimize2): Convert from a global struct to a
9285 subclass of rtl_opt_pass along with...
9286 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9288 (make_pass_branch_target_load_optimize2): ...new function.
9289 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9290 of rtl_opt_pass along with...
9291 (pass_data_jump): ...new pass_data instance and...
9292 (make_pass_jump): ...new function.
9293 (pass_jump2): Convert from a global struct to a subclass of
9294 rtl_opt_pass along with...
9295 (pass_data_jump2): ...new pass_data instance and...
9296 (make_pass_jump2): ...new function.
9297 * cfgexpand.c (pass_expand): Convert from a global struct to a
9298 subclass of rtl_opt_pass along with...
9299 (pass_data_expand): ...new pass_data instance and...
9300 (make_pass_expand): ...new function.
9301 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9302 of rtl_opt_pass along with...
9303 (pass_data_free_cfg): ...new pass_data instance and...
9304 (make_pass_free_cfg): ...new function.
9305 (pass_into_cfg_layout_mode): Convert from a global struct to a
9306 subclass of rtl_opt_pass along with...
9307 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9308 (make_pass_into_cfg_layout_mode): ...new function.
9309 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9310 subclass of rtl_opt_pass along with...
9311 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9312 (make_pass_outof_cfg_layout_mode): ...new function.
9313 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9314 struct to a subclass of gimple_opt_pass along with...
9315 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9316 (make_pass_build_cgraph_edges): ...new function.
9317 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9318 subclass of gimple_opt_pass along with...
9319 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9320 (make_pass_rebuild_cgraph_edges): ...new function.
9321 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9322 subclass of gimple_opt_pass along with...
9323 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9325 (make_pass_remove_cgraph_callee_edges): ...new function.
9326 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9327 struct to a subclass of rtl_opt_pass along with...
9328 (pass_data_stack_adjustments): ...new pass_data instance and...
9329 (make_pass_stack_adjustments): ...new function.
9330 * combine.c (pass_combine): Convert from a global struct to a subclass
9331 of rtl_opt_pass along with...
9332 (pass_data_combine): ...new pass_data instance and...
9333 (make_pass_combine): ...new function.
9334 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9335 global struct to a subclass of rtl_opt_pass along with...
9336 (pass_data_compare_elim_after_reload): ...new pass_data instance
9338 (make_pass_compare_elim_after_reload): ...new function.
9339 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9340 of rtl_opt_pass along with...
9341 (pass_data_rtl_cprop): ...new pass_data instance and...
9342 (make_pass_rtl_cprop): ...new function.
9343 * cse.c (pass_cse): Convert from a global struct to a subclass of
9344 rtl_opt_pass along with...
9345 (pass_data_cse): ...new pass_data instance and...
9346 (make_pass_cse): ...new function.
9347 (pass_cse2): Convert from a global struct to a subclass of
9348 rtl_opt_pass along with...
9349 (pass_data_cse2): ...new pass_data instance and...
9350 (make_pass_cse2): ...new function.
9351 (pass_cse_after_global_opts): Convert from a global struct to a
9352 subclass of rtl_opt_pass along with...
9353 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9354 (make_pass_cse_after_global_opts): ...new function.
9355 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9356 of rtl_opt_pass along with...
9357 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9358 (make_pass_ud_rtl_dce): ...new function.
9359 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9360 rtl_opt_pass along with...
9361 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9362 (make_pass_fast_rtl_dce): ...new function.
9363 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9364 a subclass of rtl_opt_pass along with...
9365 (pass_data_df_initialize_opt): ...new pass_data instance and...
9366 (make_pass_df_initialize_opt): ...new function.
9367 (pass_df_initialize_no_opt): Convert from a global struct to a
9368 subclass of rtl_opt_pass along with...
9369 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9370 (make_pass_df_initialize_no_opt): ...new function.
9371 (pass_df_finish): Convert from a global struct to a subclass of
9372 rtl_opt_pass along with...
9373 (pass_data_df_finish): ...new pass_data instance and...
9374 (make_pass_df_finish): ...new function.
9375 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9376 rtl_opt_pass along with...
9377 (pass_data_rtl_dse1): ...new pass_data instance and...
9378 (make_pass_rtl_dse1): ...new function.
9379 (pass_rtl_dse2): Convert from a global struct to a subclass of
9380 rtl_opt_pass along with...
9381 (pass_data_rtl_dse2): ...new pass_data instance and...
9382 (make_pass_rtl_dse2): ...new function.
9383 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9384 subclass of rtl_opt_pass along with...
9385 (pass_data_dwarf2_frame): ...new pass_data instance and...
9386 (make_pass_dwarf2_frame): ...new function.
9387 * except.c (pass_set_nothrow_function_flags): Convert from a global
9388 struct to a subclass of rtl_opt_pass along with...
9389 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9391 (make_pass_set_nothrow_function_flags): ...new function.
9392 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9393 subclass of rtl_opt_pass along with...
9394 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9396 (make_pass_convert_to_eh_region_ranges): ...new function.
9397 * final.c (pass_compute_alignments): Convert from a global struct to a
9398 subclass of rtl_opt_pass along with...
9399 (pass_data_compute_alignments): ...new pass_data instance and...
9400 (make_pass_compute_alignments): ...new function.
9401 (pass_final): Convert from a global struct to a subclass of
9402 rtl_opt_pass along with...
9403 (pass_data_final): ...new pass_data instance and...
9404 (make_pass_final): ...new function.
9405 (pass_shorten_branches): Convert from a global struct to a subclass of
9406 rtl_opt_pass along with...
9407 (pass_data_shorten_branches): ...new pass_data instance and...
9408 (make_pass_shorten_branches): ...new function.
9409 (pass_clean_state): Convert from a global struct to a subclass of
9410 rtl_opt_pass along with...
9411 (pass_data_clean_state): ...new pass_data instance and...
9412 (make_pass_clean_state): ...new function.
9413 * function.c (pass_instantiate_virtual_regs): Convert from a global
9414 struct to a subclass of rtl_opt_pass along with...
9415 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9416 (make_pass_instantiate_virtual_regs): ...new function.
9417 (pass_leaf_regs): Convert from a global struct to a subclass of
9418 rtl_opt_pass along with...
9419 (pass_data_leaf_regs): ...new pass_data instance and...
9420 (make_pass_leaf_regs): ...new function.
9421 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9422 subclass of rtl_opt_pass along with...
9423 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9425 (make_pass_thread_prologue_and_epilogue): ...new function.
9426 (pass_match_asm_constraints): Convert from a global struct to a
9427 subclass of rtl_opt_pass along with...
9428 (pass_data_match_asm_constraints): ...new pass_data instance and...
9429 (make_pass_match_asm_constraints): ...new function.
9430 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9431 subclass of rtl_opt_pass along with...
9432 (pass_data_rtl_fwprop): ...new pass_data instance and...
9433 (make_pass_rtl_fwprop): ...new function.
9434 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9435 rtl_opt_pass along with...
9436 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9437 (make_pass_rtl_fwprop_addr): ...new function.
9438 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9439 rtl_opt_pass along with...
9440 (pass_data_rtl_pre): ...new pass_data instance and...
9441 (make_pass_rtl_pre): ...new function.
9442 (pass_rtl_hoist): Convert from a global struct to a subclass of
9443 rtl_opt_pass along with...
9444 (pass_data_rtl_hoist): ...new pass_data instance and...
9445 (make_pass_rtl_hoist): ...new function.
9446 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9447 subclass of gimple_opt_pass along with...
9448 (pass_data_lower_cf): ...new pass_data instance and...
9449 (make_pass_lower_cf): ...new function.
9450 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9451 from a global struct to a subclass of gimple_opt_pass along with...
9452 (pass_data_strength_reduction): ...new pass_data instance and...
9453 (make_pass_strength_reduction): ...new function.
9454 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9455 of rtl_opt_pass along with...
9456 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9457 (make_pass_rtl_ifcvt): ...new function.
9458 (pass_if_after_combine): Convert from a global struct to a subclass of
9459 rtl_opt_pass along with...
9460 (pass_data_if_after_combine): ...new pass_data instance and...
9461 (make_pass_if_after_combine): ...new function.
9462 (pass_if_after_reload): Convert from a global struct to a subclass of
9463 rtl_opt_pass along with...
9464 (pass_data_if_after_reload): ...new pass_data instance and...
9465 (make_pass_if_after_reload): ...new function.
9466 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9467 a subclass of rtl_opt_pass along with...
9468 (pass_data_initialize_regs): ...new pass_data instance and...
9469 (make_pass_initialize_regs): ...new function.
9470 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9471 of ipa_opt_pass_d along with...
9472 (pass_data_ipa_cp): ...new pass_data instance and...
9473 (make_pass_ipa_cp): ...new function.
9474 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9475 global struct to a subclass of gimple_opt_pass along with...
9476 (pass_data_inline_parameters): ...new pass_data instance and...
9477 (make_pass_inline_parameters): ...new function.
9478 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9479 subclass of gimple_opt_pass along with...
9480 (pass_data_early_inline): ...new pass_data instance and...
9481 (make_pass_early_inline): ...new function.
9482 (pass_ipa_inline): Convert from a global struct to a subclass of
9483 ipa_opt_pass_d along with...
9484 (pass_data_ipa_inline): ...new pass_data instance and...
9485 (make_pass_ipa_inline): ...new function.
9486 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9487 struct to a subclass of gimple_opt_pass along with...
9488 (pass_data_local_pure_const): ...new pass_data instance and...
9489 (make_pass_local_pure_const): ...new function.
9490 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9491 ipa_opt_pass_d along with...
9492 (pass_data_ipa_pure_const): ...new pass_data instance and...
9493 (make_pass_ipa_pure_const): ...new function.
9494 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9495 to a subclass of ipa_opt_pass_d along with...
9496 (pass_data_ipa_reference): ...new pass_data instance and...
9497 (make_pass_ipa_reference): ...new function.
9498 * ipa-split.c (pass_split_functions): Convert from a global struct to
9499 a subclass of gimple_opt_pass along with...
9500 (pass_data_split_functions): ...new pass_data instance and...
9501 (make_pass_split_functions): ...new function.
9502 (pass_feedback_split_functions): Convert from a global struct to a
9503 subclass of gimple_opt_pass along with...
9504 (pass_data_feedback_split_functions): ...new pass_data instance and...
9505 (make_pass_feedback_split_functions): ...new function.
9506 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9507 global struct to a subclass of simple_ipa_opt_pass along with...
9508 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9510 (make_pass_ipa_function_and_variable_visibility): ...new function.
9511 (pass_ipa_free_inline_summary): Convert from a global struct to a
9512 subclass of simple_ipa_opt_pass along with...
9513 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9514 (make_pass_ipa_free_inline_summary): ...new function.
9515 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9516 subclass of ipa_opt_pass_d along with...
9517 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9519 (make_pass_ipa_whole_program_visibility): ...new function.
9520 (pass_ipa_profile): Convert from a global struct to a subclass of
9521 ipa_opt_pass_d along with...
9522 (pass_data_ipa_profile): ...new pass_data instance and...
9523 (make_pass_ipa_profile): ...new function.
9524 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9525 ipa_opt_pass_d along with...
9526 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9527 (make_pass_ipa_cdtor_merge): ...new function.
9528 * ira.c (pass_ira): Convert from a global struct to a subclass of
9529 rtl_opt_pass along with...
9530 (pass_data_ira): ...new pass_data instance and...
9531 (make_pass_ira): ...new function.
9532 (pass_reload): Convert from a global struct to a subclass of
9533 rtl_opt_pass along with...
9534 (pass_data_reload): ...new pass_data instance and...
9535 (make_pass_reload): ...new function.
9536 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9537 subclass of rtl_opt_pass along with...
9538 (pass_data_cleanup_barriers): ...new pass_data instance and...
9539 (make_pass_cleanup_barriers): ...new function.
9540 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9541 of rtl_opt_pass along with...
9542 (pass_data_loop2): ...new pass_data instance and...
9543 (make_pass_loop2): ...new function.
9544 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9545 rtl_opt_pass along with...
9546 (pass_data_rtl_loop_init): ...new pass_data instance and...
9547 (make_pass_rtl_loop_init): ...new function.
9548 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9549 rtl_opt_pass along with...
9550 (pass_data_rtl_loop_done): ...new pass_data instance and...
9551 (make_pass_rtl_loop_done): ...new function.
9552 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9553 subclass of rtl_opt_pass along with...
9554 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9555 (make_pass_rtl_move_loop_invariants): ...new function.
9556 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9557 rtl_opt_pass along with...
9558 (pass_data_rtl_unswitch): ...new pass_data instance and...
9559 (make_pass_rtl_unswitch): ...new function.
9560 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9561 subclass of rtl_opt_pass along with...
9562 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9564 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9565 (pass_rtl_doloop): Convert from a global struct to a subclass of
9566 rtl_opt_pass along with...
9567 (pass_data_rtl_doloop): ...new pass_data instance and...
9568 (make_pass_rtl_doloop): ...new function.
9569 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9570 a subclass of rtl_opt_pass along with...
9571 (pass_data_lower_subreg): ...new pass_data instance and...
9572 (make_pass_lower_subreg): ...new function.
9573 (pass_lower_subreg2): Convert from a global struct to a subclass of
9574 rtl_opt_pass along with...
9575 (pass_data_lower_subreg2): ...new pass_data instance and...
9576 (make_pass_lower_subreg2): ...new function.
9577 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9578 struct to a subclass of ipa_opt_pass_d along with...
9579 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9580 (make_pass_ipa_lto_gimple_out): ...new function.
9581 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9582 of ipa_opt_pass_d along with...
9583 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9584 (make_pass_ipa_lto_finish_out): ...new function.
9585 * mode-switching.c (pass_mode_switching): Convert from a global struct
9586 to a subclass of rtl_opt_pass along with...
9587 (pass_data_mode_switching): ...new pass_data instance and...
9588 (make_pass_mode_switching): ...new function.
9589 * modulo-sched.c (pass_sms): Convert from a global struct to a
9590 subclass of rtl_opt_pass along with...
9591 (pass_data_sms): ...new pass_data instance and...
9592 (make_pass_sms): ...new function.
9593 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9594 subclass of gimple_opt_pass along with...
9595 (pass_data_expand_omp): ...new pass_data instance and...
9596 (make_pass_expand_omp): ...new function.
9597 (pass_lower_omp): Convert from a global struct to a subclass of
9598 gimple_opt_pass along with...
9599 (pass_data_lower_omp): ...new pass_data instance and...
9600 (make_pass_lower_omp): ...new function.
9601 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9602 of gimple_opt_pass along with...
9603 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9604 (make_pass_diagnose_omp_blocks): ...new function.
9605 * passes.c (pass_early_local_passes): Convert from a global struct to
9606 a subclass of simple_ipa_opt_pass along with...
9607 (pass_data_early_local_passes): ...new pass_data instance and...
9608 (make_pass_early_local_passes): ...new function.
9609 (pass_all_early_optimizations): Convert from a global struct to a
9610 subclass of gimple_opt_pass along with...
9611 (pass_data_all_early_optimizations): ...new pass_data instance and...
9612 (make_pass_all_early_optimizations): ...new function.
9613 (pass_all_optimizations): Convert from a global struct to a subclass
9614 of gimple_opt_pass along with...
9615 (pass_data_all_optimizations): ...new pass_data instance and...
9616 (make_pass_all_optimizations): ...new function.
9617 (pass_all_optimizations_g): Convert from a global struct to a subclass
9618 of gimple_opt_pass along with...
9619 (pass_data_all_optimizations_g): ...new pass_data instance and...
9620 (make_pass_all_optimizations_g): ...new function.
9621 (pass_rest_of_compilation): Convert from a global struct to a subclass
9622 of rtl_opt_pass along with...
9623 (pass_data_rest_of_compilation): ...new pass_data instance and...
9624 (make_pass_rest_of_compilation): ...new function.
9625 (pass_postreload): Convert from a global struct to a subclass of
9626 rtl_opt_pass along with...
9627 (pass_data_postreload): ...new pass_data instance and...
9628 (make_pass_postreload): ...new function.
9629 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9630 subclass of rtl_opt_pass along with...
9631 (pass_data_gcse2): ...new pass_data instance and...
9632 (make_pass_gcse2): ...new function.
9633 * postreload.c (pass_postreload_cse): Convert from a global struct to
9634 a subclass of rtl_opt_pass along with...
9635 (pass_data_postreload_cse): ...new pass_data instance and...
9636 (make_pass_postreload_cse): ...new function.
9637 * predict.c (pass_profile): Convert from a global struct to a subclass
9638 of gimple_opt_pass along with...
9639 (pass_data_profile): ...new pass_data instance and...
9640 (make_pass_profile): ...new function.
9641 (pass_strip_predict_hints): Convert from a global struct to a subclass
9642 of gimple_opt_pass along with...
9643 (pass_data_strip_predict_hints): ...new pass_data instance and...
9644 (make_pass_strip_predict_hints): ...new function.
9645 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9646 of rtl_opt_pass along with...
9647 (pass_data_peephole2): ...new pass_data instance and...
9648 (make_pass_peephole2): ...new function.
9649 (pass_split_all_insns): Convert from a global struct to a subclass of
9650 rtl_opt_pass along with...
9651 (pass_data_split_all_insns): ...new pass_data instance and...
9652 (make_pass_split_all_insns): ...new function.
9653 (pass_split_after_reload): Convert from a global struct to a subclass
9654 of rtl_opt_pass along with...
9655 (pass_data_split_after_reload): ...new pass_data instance and...
9656 (make_pass_split_after_reload): ...new function.
9657 (pass_split_before_regstack): Convert from a global struct to a
9658 subclass of rtl_opt_pass along with...
9659 (pass_data_split_before_regstack): ...new pass_data instance and...
9660 (make_pass_split_before_regstack): ...new function.
9661 (pass_split_before_sched2): Convert from a global struct to a subclass
9662 of rtl_opt_pass along with...
9663 (pass_data_split_before_sched2): ...new pass_data instance and...
9664 (make_pass_split_before_sched2): ...new function.
9665 (pass_split_for_shorten_branches): Convert from a global struct to a
9666 subclass of rtl_opt_pass along with...
9667 (pass_data_split_for_shorten_branches): ...new pass_data instance
9669 (make_pass_split_for_shorten_branches): ...new function.
9670 * ree.c (pass_ree): Convert from a global struct to a subclass of
9671 rtl_opt_pass along with...
9672 (pass_data_ree): ...new pass_data instance and...
9673 (make_pass_ree): ...new function.
9674 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9675 subclass of rtl_opt_pass along with...
9676 (pass_data_stack_regs): ...new pass_data instance and...
9677 (make_pass_stack_regs): ...new function.
9678 (pass_stack_regs_run): Convert from a global struct to a subclass of
9679 rtl_opt_pass along with...
9680 (pass_data_stack_regs_run): ...new pass_data instance and...
9681 (make_pass_stack_regs_run): ...new function.
9682 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9683 subclass of rtl_opt_pass along with...
9684 (pass_data_cprop_hardreg): ...new pass_data instance and...
9685 (make_pass_cprop_hardreg): ...new function.
9686 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9687 subclass of rtl_opt_pass along with...
9688 (pass_data_reginfo_init): ...new pass_data instance and...
9689 (make_pass_reginfo_init): ...new function.
9690 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9691 of rtl_opt_pass along with...
9692 (pass_data_regmove): ...new pass_data instance and...
9693 (make_pass_regmove): ...new function.
9694 * regrename.c (pass_regrename): Convert from a global struct to a
9695 subclass of rtl_opt_pass along with...
9696 (pass_data_regrename): ...new pass_data instance and...
9697 (make_pass_regrename): ...new function.
9698 * reorg.c (pass_delay_slots): Convert from a global struct to a
9699 subclass of rtl_opt_pass along with...
9700 (pass_data_delay_slots): ...new pass_data instance and...
9701 (make_pass_delay_slots): ...new function.
9702 (pass_machine_reorg): Convert from a global struct to a subclass of
9703 rtl_opt_pass along with...
9704 (pass_data_machine_reorg): ...new pass_data instance and...
9705 (make_pass_machine_reorg): ...new function.
9706 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9707 of rtl_opt_pass along with...
9708 (pass_data_sched): ...new pass_data instance and...
9709 (make_pass_sched): ...new function.
9710 (pass_sched2): Convert from a global struct to a subclass of
9711 rtl_opt_pass along with...
9712 (pass_data_sched2): ...new pass_data instance and...
9713 (make_pass_sched2): ...new function.
9714 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9715 to a subclass of rtl_opt_pass along with...
9716 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9717 (make_pass_stack_ptr_mod): ...new function.
9718 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9719 to a subclass of rtl_opt_pass along with...
9720 (pass_data_rtl_store_motion): ...new pass_data instance and...
9721 (make_pass_rtl_store_motion): ...new function.
9722 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9723 of gimple_opt_pass along with...
9724 (pass_data_tracer): ...new pass_data instance and...
9725 (make_pass_tracer): ...new function.
9726 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9727 to a subclass of gimple_opt_pass along with...
9728 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9729 (make_pass_diagnose_tm_blocks): ...new function.
9730 (pass_lower_tm): Convert from a global struct to a subclass of
9731 gimple_opt_pass along with...
9732 (pass_data_lower_tm): ...new pass_data instance and...
9733 (make_pass_lower_tm): ...new function.
9734 (pass_tm_init): Convert from a global struct to a subclass of
9735 gimple_opt_pass along with...
9736 (pass_data_tm_init): ...new pass_data instance and...
9737 (make_pass_tm_init): ...new function.
9738 (pass_tm_mark): Convert from a global struct to a subclass of
9739 gimple_opt_pass along with...
9740 (pass_data_tm_mark): ...new pass_data instance and...
9741 (make_pass_tm_mark): ...new function.
9742 (pass_tm_edges): Convert from a global struct to a subclass of
9743 gimple_opt_pass along with...
9744 (pass_data_tm_edges): ...new pass_data instance and...
9745 (make_pass_tm_edges): ...new function.
9746 (pass_tm_memopt): Convert from a global struct to a subclass of
9747 gimple_opt_pass along with...
9748 (pass_data_tm_memopt): ...new pass_data instance and...
9749 (make_pass_tm_memopt): ...new function.
9750 (pass_ipa_tm): Convert from a global struct to a subclass of
9751 simple_ipa_opt_pass along with...
9752 (pass_data_ipa_tm): ...new pass_data instance and...
9753 (make_pass_ipa_tm): ...new function.
9754 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9755 subclass of gimple_opt_pass along with...
9756 (pass_data_call_cdce): ...new pass_data instance and...
9757 (make_pass_call_cdce): ...new function.
9758 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9759 subclass of gimple_opt_pass along with...
9760 (pass_data_build_cfg): ...new pass_data instance and...
9761 (make_pass_build_cfg): ...new function.
9762 (pass_split_crit_edges): Convert from a global struct to a subclass of
9763 gimple_opt_pass along with...
9764 (pass_data_split_crit_edges): ...new pass_data instance and...
9765 (make_pass_split_crit_edges): ...new function.
9766 (pass_warn_function_return): Convert from a global struct to a
9767 subclass of gimple_opt_pass along with...
9768 (pass_data_warn_function_return): ...new pass_data instance and...
9769 (make_pass_warn_function_return): ...new function.
9770 (pass_warn_function_noreturn): Convert from a global struct to a
9771 subclass of gimple_opt_pass along with...
9772 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9773 (make_pass_warn_function_noreturn): ...new function.
9774 (pass_warn_unused_result): Convert from a global struct to a subclass
9775 of gimple_opt_pass along with...
9776 (pass_data_warn_unused_result): ...new pass_data instance and...
9777 (make_pass_warn_unused_result): ...new function.
9778 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9779 a subclass of gimple_opt_pass along with...
9780 (pass_data_merge_phi): ...new pass_data instance and...
9781 (make_pass_merge_phi): ...new function.
9782 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9783 a subclass of gimple_opt_pass along with...
9784 (pass_data_lower_complex): ...new pass_data instance and...
9785 (make_pass_lower_complex): ...new function.
9786 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9787 gimple_opt_pass along with...
9788 (pass_data_lower_complex_O0): ...new pass_data instance and...
9789 (make_pass_lower_complex_O0): ...new function.
9790 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9791 subclass of gimple_opt_pass along with...
9792 (pass_data_lower_eh): ...new pass_data instance and...
9793 (make_pass_lower_eh): ...new function.
9794 (pass_refactor_eh): Convert from a global struct to a subclass of
9795 gimple_opt_pass along with...
9796 (pass_data_refactor_eh): ...new pass_data instance and...
9797 (make_pass_refactor_eh): ...new function.
9798 (pass_lower_resx): Convert from a global struct to a subclass of
9799 gimple_opt_pass along with...
9800 (pass_data_lower_resx): ...new pass_data instance and...
9801 (make_pass_lower_resx): ...new function.
9802 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9803 of gimple_opt_pass along with...
9804 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9805 (make_pass_lower_eh_dispatch): ...new function.
9806 (pass_cleanup_eh): Convert from a global struct to a subclass of
9807 gimple_opt_pass along with...
9808 (pass_data_cleanup_eh): ...new pass_data instance and...
9809 (make_pass_cleanup_eh): ...new function.
9810 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9811 to a subclass of simple_ipa_opt_pass along with...
9812 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9813 (make_pass_ipa_lower_emutls): ...new function.
9814 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9815 a subclass of gimple_opt_pass along with...
9816 (pass_data_if_conversion): ...new pass_data instance and...
9817 (make_pass_if_conversion): ...new function.
9818 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9819 subclass of gimple_opt_pass along with...
9820 (pass_data_build_ssa): ...new pass_data instance and...
9821 (make_pass_build_ssa): ...new function.
9822 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9823 global struct to a subclass of gimple_opt_pass along with...
9824 (pass_data_loop_distribution): ...new pass_data instance and...
9825 (make_pass_loop_distribution): ...new function.
9826 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9827 subclass of gimple_opt_pass along with...
9828 (pass_data_mudflap_1): ...new pass_data instance and...
9829 (make_pass_mudflap_1): ...new function.
9830 (pass_mudflap_2): Convert from a global struct to a subclass of
9831 gimple_opt_pass along with...
9832 (pass_data_mudflap_2): ...new pass_data instance and...
9833 (make_pass_mudflap_2): ...new function.
9834 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9835 subclass of gimple_opt_pass along with...
9836 (pass_data_mudflap_1): ...new pass_data instance and...
9837 (make_pass_mudflap_1): ...new function.
9838 (pass_mudflap_2): Convert from a global struct to a subclass of
9839 gimple_opt_pass along with...
9840 (pass_data_mudflap_2): ...new pass_data instance and...
9841 (make_pass_mudflap_2): ...new function.
9842 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9843 gimple_opt_pass along with...
9844 (pass_data_nrv): ...new pass_data instance and...
9845 (make_pass_nrv): ...new function.
9846 (pass_return_slot): Convert from a global struct to a subclass of
9847 gimple_opt_pass along with...
9848 (pass_data_return_slot): ...new pass_data instance and...
9849 (make_pass_return_slot): ...new function.
9850 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9851 to a subclass of gimple_opt_pass along with...
9852 (pass_data_object_sizes): ...new pass_data instance and...
9853 (make_pass_object_sizes): ...new function.
9854 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9855 global struct to a subclass of gimple_opt_pass along with...
9856 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9858 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9859 (pass_fixup_cfg): Convert from a global struct to a subclass of
9860 gimple_opt_pass along with...
9861 (pass_data_fixup_cfg): ...new pass_data instance and...
9862 (make_pass_fixup_cfg): ...new function.
9863 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9864 (make_pass_mudflap_1): ...new function.
9865 (pass_mudflap_2): Replace declaration with that of...
9866 (make_pass_mudflap_2): ...new function.
9867 (pass_asan): Replace declaration with that of...
9868 (make_pass_asan): ...new function.
9869 (pass_asan_O0): Replace declaration with that of...
9870 (make_pass_asan_O0): ...new function.
9871 (pass_tsan): Replace declaration with that of...
9872 (make_pass_tsan): ...new function.
9873 (pass_tsan_O0): Replace declaration with that of...
9874 (make_pass_tsan_O0): ...new function.
9875 (pass_lower_cf): Replace declaration with that of...
9876 (make_pass_lower_cf): ...new function.
9877 (pass_refactor_eh): Replace declaration with that of...
9878 (make_pass_refactor_eh): ...new function.
9879 (pass_lower_eh): Replace declaration with that of...
9880 (make_pass_lower_eh): ...new function.
9881 (pass_lower_eh_dispatch): Replace declaration with that of...
9882 (make_pass_lower_eh_dispatch): ...new function.
9883 (pass_lower_resx): Replace declaration with that of...
9884 (make_pass_lower_resx): ...new function.
9885 (pass_build_cfg): Replace declaration with that of...
9886 (make_pass_build_cfg): ...new function.
9887 (pass_early_tree_profile): Replace declaration with that of...
9888 (make_pass_early_tree_profile): ...new function.
9889 (pass_cleanup_eh): Replace declaration with that of...
9890 (make_pass_cleanup_eh): ...new function.
9891 (pass_sra): Replace declaration with that of...
9892 (make_pass_sra): ...new function.
9893 (pass_sra_early): Replace declaration with that of...
9894 (make_pass_sra_early): ...new function.
9895 (pass_early_ipa_sra): Replace declaration with that of...
9896 (make_pass_early_ipa_sra): ...new function.
9897 (pass_tail_recursion): Replace declaration with that of...
9898 (make_pass_tail_recursion): ...new function.
9899 (pass_tail_calls): Replace declaration with that of...
9900 (make_pass_tail_calls): ...new function.
9901 (pass_tree_loop): Replace declaration with that of...
9902 (make_pass_tree_loop): ...new function.
9903 (pass_tree_loop_init): Replace declaration with that of...
9904 (make_pass_tree_loop_init): ...new function.
9905 (pass_lim): Replace declaration with that of...
9906 (make_pass_lim): ...new function.
9907 (pass_tree_unswitch): Replace declaration with that of...
9908 (make_pass_tree_unswitch): ...new function.
9909 (pass_predcom): Replace declaration with that of...
9910 (make_pass_predcom): ...new function.
9911 (pass_iv_canon): Replace declaration with that of...
9912 (make_pass_iv_canon): ...new function.
9913 (pass_scev_cprop): Replace declaration with that of...
9914 (make_pass_scev_cprop): ...new function.
9915 (pass_empty_loop): Replace declaration with that of...
9916 (make_pass_empty_loop): ...new function.
9917 (pass_record_bounds): Replace declaration with that of...
9918 (make_pass_record_bounds): ...new function.
9919 (pass_graphite): Replace declaration with that of...
9920 (make_pass_graphite): ...new function.
9921 (pass_graphite_transforms): Replace declaration with that of...
9922 (make_pass_graphite_transforms): ...new function.
9923 (pass_if_conversion): Replace declaration with that of...
9924 (make_pass_if_conversion): ...new function.
9925 (pass_loop_distribution): Replace declaration with that of...
9926 (make_pass_loop_distribution): ...new function.
9927 (pass_vectorize): Replace declaration with that of...
9928 (make_pass_vectorize): ...new function.
9929 (pass_slp_vectorize): Replace declaration with that of...
9930 (make_pass_slp_vectorize): ...new function.
9931 (pass_complete_unroll): Replace declaration with that of...
9932 (make_pass_complete_unroll): ...new function.
9933 (pass_complete_unrolli): Replace declaration with that of...
9934 (make_pass_complete_unrolli): ...new function.
9935 (pass_parallelize_loops): Replace declaration with that of...
9936 (make_pass_parallelize_loops): ...new function.
9937 (pass_loop_prefetch): Replace declaration with that of...
9938 (make_pass_loop_prefetch): ...new function.
9939 (pass_iv_optimize): Replace declaration with that of...
9940 (make_pass_iv_optimize): ...new function.
9941 (pass_tree_loop_done): Replace declaration with that of...
9942 (make_pass_tree_loop_done): ...new function.
9943 (pass_ch): Replace declaration with that of...
9944 (make_pass_ch): ...new function.
9945 (pass_ccp): Replace declaration with that of...
9946 (make_pass_ccp): ...new function.
9947 (pass_phi_only_cprop): Replace declaration with that of...
9948 (make_pass_phi_only_cprop): ...new function.
9949 (pass_build_ssa): Replace declaration with that of...
9950 (make_pass_build_ssa): ...new function.
9951 (pass_build_alias): Replace declaration with that of...
9952 (make_pass_build_alias): ...new function.
9953 (pass_build_ealias): Replace declaration with that of...
9954 (make_pass_build_ealias): ...new function.
9955 (pass_dominator): Replace declaration with that of...
9956 (make_pass_dominator): ...new function.
9957 (pass_dce): Replace declaration with that of...
9958 (make_pass_dce): ...new function.
9959 (pass_dce_loop): Replace declaration with that of...
9960 (make_pass_dce_loop): ...new function.
9961 (pass_cd_dce): Replace declaration with that of...
9962 (make_pass_cd_dce): ...new function.
9963 (pass_call_cdce): Replace declaration with that of...
9964 (make_pass_call_cdce): ...new function.
9965 (pass_merge_phi): Replace declaration with that of...
9966 (make_pass_merge_phi): ...new function.
9967 (pass_split_crit_edges): Replace declaration with that of...
9968 (make_pass_split_crit_edges): ...new function.
9969 (pass_pre): Replace declaration with that of...
9970 (make_pass_pre): ...new function.
9971 (pass_profile): Replace declaration with that of...
9972 (make_pass_profile): ...new function.
9973 (pass_strip_predict_hints): Replace declaration with that of...
9974 (make_pass_strip_predict_hints): ...new function.
9975 (pass_lower_complex_O0): Replace declaration with that of...
9976 (make_pass_lower_complex_O0): ...new function.
9977 (pass_lower_complex): Replace declaration with that of...
9978 (make_pass_lower_complex): ...new function.
9979 (pass_lower_vector): Replace declaration with that of...
9980 (make_pass_lower_vector): ...new function.
9981 (pass_lower_vector_ssa): Replace declaration with that of...
9982 (make_pass_lower_vector_ssa): ...new function.
9983 (pass_lower_omp): Replace declaration with that of...
9984 (make_pass_lower_omp): ...new function.
9985 (pass_diagnose_omp_blocks): Replace declaration with that of...
9986 (make_pass_diagnose_omp_blocks): ...new function.
9987 (pass_expand_omp): Replace declaration with that of...
9988 (make_pass_expand_omp): ...new function.
9989 (pass_expand_omp_ssa): Replace declaration with that of...
9990 (make_pass_expand_omp_ssa): ...new function.
9991 (pass_object_sizes): Replace declaration with that of...
9992 (make_pass_object_sizes): ...new function.
9993 (pass_strlen): Replace declaration with that of...
9994 (make_pass_strlen): ...new function.
9995 (pass_fold_builtins): Replace declaration with that of...
9996 (make_pass_fold_builtins): ...new function.
9997 (pass_stdarg): Replace declaration with that of...
9998 (make_pass_stdarg): ...new function.
9999 (pass_early_warn_uninitialized): Replace declaration with that of...
10000 (make_pass_early_warn_uninitialized): ...new function.
10001 (pass_late_warn_uninitialized): Replace declaration with that of...
10002 (make_pass_late_warn_uninitialized): ...new function.
10003 (pass_cse_reciprocals): Replace declaration with that of...
10004 (make_pass_cse_reciprocals): ...new function.
10005 (pass_cse_sincos): Replace declaration with that of...
10006 (make_pass_cse_sincos): ...new function.
10007 (pass_optimize_bswap): Replace declaration with that of...
10008 (make_pass_optimize_bswap): ...new function.
10009 (pass_optimize_widening_mul): Replace declaration with that of...
10010 (make_pass_optimize_widening_mul): ...new function.
10011 (pass_warn_function_return): Replace declaration with that of...
10012 (make_pass_warn_function_return): ...new function.
10013 (pass_warn_function_noreturn): Replace declaration with that of...
10014 (make_pass_warn_function_noreturn): ...new function.
10015 (pass_cselim): Replace declaration with that of...
10016 (make_pass_cselim): ...new function.
10017 (pass_phiopt): Replace declaration with that of...
10018 (make_pass_phiopt): ...new function.
10019 (pass_forwprop): Replace declaration with that of...
10020 (make_pass_forwprop): ...new function.
10021 (pass_phiprop): Replace declaration with that of...
10022 (make_pass_phiprop): ...new function.
10023 (pass_tree_ifcombine): Replace declaration with that of...
10024 (make_pass_tree_ifcombine): ...new function.
10025 (pass_dse): Replace declaration with that of...
10026 (make_pass_dse): ...new function.
10027 (pass_nrv): Replace declaration with that of...
10028 (make_pass_nrv): ...new function.
10029 (pass_rename_ssa_copies): Replace declaration with that of...
10030 (make_pass_rename_ssa_copies): ...new function.
10031 (pass_sink_code): Replace declaration with that of...
10032 (make_pass_sink_code): ...new function.
10033 (pass_fre): Replace declaration with that of...
10034 (make_pass_fre): ...new function.
10035 (pass_check_data_deps): Replace declaration with that of...
10036 (make_pass_check_data_deps): ...new function.
10037 (pass_copy_prop): Replace declaration with that of...
10038 (make_pass_copy_prop): ...new function.
10039 (pass_vrp): Replace declaration with that of...
10040 (make_pass_vrp): ...new function.
10041 (pass_uncprop): Replace declaration with that of...
10042 (make_pass_uncprop): ...new function.
10043 (pass_return_slot): Replace declaration with that of...
10044 (make_pass_return_slot): ...new function.
10045 (pass_reassoc): Replace declaration with that of...
10046 (make_pass_reassoc): ...new function.
10047 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10048 (make_pass_rebuild_cgraph_edges): ...new function.
10049 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10050 (make_pass_remove_cgraph_callee_edges): ...new function.
10051 (pass_build_cgraph_edges): Replace declaration with that of...
10052 (make_pass_build_cgraph_edges): ...new function.
10053 (pass_local_pure_const): Replace declaration with that of...
10054 (make_pass_local_pure_const): ...new function.
10055 (pass_tracer): Replace declaration with that of...
10056 (make_pass_tracer): ...new function.
10057 (pass_warn_unused_result): Replace declaration with that of...
10058 (make_pass_warn_unused_result): ...new function.
10059 (pass_diagnose_tm_blocks): Replace declaration with that of...
10060 (make_pass_diagnose_tm_blocks): ...new function.
10061 (pass_lower_tm): Replace declaration with that of...
10062 (make_pass_lower_tm): ...new function.
10063 (pass_tm_init): Replace declaration with that of...
10064 (make_pass_tm_init): ...new function.
10065 (pass_tm_mark): Replace declaration with that of...
10066 (make_pass_tm_mark): ...new function.
10067 (pass_tm_memopt): Replace declaration with that of...
10068 (make_pass_tm_memopt): ...new function.
10069 (pass_tm_edges): Replace declaration with that of...
10070 (make_pass_tm_edges): ...new function.
10071 (pass_split_functions): Replace declaration with that of...
10072 (make_pass_split_functions): ...new function.
10073 (pass_feedback_split_functions): Replace declaration with that of...
10074 (make_pass_feedback_split_functions): ...new function.
10075 (pass_strength_reduction): Replace declaration with that of...
10076 (make_pass_strength_reduction): ...new function.
10077 (pass_ipa_lower_emutls): Replace declaration with that of...
10078 (make_pass_ipa_lower_emutls): ...new function.
10079 (pass_ipa_function_and_variable_visibility): Replace declaration with
10081 (make_pass_ipa_function_and_variable_visibility): ...new function.
10082 (pass_ipa_tree_profile): Replace declaration with that of...
10083 (make_pass_ipa_tree_profile): ...new function.
10084 (pass_early_local_passes): Replace declaration with that of...
10085 (make_pass_early_local_passes): ...new function.
10086 (pass_ipa_whole_program_visibility): Replace declaration with that
10088 (make_pass_ipa_whole_program_visibility): ...new function.
10089 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10090 (make_pass_ipa_lto_gimple_out): ...new function.
10091 (pass_ipa_increase_alignment): Replace declaration with that of...
10092 (make_pass_ipa_increase_alignment): ...new function.
10093 (pass_ipa_inline): Replace declaration with that of...
10094 (make_pass_ipa_inline): ...new function.
10095 (pass_ipa_free_lang_data): Replace declaration with that of...
10096 (make_pass_ipa_free_lang_data): ...new function.
10097 (pass_ipa_free_inline_summary): Replace declaration with that of...
10098 (make_pass_ipa_free_inline_summary): ...new function.
10099 (pass_ipa_cp): Replace declaration with that of...
10100 (make_pass_ipa_cp): ...new function.
10101 (pass_ipa_reference): Replace declaration with that of...
10102 (make_pass_ipa_reference): ...new function.
10103 (pass_ipa_pure_const): Replace declaration with that of...
10104 (make_pass_ipa_pure_const): ...new function.
10105 (pass_ipa_pta): Replace declaration with that of...
10106 (make_pass_ipa_pta): ...new function.
10107 (pass_ipa_lto_finish_out): Replace declaration with that of...
10108 (make_pass_ipa_lto_finish_out): ...new function.
10109 (pass_ipa_tm): Replace declaration with that of...
10110 (make_pass_ipa_tm): ...new function.
10111 (pass_ipa_profile): Replace declaration with that of...
10112 (make_pass_ipa_profile): ...new function.
10113 (pass_ipa_cdtor_merge): Replace declaration with that of...
10114 (make_pass_ipa_cdtor_merge): ...new function.
10115 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10117 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10118 (pass_init_datastructures): Replace declaration with that of...
10119 (make_pass_init_datastructures): ...new function.
10120 (pass_fixup_cfg): Replace declaration with that of...
10121 (make_pass_fixup_cfg): ...new function.
10122 (pass_expand): Replace declaration with that of...
10123 (make_pass_expand): ...new function.
10124 (pass_instantiate_virtual_regs): Replace declaration with that of...
10125 (make_pass_instantiate_virtual_regs): ...new function.
10126 (pass_rtl_fwprop): Replace declaration with that of...
10127 (make_pass_rtl_fwprop): ...new function.
10128 (pass_rtl_fwprop_addr): Replace declaration with that of...
10129 (make_pass_rtl_fwprop_addr): ...new function.
10130 (pass_jump): Replace declaration with that of...
10131 (make_pass_jump): ...new function.
10132 (pass_jump2): Replace declaration with that of...
10133 (make_pass_jump2): ...new function.
10134 (pass_lower_subreg): Replace declaration with that of...
10135 (make_pass_lower_subreg): ...new function.
10136 (pass_cse): Replace declaration with that of...
10137 (make_pass_cse): ...new function.
10138 (pass_fast_rtl_dce): Replace declaration with that of...
10139 (make_pass_fast_rtl_dce): ...new function.
10140 (pass_ud_rtl_dce): Replace declaration with that of...
10141 (make_pass_ud_rtl_dce): ...new function.
10142 (pass_rtl_dce): Replace declaration with that of...
10143 (make_pass_rtl_dce): ...new function.
10144 (pass_rtl_dse1): Replace declaration with that of...
10145 (make_pass_rtl_dse1): ...new function.
10146 (pass_rtl_dse2): Replace declaration with that of...
10147 (make_pass_rtl_dse2): ...new function.
10148 (pass_rtl_dse3): Replace declaration with that of...
10149 (make_pass_rtl_dse3): ...new function.
10150 (pass_rtl_cprop): Replace declaration with that of...
10151 (make_pass_rtl_cprop): ...new function.
10152 (pass_rtl_pre): Replace declaration with that of...
10153 (make_pass_rtl_pre): ...new function.
10154 (pass_rtl_hoist): Replace declaration with that of...
10155 (make_pass_rtl_hoist): ...new function.
10156 (pass_rtl_store_motion): Replace declaration with that of...
10157 (make_pass_rtl_store_motion): ...new function.
10158 (pass_cse_after_global_opts): Replace declaration with that of...
10159 (make_pass_cse_after_global_opts): ...new function.
10160 (pass_rtl_ifcvt): Replace declaration with that of...
10161 (make_pass_rtl_ifcvt): ...new function.
10162 (pass_into_cfg_layout_mode): Replace declaration with that of...
10163 (make_pass_into_cfg_layout_mode): ...new function.
10164 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10165 (make_pass_outof_cfg_layout_mode): ...new function.
10166 (pass_loop2): Replace declaration with that of...
10167 (make_pass_loop2): ...new function.
10168 (pass_rtl_loop_init): Replace declaration with that of...
10169 (make_pass_rtl_loop_init): ...new function.
10170 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10171 (make_pass_rtl_move_loop_invariants): ...new function.
10172 (pass_rtl_unswitch): Replace declaration with that of...
10173 (make_pass_rtl_unswitch): ...new function.
10174 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10175 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10176 (pass_rtl_doloop): Replace declaration with that of...
10177 (make_pass_rtl_doloop): ...new function.
10178 (pass_rtl_loop_done): Replace declaration with that of...
10179 (make_pass_rtl_loop_done): ...new function.
10180 (pass_web): Replace declaration with that of...
10181 (make_pass_web): ...new function.
10182 (pass_cse2): Replace declaration with that of...
10183 (make_pass_cse2): ...new function.
10184 (pass_df_initialize_opt): Replace declaration with that of...
10185 (make_pass_df_initialize_opt): ...new function.
10186 (pass_df_initialize_no_opt): Replace declaration with that of...
10187 (make_pass_df_initialize_no_opt): ...new function.
10188 (pass_reginfo_init): Replace declaration with that of...
10189 (make_pass_reginfo_init): ...new function.
10190 (pass_inc_dec): Replace declaration with that of...
10191 (make_pass_inc_dec): ...new function.
10192 (pass_stack_ptr_mod): Replace declaration with that of...
10193 (make_pass_stack_ptr_mod): ...new function.
10194 (pass_initialize_regs): Replace declaration with that of...
10195 (make_pass_initialize_regs): ...new function.
10196 (pass_combine): Replace declaration with that of...
10197 (make_pass_combine): ...new function.
10198 (pass_if_after_combine): Replace declaration with that of...
10199 (make_pass_if_after_combine): ...new function.
10200 (pass_ree): Replace declaration with that of...
10201 (make_pass_ree): ...new function.
10202 (pass_partition_blocks): Replace declaration with that of...
10203 (make_pass_partition_blocks): ...new function.
10204 (pass_match_asm_constraints): Replace declaration with that of...
10205 (make_pass_match_asm_constraints): ...new function.
10206 (pass_regmove): Replace declaration with that of...
10207 (make_pass_regmove): ...new function.
10208 (pass_split_all_insns): Replace declaration with that of...
10209 (make_pass_split_all_insns): ...new function.
10210 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10211 (make_pass_fast_rtl_byte_dce): ...new function.
10212 (pass_lower_subreg2): Replace declaration with that of...
10213 (make_pass_lower_subreg2): ...new function.
10214 (pass_mode_switching): Replace declaration with that of...
10215 (make_pass_mode_switching): ...new function.
10216 (pass_sms): Replace declaration with that of...
10217 (make_pass_sms): ...new function.
10218 (pass_sched): Replace declaration with that of...
10219 (make_pass_sched): ...new function.
10220 (pass_ira): Replace declaration with that of...
10221 (make_pass_ira): ...new function.
10222 (pass_reload): Replace declaration with that of...
10223 (make_pass_reload): ...new function.
10224 (pass_clean_state): Replace declaration with that of...
10225 (make_pass_clean_state): ...new function.
10226 (pass_branch_prob): Replace declaration with that of...
10227 (make_pass_branch_prob): ...new function.
10228 (pass_value_profile_transformations): Replace declaration with that
10230 (make_pass_value_profile_transformations): ...new function.
10231 (pass_postreload_cse): Replace declaration with that of...
10232 (make_pass_postreload_cse): ...new function.
10233 (pass_gcse2): Replace declaration with that of...
10234 (make_pass_gcse2): ...new function.
10235 (pass_split_after_reload): Replace declaration with that of...
10236 (make_pass_split_after_reload): ...new function.
10237 (pass_branch_target_load_optimize1): Replace declaration with that
10239 (make_pass_branch_target_load_optimize1): ...new function.
10240 (pass_thread_prologue_and_epilogue): Replace declaration with that
10242 (make_pass_thread_prologue_and_epilogue): ...new function.
10243 (pass_stack_adjustments): Replace declaration with that of...
10244 (make_pass_stack_adjustments): ...new function.
10245 (pass_peephole2): Replace declaration with that of...
10246 (make_pass_peephole2): ...new function.
10247 (pass_if_after_reload): Replace declaration with that of...
10248 (make_pass_if_after_reload): ...new function.
10249 (pass_regrename): Replace declaration with that of...
10250 (make_pass_regrename): ...new function.
10251 (pass_cprop_hardreg): Replace declaration with that of...
10252 (make_pass_cprop_hardreg): ...new function.
10253 (pass_reorder_blocks): Replace declaration with that of...
10254 (make_pass_reorder_blocks): ...new function.
10255 (pass_branch_target_load_optimize2): Replace declaration with that
10257 (make_pass_branch_target_load_optimize2): ...new function.
10258 (pass_leaf_regs): Replace declaration with that of...
10259 (make_pass_leaf_regs): ...new function.
10260 (pass_split_before_sched2): Replace declaration with that of...
10261 (make_pass_split_before_sched2): ...new function.
10262 (pass_compare_elim_after_reload): Replace declaration with that of...
10263 (make_pass_compare_elim_after_reload): ...new function.
10264 (pass_sched2): Replace declaration with that of...
10265 (make_pass_sched2): ...new function.
10266 (pass_stack_regs): Replace declaration with that of...
10267 (make_pass_stack_regs): ...new function.
10268 (pass_stack_regs_run): Replace declaration with that of...
10269 (make_pass_stack_regs_run): ...new function.
10270 (pass_df_finish): Replace declaration with that of...
10271 (make_pass_df_finish): ...new function.
10272 (pass_compute_alignments): Replace declaration with that of...
10273 (make_pass_compute_alignments): ...new function.
10274 (pass_duplicate_computed_gotos): Replace declaration with that of...
10275 (make_pass_duplicate_computed_gotos): ...new function.
10276 (pass_variable_tracking): Replace declaration with that of...
10277 (make_pass_variable_tracking): ...new function.
10278 (pass_free_cfg): Replace declaration with that of...
10279 (make_pass_free_cfg): ...new function.
10280 (pass_machine_reorg): Replace declaration with that of...
10281 (make_pass_machine_reorg): ...new function.
10282 (pass_cleanup_barriers): Replace declaration with that of...
10283 (make_pass_cleanup_barriers): ...new function.
10284 (pass_delay_slots): Replace declaration with that of...
10285 (make_pass_delay_slots): ...new function.
10286 (pass_split_for_shorten_branches): Replace declaration with that of...
10287 (make_pass_split_for_shorten_branches): ...new function.
10288 (pass_split_before_regstack): Replace declaration with that of...
10289 (make_pass_split_before_regstack): ...new function.
10290 (pass_convert_to_eh_region_ranges): Replace declaration with that
10292 (make_pass_convert_to_eh_region_ranges): ...new function.
10293 (pass_shorten_branches): Replace declaration with that of...
10294 (make_pass_shorten_branches): ...new function.
10295 (pass_set_nothrow_function_flags): Replace declaration with that of...
10296 (make_pass_set_nothrow_function_flags): ...new function.
10297 (pass_dwarf2_frame): Replace declaration with that of...
10298 (make_pass_dwarf2_frame): ...new function.
10299 (pass_final): Replace declaration with that of...
10300 (make_pass_final): ...new function.
10301 (pass_rtl_seqabstr): Replace declaration with that of...
10302 (make_pass_rtl_seqabstr): ...new function.
10303 (pass_release_ssa_names): Replace declaration with that of...
10304 (make_pass_release_ssa_names): ...new function.
10305 (pass_early_inline): Replace declaration with that of...
10306 (make_pass_early_inline): ...new function.
10307 (pass_inline_parameters): Replace declaration with that of...
10308 (make_pass_inline_parameters): ...new function.
10309 (pass_update_address_taken): Replace declaration with that of...
10310 (make_pass_update_address_taken): ...new function.
10311 (pass_convert_switch): Replace declaration with that of...
10312 (make_pass_convert_switch): ...new function.
10313 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10314 to a subclass of simple_ipa_opt_pass along with...
10315 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10316 (make_pass_ipa_tree_profile): ...new function.
10317 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10318 subclass of gimple_opt_pass along with...
10319 (pass_data_sra_early): ...new pass_data instance and...
10320 (make_pass_sra_early): ...new function.
10321 (pass_sra): Convert from a global struct to a subclass of
10322 gimple_opt_pass along with...
10323 (pass_data_sra): ...new pass_data instance and...
10324 (make_pass_sra): ...new function.
10325 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10326 gimple_opt_pass along with...
10327 (pass_data_early_ipa_sra): ...new pass_data instance and...
10328 (make_pass_early_ipa_sra): ...new function.
10329 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10330 subclass of gimple_opt_pass along with...
10331 (pass_data_ccp): ...new pass_data instance and...
10332 (make_pass_ccp): ...new function.
10333 (pass_fold_builtins): Convert from a global struct to a subclass of
10334 gimple_opt_pass along with...
10335 (pass_data_fold_builtins): ...new pass_data instance and...
10336 (make_pass_fold_builtins): ...new function.
10337 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10338 subclass of gimple_opt_pass along with...
10339 (pass_data_copy_prop): ...new pass_data instance and...
10340 (make_pass_copy_prop): ...new function.
10341 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10342 global struct to a subclass of gimple_opt_pass along with...
10343 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10344 (make_pass_rename_ssa_copies): ...new function.
10345 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10346 subclass of gimple_opt_pass along with...
10347 (pass_data_dce): ...new pass_data instance and...
10348 (make_pass_dce): ...new function.
10349 (pass_dce_loop): Convert from a global struct to a subclass of
10350 gimple_opt_pass along with...
10351 (pass_data_dce_loop): ...new pass_data instance and...
10352 (make_pass_dce_loop): ...new function.
10353 (pass_cd_dce): Convert from a global struct to a subclass of
10354 gimple_opt_pass along with...
10355 (pass_data_cd_dce): ...new pass_data instance and...
10356 (make_pass_cd_dce): ...new function.
10357 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10358 subclass of gimple_opt_pass along with...
10359 (pass_data_dominator): ...new pass_data instance and...
10360 (make_pass_dominator): ...new function.
10361 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10362 gimple_opt_pass along with...
10363 (pass_data_phi_only_cprop): ...new pass_data instance and...
10364 (make_pass_phi_only_cprop): ...new function.
10365 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10366 subclass of gimple_opt_pass along with...
10367 (pass_data_dse): ...new pass_data instance and...
10368 (make_pass_dse): ...new function.
10369 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10370 a subclass of gimple_opt_pass along with...
10371 (pass_data_forwprop): ...new pass_data instance and...
10372 (make_pass_forwprop): ...new function.
10373 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10374 struct to a subclass of gimple_opt_pass along with...
10375 (pass_data_tree_ifcombine): ...new pass_data instance and...
10376 (make_pass_tree_ifcombine): ...new function.
10377 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10378 subclass of gimple_opt_pass along with...
10379 (pass_data_ch): ...new pass_data instance and...
10380 (make_pass_ch): ...new function.
10381 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10382 subclass of gimple_opt_pass along with...
10383 (pass_data_tree_loop): ...new pass_data instance and...
10384 (make_pass_tree_loop): ...new function.
10385 (pass_tree_loop_init): Convert from a global struct to a subclass of
10386 gimple_opt_pass along with...
10387 (pass_data_tree_loop_init): ...new pass_data instance and...
10388 (make_pass_tree_loop_init): ...new function.
10389 (pass_lim): Convert from a global struct to a subclass of
10390 gimple_opt_pass along with...
10391 (pass_data_lim): ...new pass_data instance and...
10392 (make_pass_lim): ...new function.
10393 (pass_tree_unswitch): Convert from a global struct to a subclass of
10394 gimple_opt_pass along with...
10395 (pass_data_tree_unswitch): ...new pass_data instance and...
10396 (make_pass_tree_unswitch): ...new function.
10397 (pass_predcom): Convert from a global struct to a subclass of
10398 gimple_opt_pass along with...
10399 (pass_data_predcom): ...new pass_data instance and...
10400 (make_pass_predcom): ...new function.
10401 (pass_vectorize): Convert from a global struct to a subclass of
10402 gimple_opt_pass along with...
10403 (pass_data_vectorize): ...new pass_data instance and...
10404 (make_pass_vectorize): ...new function.
10405 (pass_graphite): Convert from a global struct to a subclass of
10406 gimple_opt_pass along with...
10407 (pass_data_graphite): ...new pass_data instance and...
10408 (make_pass_graphite): ...new function.
10409 (pass_graphite_transforms): Convert from a global struct to a subclass
10410 of gimple_opt_pass along with...
10411 (pass_data_graphite_transforms): ...new pass_data instance and...
10412 (make_pass_graphite_transforms): ...new function.
10413 (pass_check_data_deps): Convert from a global struct to a subclass of
10414 gimple_opt_pass along with...
10415 (pass_data_check_data_deps): ...new pass_data instance and...
10416 (make_pass_check_data_deps): ...new function.
10417 (pass_iv_canon): Convert from a global struct to a subclass of
10418 gimple_opt_pass along with...
10419 (pass_data_iv_canon): ...new pass_data instance and...
10420 (make_pass_iv_canon): ...new function.
10421 (pass_scev_cprop): Convert from a global struct to a subclass of
10422 gimple_opt_pass along with...
10423 (pass_data_scev_cprop): ...new pass_data instance and...
10424 (make_pass_scev_cprop): ...new function.
10425 (pass_record_bounds): Convert from a global struct to a subclass of
10426 gimple_opt_pass along with...
10427 (pass_data_record_bounds): ...new pass_data instance and...
10428 (make_pass_record_bounds): ...new function.
10429 (pass_complete_unroll): Convert from a global struct to a subclass of
10430 gimple_opt_pass along with...
10431 (pass_data_complete_unroll): ...new pass_data instance and...
10432 (make_pass_complete_unroll): ...new function.
10433 (pass_complete_unrolli): Convert from a global struct to a subclass of
10434 gimple_opt_pass along with...
10435 (pass_data_complete_unrolli): ...new pass_data instance and...
10436 (make_pass_complete_unrolli): ...new function.
10437 (pass_parallelize_loops): Convert from a global struct to a subclass
10438 of gimple_opt_pass along with...
10439 (pass_data_parallelize_loops): ...new pass_data instance and...
10440 (make_pass_parallelize_loops): ...new function.
10441 (pass_loop_prefetch): Convert from a global struct to a subclass of
10442 gimple_opt_pass along with...
10443 (pass_data_loop_prefetch): ...new pass_data instance and...
10444 (make_pass_loop_prefetch): ...new function.
10445 (pass_iv_optimize): Convert from a global struct to a subclass of
10446 gimple_opt_pass along with...
10447 (pass_data_iv_optimize): ...new pass_data instance and...
10448 (make_pass_iv_optimize): ...new function.
10449 (pass_tree_loop_done): Convert from a global struct to a subclass of
10450 gimple_opt_pass along with...
10451 (pass_data_tree_loop_done): ...new pass_data instance and...
10452 (make_pass_tree_loop_done): ...new function.
10453 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10454 struct to a subclass of gimple_opt_pass along with...
10455 (pass_data_cse_reciprocals): ...new pass_data instance and...
10456 (make_pass_cse_reciprocals): ...new function.
10457 (pass_cse_sincos): Convert from a global struct to a subclass of
10458 gimple_opt_pass along with...
10459 (pass_data_cse_sincos): ...new pass_data instance and...
10460 (make_pass_cse_sincos): ...new function.
10461 (pass_optimize_bswap): Convert from a global struct to a subclass of
10462 gimple_opt_pass along with...
10463 (pass_data_optimize_bswap): ...new pass_data instance and...
10464 (make_pass_optimize_bswap): ...new function.
10465 (pass_optimize_widening_mul): Convert from a global struct to a
10466 subclass of gimple_opt_pass along with...
10467 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10468 (make_pass_optimize_widening_mul): ...new function.
10469 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10470 subclass of gimple_opt_pass along with...
10471 (pass_data_phiopt): ...new pass_data instance and...
10472 (make_pass_phiopt): ...new function.
10473 (pass_cselim): Convert from a global struct to a subclass of
10474 gimple_opt_pass along with...
10475 (pass_data_cselim): ...new pass_data instance and...
10476 (make_pass_cselim): ...new function.
10477 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10478 subclass of gimple_opt_pass along with...
10479 (pass_data_phiprop): ...new pass_data instance and...
10480 (make_pass_phiprop): ...new function.
10481 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10482 subclass of gimple_opt_pass along with...
10483 (pass_data_pre): ...new pass_data instance and...
10484 (make_pass_pre): ...new function.
10485 (pass_fre): Convert from a global struct to a subclass of
10486 gimple_opt_pass along with...
10487 (pass_data_fre): ...new pass_data instance and...
10488 (make_pass_fre): ...new function.
10489 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10490 subclass of gimple_opt_pass along with...
10491 (pass_data_reassoc): ...new pass_data instance and...
10492 (make_pass_reassoc): ...new function.
10493 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10494 subclass of gimple_opt_pass along with...
10495 (pass_data_sink_code): ...new pass_data instance and...
10496 (make_pass_sink_code): ...new function.
10497 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10498 subclass of gimple_opt_pass along with...
10499 (pass_data_strlen): ...new pass_data instance and...
10500 (make_pass_strlen): ...new function.
10501 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10502 struct to a subclass of gimple_opt_pass along with...
10503 (pass_data_build_alias): ...new pass_data instance and...
10504 (make_pass_build_alias): ...new function.
10505 (pass_build_ealias): Convert from a global struct to a subclass of
10506 gimple_opt_pass along with...
10507 (pass_data_build_ealias): ...new pass_data instance and...
10508 (make_pass_build_ealias): ...new function.
10509 (pass_ipa_pta): Convert from a global struct to a subclass of
10510 simple_ipa_opt_pass along with...
10511 (pass_data_ipa_pta): ...new pass_data instance and...
10512 (make_pass_ipa_pta): ...new function.
10513 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10514 subclass of gimple_opt_pass along with...
10515 (pass_data_uncprop): ...new pass_data instance and...
10516 (make_pass_uncprop): ...new function.
10517 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10518 global struct to a subclass of gimple_opt_pass along with...
10519 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10520 (make_pass_late_warn_uninitialized): ...new function.
10521 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10522 to a subclass of gimple_opt_pass along with...
10523 (pass_data_init_datastructures): ...new pass_data instance and...
10524 (make_pass_init_datastructures): ...new function.
10525 (pass_early_warn_uninitialized): Convert from a global struct to a
10526 subclass of gimple_opt_pass along with...
10527 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10528 (make_pass_early_warn_uninitialized): ...new function.
10529 (pass_update_address_taken): Convert from a global struct to a
10530 subclass of gimple_opt_pass along with...
10531 (pass_data_update_address_taken): ...new pass_data instance and...
10532 (make_pass_update_address_taken): ...new function.
10533 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10534 struct to a subclass of gimple_opt_pass along with...
10535 (pass_data_release_ssa_names): ...new pass_data instance and...
10536 (make_pass_release_ssa_names): ...new function.
10537 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10538 subclass of gimple_opt_pass along with...
10539 (pass_data_stdarg): ...new pass_data instance and...
10540 (make_pass_stdarg): ...new function.
10541 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10542 global struct to a subclass of gimple_opt_pass along with...
10543 (pass_data_convert_switch): ...new pass_data instance and...
10544 (make_pass_convert_switch): ...new function.
10545 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10546 to a subclass of gimple_opt_pass along with...
10547 (pass_data_tail_recursion): ...new pass_data instance and...
10548 (make_pass_tail_recursion): ...new function.
10549 (pass_tail_calls): Convert from a global struct to a subclass of
10550 gimple_opt_pass along with...
10551 (pass_data_tail_calls): ...new pass_data instance and...
10552 (make_pass_tail_calls): ...new function.
10553 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10554 struct to a subclass of gimple_opt_pass along with...
10555 (pass_data_lower_vector): ...new pass_data instance and...
10556 (make_pass_lower_vector): ...new function.
10557 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10558 gimple_opt_pass along with...
10559 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10560 (make_pass_lower_vector_ssa): ...new function.
10561 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10562 to a subclass of gimple_opt_pass along with...
10563 (pass_data_slp_vectorize): ...new pass_data instance and...
10564 (make_pass_slp_vectorize): ...new function.
10565 (pass_ipa_increase_alignment): Convert from a global struct to a
10566 subclass of simple_ipa_opt_pass along with...
10567 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10568 (make_pass_ipa_increase_alignment): ...new function.
10569 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10570 gimple_opt_pass along with...
10571 (pass_data_vrp): ...new pass_data instance and...
10572 (make_pass_vrp): ...new function.
10573 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10574 subclass of simple_ipa_opt_pass along with...
10575 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10576 (make_pass_ipa_free_lang_data): ...new function.
10577 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10578 gimple_opt_pass along with...
10579 (pass_data_tsan): ...new pass_data instance and...
10580 (make_pass_tsan): ...new function.
10581 (pass_tsan_O0): Convert from a global struct to a subclass of
10582 gimple_opt_pass along with...
10583 (pass_data_tsan_O0): ...new pass_data instance and...
10584 (make_pass_tsan_O0): ...new function.
10585 * var-tracking.c (pass_variable_tracking): Convert from a global
10586 struct to a subclass of rtl_opt_pass along with...
10587 (pass_data_variable_tracking): ...new pass_data instance and...
10588 (make_pass_variable_tracking): ...new function.
10589 * web.c (pass_web): Convert from a global struct to a subclass of
10590 rtl_opt_pass along with...
10591 (pass_data_web): ...new pass_data instance and...
10592 (make_pass_web): ...new function.
10593 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10594 declaration with that of...
10595 (make_pass_mode_switch_use): ...new function.
10596 (pass_resolve_sw_modes): Replace declaration with that of...
10597 (make_pass_resolve_sw_modes): ...new function.
10598 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10599 from a global struct to a subclass of rtl_opt_pass along with...
10600 (pass_data_mode_switch_use): ...new pass_data instance and...
10601 (make_pass_mode_switch_use): ...new function.
10602 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10603 from a global struct to a subclass of rtl_opt_pass along with...
10604 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10605 (make_pass_resolve_sw_modes): ...new function.
10606 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10607 struct to a subclass of rtl_opt_pass along with...
10608 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10609 (make_pass_insert_vzeroupper): ...new function.
10610 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10611 global struct to a subclass of rtl_opt_pass along with...
10612 (pass_data_work_around_errata): ...new pass_data instance and...
10613 (make_pass_work_around_errata): ...new function.
10614 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10615 struct to a subclass of rtl_opt_pass along with...
10616 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10617 (make_pass_mips_machine_reorg2): ...new function.
10619 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10621 * passes.c (pass_manager::operator new): New.
10623 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10625 Handwritten part of conversion of passes to C++ classes.
10627 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10628 (toplev.o): Add dep on PASS_MANAGER_H.
10629 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10630 of early local pases to reflect this moving from a global to a
10631 member of gcc::pass_manager.
10632 (cgraph_add_new_function): Likewise.
10633 * lto-cgraph.c (lto_output_node): Update for conversion of
10634 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10635 * passes.c (opt_pass::clone): New.
10636 (opt_pass::gate): New.
10637 (opt_pass::execute): New.
10638 (opt_pass::opt_pass): New.
10639 (pass_manager::execute_early_local_passes): New.
10640 (pass_manager::execute_pass_mode_switching): new.
10641 (finish_optimization_passes): Convert to...
10642 (pass_manager::finish_optimization_passes): ...this.
10643 (finish_optimization_passes): Update for conversion of passes to
10645 (register_dump_files_1): Use has_gate since we cannot portably
10646 check a vtable entry against NULL.
10647 (dump_one_pass): Likewise.
10648 (ipa_write_summaries_2): Likewise.
10649 (ipa_write_optimization_summaries_1): Likewise.
10650 (ipa_read_summaries_1): Likewise.
10651 (ipa_read_optimization_summaries_1): Likewise.
10652 (execute_ipa_stmt_fixups): Likewise.
10653 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10654 pass-creation functions rather than wiring up globals, and
10655 storing the results in fields of pass_manager generated using
10656 pass-instances.def.
10657 (pass_manager::dump_profile_report): Update for conversion of
10658 passes to C++ classes.
10659 (pass_manager::execute_ipa_summary_passes): Likewise.
10660 (execute_one_ipa_transform_pass): Likewise.
10661 (execute_one_pass): Use has_gate and has_execute since we cannot
10662 portably check a vtable entry against NULL.
10663 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10664 (pass_manager): Use pass-instances.def to add fields for the
10665 various pass instances.
10666 * toplev.c (finalize): Update for move of
10667 finish_optimization_passes to a method of gcc::pass_manager.
10668 * toplev.h (finish_optimization_passes): Move to method of class
10670 * tree-pass.h (struct pass_data): New.
10671 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10672 (opt_pass::gate): Convert to virtual function.
10673 (opt_pass::~opt_pass): New.
10674 (opt_pass::clone): New.
10675 (opt_pass::execute): Convert to virtual function.
10676 (opt_pass::opt_pass): New.
10677 (opt_pass::ctxt_): new.
10678 (gimple_opt_pass): Convert to subclass of opt_pass.
10679 (gimple_opt_pass::gimple_opt_pass): New.
10680 (rtl_opt_pass): Convert to subclass of opt_pass.
10681 (rtl_opt_pass::rtl_opt_pass): New.
10682 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10683 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10684 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10685 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10686 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10687 invocation of pass_mode_switching to reflect this moving from a
10688 global to a member of gcc::pass_manager.
10689 (ix86_option_override): Rework how pass_insert_vzeroupper is
10690 added to the pass_manager to reflect autogenerated changes.
10691 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10693 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10695 PR rtl-optimization/57708
10696 * recog.c (peep2_find_free_register): Validate all regs in a
10699 2013-08-05 Jan Hubicka <jh@suse.cz>
10702 * cgraph.c (verify_cgraph_node): Accept local flags from other
10704 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10705 (function_and_variable_visibility): Likewise.
10706 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10708 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10710 * graph.c (init_graph_slim_pretty_print): Remove.
10711 (print_graph_cfg): Do not call it. Use local pretty printer.
10712 (start_graph_dump): Likewise.
10714 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10716 * gimple-pretty-print.c (buffer): Remove.
10717 (initialized): Likewise.
10718 (maybe_init_pretty_print): Likewise.
10719 (print_gimple_stmt): Do not call it. Use non-static local
10720 pretty_printer variable.
10721 (print_gimple_expr): Likewise.
10722 (print_gimple_seq): Likewise.
10723 (gimple_dump_bb): Likewise.
10725 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10727 * asan.c (asan_pp): Remove.
10728 (asan_pp_initialized): Likewise.
10729 (asan_pp_initialize): Likewise.
10730 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10731 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10732 (asan_add_global): Likewise.
10734 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10736 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10737 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10738 * pretty-print.c (pp_formatted_text_data): Likewise.
10739 (pp_indent): Rename from pp_base_indent.
10740 (pp_format): Rename from pp_base_format.
10741 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10742 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10743 (pp_flush): Rename from pp_base_flush.
10744 (pp_set_line_maximum_length): Rename from
10745 pp_base_set_line_maximum_length.
10746 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10747 (pp_set_prefix): Rename from pp_base_set_prefix.
10748 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10749 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10750 (pp_append_text): Rename from pp_base_append_text.
10751 (pp_formatted_text): Rename from pp_base_formatted_text.
10752 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10753 (pp_remaining_character_count_for_line): Rename from
10754 pp_base_remaining_character_count_for_line.
10755 (pp_newline): Rename from pp_base_newline.
10756 (pp_character): Rename from pp_base_character.
10757 (pp_string): Rename from pp_base_string.
10758 (pp_maybe_space): Rename from pp_base_maybe_space.
10759 * asan.c (asan_pp_string): Adjust.
10760 (asan_emit_stack_protection): Likewise.
10761 (asan_add_global): Likewise.
10762 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10763 * tree-mudflap.c (mf_varname_tree): Likewise.
10764 * tree-pretty-print.c (pp_tree_identifier): Rename from
10765 pp_base_tree_identifier.
10766 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10767 Declare as function.
10769 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10771 * pretty-print.h (pp_bar_bar): New.
10772 (pp_ampersand_ampersand): Likewise.
10773 (pp_less_equal): Likewise.
10774 (pp_greater_equal): Likewise.
10775 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10776 printer functions instead of pp_string or operators and punctuators.
10777 (dump_gimple_call): Likewise.
10778 (dump_gimple_omp_for): Likewise.
10779 (dump_gimple_transaction): Likewise.
10780 (dump_gimple_phi): Likewise.
10781 (pp_gimple_stmt_1): Likewise.
10782 * sched-vis.c (print_insn): Likewise.
10783 * tree-mudflap.c (mf_varname_tree): Likewise.
10784 * tree-pretty-print.c (dump_block_node): Likewise.
10785 (dump_generic_node): Likewise.
10787 2013-08-02 Jan Hubicka <jh@suse.cz>
10789 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10791 * lto-streamer-out.c (tree_is_indexable): Results decls and
10792 parm decls are not indexable.
10793 (DFS_write_tree_body): Do not follow args and results.
10794 (hash_tree): Likewise.
10795 (output_functions): Rearrange so struct function is needed
10796 only when real body is output; be able to also ouptut abstract
10797 functions; output DECL_ARGUMENTS and DECL_RESULT.
10798 (lto_output): When not in WPA, ale store abstract functions.
10799 (write_symbol): Do not care about RESULT_DECL.
10800 (output_symbol_p): Handle correctly sbtract decls.
10801 * lto-streamer-in.c (input_function): Rearrange so struct
10802 function can be NULL at entry; allow streaming of
10803 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10804 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10805 sanity check during LTO.
10806 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10807 RESULT_DECl and DECL_ARGUMENTS.
10808 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10811 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10813 * pretty-print.h (pp_underscore): New.
10815 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10816 printer functions instead of pp_character.
10817 (dump_binary_rhs): Likewise.
10818 (dump_ternary_rhs): Likewise.
10819 (dump_gimple_call_args): Likewise.
10820 (pp_points_to_solution): Likewise.
10821 (dump_gimple_call): Likewise.
10822 (dump_gimple_switch): Likewise.
10823 (dump_gimple_cond): Likewise.
10824 (dump_gimple_bind): Likewise.
10825 (dump_gimple_try): Likewise.
10826 (dump_gimple_omp_for): Likewise.
10827 (dump_gimple_omp_continue): Likewise.
10828 (dump_gimple_omp_single): Likewise.
10829 (dump_gimple_omp_sections): Likewise.
10830 (dump_gimple_omp_block): Likewise.
10831 (dump_gimple_omp_critical): Likewise.
10832 (dump_gimple_transaction): Likewise.
10833 (dump_gimple_asm): Likewise.
10834 (dump_gimple_phi): Likewise.
10835 (dump_gimple_omp_parallel): Likewise.
10836 (dump_gimple_omp_task): Likewise.
10837 (dump_gimple_omp_atomic_load): Likewise.
10838 (dump_gimple_omp_atomic_store): Likewise.
10839 (dump_gimple_mem_ops): Likewise.
10840 (pp_gimple_stmt_1): Likewise.
10841 (pp_cfg_jump): Likewise.
10842 (dump_implicit_edges): Likewise.
10843 (gimple_dump_bb_for_graph): Likewise.
10844 * graph.c (draw_cfg_node): Likewise.
10845 * langhooks.c (lhd_print_error_function): Likewise.
10846 * sched-vis.c (print_exp): Likewise.
10847 (print_value): Likewise.
10848 (print_pattern): Likewise.
10849 (print_insn): Likewise.
10850 (rtl_dump_bb_for_graph): Likewise.
10851 * tree-pretty-print.c (dump_function_declaration): Likewise.
10852 (dump_array_domain): Likewise.
10853 (dump_omp_clause): Likewise.
10854 (dump_location): Likewise.
10855 (dump_generic_node): Likewise.
10856 (print_struct_decl): Likewise.
10857 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10859 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10861 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10862 candidate table when replacing a candidate statement.
10863 (replace_rhs_if_not_dup): Likewise.
10864 (replace_one_candidate): Likewise.
10866 2013-08-02 Jan Hubicka <jh@suse.cz>
10867 Martin Liska <marxin.liska@gmail.com>
10869 * cgraphunit.c (add_new_function): Fix logic when adding from
10871 (assemble_thunk): Rename to ...
10872 (expand_thunk); .. this one; export; get it working with
10873 general functions; make produced gimple valid.
10874 * cgraph.h (expand_thunk): Declare.
10876 2013-08-02 Jan Hubicka <jh@suse.cz>
10878 * ipa-cp.c (gather_context_independent_values): Use
10879 ipa_get_param_move_cost.
10880 (get_replacement_map): Remove PARAM; move parameter folding
10882 (create_specialized_node): Update.
10883 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10884 assert that we have gimple body; update move_cost.
10885 (count_formal_params): Assert that we have gimple body.
10886 (ipa_dump_param): New function.
10887 (ipa_alloc_node_params): Break out from ...
10888 (ipa_initialize_node_params): ... here.
10889 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10890 (ipa_write_node_info): Stream move costs.
10891 (ipa_read_node_info): Read move costs.
10892 (ipa_update_after_lto_read): Do not recompute node params.
10893 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10894 (ipa_get_param): Check we are not in WPA.
10895 (ipa_get_param_move_cost): New.
10896 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10897 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10898 parm numbers to be present.
10900 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10902 PR rtl-optimization/58048
10903 * lra-constraints.c (process_alt_operands): Don't check asm
10904 operand on register.
10906 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10908 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10909 the implied StoreLoad barrier for atomic operations if before.
10911 2013-08-02 Jan Hubicka <jh@suse.cz>
10912 Martin Liska <marxin.liska@gmail.com>
10914 * cgraph.c (cgraph_function_body_availability): Do not check
10916 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10917 symtab_node_availability): Declare.
10918 * ipa.c (can_replace_by_local_alias): New.
10919 (function_and_variable_visibility): Use it.
10920 * symtab.c (symtab_for_node_and_aliases,
10921 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
10923 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10925 PR rtl-optimization/57963
10926 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
10927 (lra_constraints): Use them.
10929 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10931 * config/arm/types.md (define_attr "type"): Add "load_acq"
10933 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10935 (cortex_a53_store1): Likewise.
10937 2013-08-01 Jan Hubicka <jh@suse.cz>
10939 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10940 partitions are not needed.
10942 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
10944 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10945 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10946 MAYBE_NON_Q_CLASS_P where appropriate.
10948 2013-08-01 Jan Hubicka <jh@suse.cz>
10950 * cgraph.h (release_function_body): Declare.
10951 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10952 of unused delcarations.
10954 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10956 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10957 RTL form when subtracting a constant.
10959 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10961 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10962 Generate canonical plus rtx with negated immediate instead of minus
10964 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10966 2013-08-01 Jan Hubicka <jh@suse.cz>
10968 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10969 (cgraph_release_function_body): Likewise.
10970 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10971 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10972 to used_as_abstract_origin.
10973 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10974 symbols used as abstract origins.
10975 * cgraphunit.c (analyze_functions): Update.
10976 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10977 used_as_abstract_origin.
10978 * tree-inline.c (tree_function_versioning): Update
10979 used_as_abstract_origin; be ready for DECL_RESULT and
10980 DECL_ARGUMENTS to be NULL.
10982 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
10983 for abstract functions.
10984 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
10987 2013-08-01 Jan Hubicka <jh@suse.cz>
10989 * profile.c (compute_value_histograms): Fix thinko.
10991 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10993 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
10994 aarch-common-protos.h to extra_headers.
10995 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10996 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
10997 * config/aarch64/t-aarch64 (aarch-common.o): Define.
10999 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11001 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11002 Include "../arm/types.md". Define "type" attribute for all patterns.
11003 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11006 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11008 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11009 to support power8 load fusion.
11010 (fusion_gpr_mem_load): Likewise.
11012 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11014 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11015 declarations for power8 load fusion.
11016 (emit_fusion_gpr_load): Likewise.
11018 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11019 tuning for power8, turn on fusion mode by default. Turn on sign
11020 extending fusion mode if normal fusion mode is on, and we are at
11022 (fusion_gpr_load_p): New function, return true if we can fuse an
11023 addis instruction with a dependent load to a GPR.
11024 (emit_fusion_gpr_load): Emit the instructions for power8 load
11027 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11028 (VSX load fusion peepholes): New peepholes to fuse together an
11029 addi instruction with a VSX load instruction.
11031 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11032 peepholes to fuse an addis instruction with a load to a GPR base
11033 register. If we are supporting sign extending fusions, convert
11034 sign extending loads to zero extending loads and add an explicit
11037 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11039 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11040 aarch-common-protos.h to extra_headers.
11041 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11042 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11043 (arm_early_store_addr_dep): Likewise.
11044 (arm_no_early_alu_shift_dep): Likewise.
11045 (arm_no_early_alu_shift_value_dep): Likewise.
11046 (arm_no_early_mul_dep): Likewise.
11047 (arm_no_early_store_addr_dep): Likewise.
11048 (arm_mac_accumulator_is_mul_result): Likewise.
11049 (arm_mac_accumulator_is_result): Likewise.
11050 * config/arm/aarch-common.c: ... here. New file.
11051 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11053 (arm_early_store_addr_dep): Likewise.
11054 (arm_no_early_alu_shift_dep): Likewise.
11055 (arm_no_early_alu_shift_value_dep): Likewise.
11056 (arm_no_early_mul_dep): Likewise.
11057 (arm_no_early_store_addr_dep): Likewise.
11058 (arm_mac_accumulator_is_mul_result): Likewise.
11059 (arm_mac_accumulator_is_result): Likewise.
11060 * config/arm/aarch-common-protos.h: ... here. New file.
11061 * config/arm/t-arm (aarch-common.o): Define.
11063 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11065 * config/arm/arm.md: Include new file "types.md".
11066 (define_attr "type"): Move from here to ...
11067 (define_attr "mul32"): Likewise.
11068 (define_attr "mul64"): Likewise.
11069 * config/arm/types.md: ... here. New file.
11071 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11073 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11074 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11076 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11078 * gen-pass-instances.awk: Fix offset of substr().
11080 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11082 * Makefile.in (pass-instances.def): New.
11083 (passes.o): Replace dependency on passes.def with one on
11086 * gen-pass-instances.awk: New.
11088 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11089 rather than passes.def, updating local definition of NEXT_PASS
11090 macro to add an extra NUM parameter (currently unused).
11092 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11094 * Makefile.in (PASS_MANAGER_H): New.
11095 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11096 (passes.o): Likewise.
11097 (statistics.o): Likewise.
11098 (cgraphunit.o): Likewise.
11099 (context.o): Depend on PASS_MANAGER_H.
11101 * pass_manager.h: New.
11103 * cgraphunit.c (cgraph_add_new_function): Update for moves
11104 of globals to fields of pass_manager.
11105 (analyze_function): Likewise.
11106 (expand_function): Likewise.
11107 (ipa_passes): Likewise.
11108 (compile): Likewise.
11110 * context.c (context::context): New.
11111 * context.h (context::context): New.
11112 (context::get_passes): New.
11113 (context::passes_): New.
11115 * lto-cgraph.c (input_node): Update for moves of globals to
11116 fields of pass_manager.
11118 * passes.c (all_passes): Remove, in favor of a field of the
11119 same name within the new class pass_manager.
11120 (all_small_ipa_passes): Likewise.
11121 (all_lowering_passes): Likewise.
11122 (all_regular_ipa_passes): Likewise.
11123 (all_late_ipa_passes): Likewise.
11124 (all_lto_gen_passes): Likewise.
11125 (passes_by_id): Likewise.
11126 (passes_by_id_size): Likewise.
11127 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11128 the new class pass_manager.
11129 (set_pass_for_id): Convert to...
11130 (pass_manager::set_pass_for_id): ...method.
11131 (get_pass_for_id): Convert to...
11132 (pass_manager::get_pass_for_id): ...method.
11133 (register_one_dump_file): Move body of implementation into...
11134 (pass_manager::register_one_dump_file): ...here.
11135 (register_dump_files_1): Convert to...
11136 (pass_manager::register_dump_files_1): ...method.
11137 (register_dump_files): Convert to...
11138 (pass_manager::register_dump_files): ...method.
11139 (create_pass_tab): Update for moves of globals to fields of
11141 (dump_passes): Move body of implementation into...
11142 (pass_manager::dump_passes): ...here.
11143 (register_pass): Move body of implementation into...
11144 (pass_manager::register_pass): ...here.
11145 (init_optimization_passes): Convert into...
11146 (pass_manager::pass_manager): ...constructor for new
11147 pass_manager class, and initialize the pass_lists array.
11148 (check_profile_consistency): Update for moves of globals to
11149 fields of pass_manager.
11150 (dump_profile_report): Move body of implementation into...
11151 (pass_manager::dump_profile_report): ...here.
11152 (ipa_write_summaries_1): Update for moves of pass lists from
11153 being globals to fields of pass_manager.
11154 (ipa_write_optimization_summaries): Likewise.
11155 (ipa_read_summaries): Likewise.
11156 (ipa_read_optimization_summaries): Likewise.
11157 (execute_all_ipa_stmt_fixups): Likewise.
11159 * statistics.c (statistics_fini): Update for moves of globals to
11160 fields of pass_manager.
11162 * toplev.c (general_init): Replace call to
11163 init_optimization_passes with construction of the pass_manager
11166 * tree-pass.h (all_passes): Remove, in favor of a field of the
11167 same name within the new class pass_manager.
11168 (all_small_ipa_passes): Likewise.
11169 (all_lowering_passes): Likewise.
11170 (all_regular_ipa_passes): Likewise.
11171 (all_lto_gen_passes): Likewise.
11172 (all_late_ipa_passes): Likewise.
11173 (passes_by_id): Likewise.
11174 (passes_by_id_size): Likewise.
11175 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11176 the new class pass_manager.
11177 (get_pass_for_id): Remove.
11179 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11181 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11184 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11186 * arm.md (mulhi3): New expand pattern.
11188 2013-07-30 Jan Hubicka <jh@suse.cz>
11189 Martin Liska <marxin.liska@gmail.com>
11191 * profile.c (compute_value_histograms): Do not ICE when
11192 there is mismatch only on some counters.
11194 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11196 PR rtl-optimization/57637
11197 * function.c (move_insn_for_shrink_wrap): Also check the
11198 GEN set of the LIVE problem for the liveness analysis
11199 if it exists, otherwise give up.
11201 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11203 PR tree-optimization/57993
11204 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11205 replaced statement in the candidate table.
11206 (phi_add_costs): Return infinite cost when the hidden basis does
11207 not dominate all phis on which the candidate is dependent.
11208 (replace_one_candidate): Record replaced statement in the
11211 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11213 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11214 (ashlv2si3): New expander.
11215 (*ashlv2si3_i): New define_insn_and_split.
11216 * predicates.md (float_operation): Allow patterns with three
11217 basic sub-patterns.
11219 PR rtl-optimization/58021
11220 * mode-switching.c (create_pre_exit): Always split off preceding
11221 insns if we are not at the basic block head.
11223 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11225 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11226 (UCLIBC_DYNAMIC_LINKER): New macro.
11227 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11229 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11230 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11232 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11233 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11234 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11235 for SF and DF modes. Use ieee_quad_format for TF mode.
11236 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11237 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11238 (mips_option_override): Handle `-mnan=legacy'.
11239 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11240 `-mabs=2008' and `-mnan=2008'.
11241 (OPTION_DEFAULT_SPECS): Add "nan" default.
11242 (ASM_SPEC): Handle `-mnan='.
11243 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11244 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11245 comment accordingly.
11246 (neg<mode>2): Likewise.
11247 * config/mips/mips.opt (mabs, mnan): New options.
11248 * doc/install.texi (Configuration): Document `--with-nan=' option.
11249 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11251 (MIPS Options): Document them.
11252 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11253 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11254 * configure: Regenerate.
11255 * config.in: Regenerate.
11257 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11259 * config/i386/i386.md (float post-reload splitters): Do not check
11260 for subregs of SSE registers.
11262 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11263 H.J. Lu <hongjiu.lu@intel.com>
11267 * config/i386/i386.md (post-reload splitter
11268 to avoid partial SSE reg dependency stalls): New pattern.
11270 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11272 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11274 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11276 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11277 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11279 2013-07-26 Cary Coutant <ccoutant@google.com>
11281 * dwarf2out.c (die_checksum_ordered): Don't include template
11282 instantiations in signature.
11283 (is_template_parameter): New function.
11284 (is_template_instantiation): New function.
11285 (generate_skeleton_bottom_up): Don't include template instantiations
11287 (generate_skeleton): Likewise.
11288 (break_out_comdat_types): Move recursive call to break out nested
11290 (prune_unused_types_mark_generic_parms_dies): Call
11291 is_template_parameter.
11293 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11295 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11296 uses vector registers.
11297 * config/aarch64/iterators.md: Add attributes rtn and vas.
11299 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11300 Richard Earnshaw <richard.earnshaw@arm.com>
11302 * combine.c (simplify_comparison): Re-canonicalize operands
11304 * config/arm/arm.md (movcond_addsi): New splitter.
11306 2013-07-25 Sterling Augustine <saugustine@google.com>
11308 * dwarf2out.c (size_of_pubnames): Move code to...
11309 (include_pubname_in_output): ...here. New.
11310 (want_pubnames): Rearrange.
11311 (output_pubnames): Call include_pubname_in_output. Move assertion.
11313 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11315 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11317 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11320 * doc/extend.texi: Remove obsolete builtins. Fix
11321 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11323 2013-07-25 Jan Hubicka <jh@suse.cz>
11325 * cgraph.c (release_function_body): Break out from ...
11326 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11327 and DECL_ARGUMENTS.
11328 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11329 old_tree in the map.
11330 (create_specialized_node): Update.
11331 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11333 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11334 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11335 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11336 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11339 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11341 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11342 addsi3_carryin_alt2_<optab>): Correct output template.
11344 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11346 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11347 Adjust for arm_restrict_it.
11348 Remove trailing whitespace.
11350 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11352 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11353 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11355 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11357 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11359 PR rtl-optimization/57960
11360 * lra-constraints.c (process_alt_operands): Use the right mode
11361 when checking strict_low.
11363 2013-07-25 Jan Hubicka <jh@suse.cz>
11365 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11366 * cgraph.c (cgraph_remove_node): Do not release function body
11367 when in cgraph streaming.
11368 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11369 in other partitions are not considered reachable; fix handling of
11372 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11374 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11376 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11381 * config/arm/arm.md ("*sibcall_insn): Replace use of
11382 Ss with US. Adjust output for v5 and v4t.
11383 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11385 * config/arm/constraints.md ("Ss"): Rename to US.
11387 2013-07-25 Terry Guo <terry.guo@arm.com>
11389 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11390 shift_add/shift_sub0/shift_sub1 RTXs.
11392 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11393 Anton Blanchard <anton@au1.ibm.com>
11395 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11396 (altivec_vpks<VI_char>ss): Likewise.
11397 (altivec_vpks<VI_char>us): Likewise.
11398 (altivec_vpku<VI_char>us): Likewise.
11399 (altivec_vpku<VI_char>um): Likewise.
11401 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11403 Introduce context class.
11405 * Makefile.in (CONTEXT_H): New.
11406 (OBJS): Add context.o.
11407 (toplev.o): Add CONTEXT_H to dependencies.
11410 * toplev.c (general_init): Create the singleton gcc::context instance.
11416 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11418 PR rtl-optimization/57968
11419 * mode-switching.c (create_pre_exit): Allow instructions that
11420 don't set a return register to need a non-exit mode.
11422 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11423 Anton Blanchard <anton@au1.ibm.com>
11425 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11426 operands to vperm for little endian.
11427 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11428 of lvsl to create the control mask for a vperm for little endian.
11430 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11431 Anton Blanchard <anton@au1.ibm.com>
11433 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11434 two operands for little-endian.
11436 2013-07-23 Steve Ellcey <sellcey@mips.com>
11438 * config/mips/mips.c (mips_case_values_threshold): New.
11439 (TARGET_CASE_VALUES_THRESHOLD): Define.
11441 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11442 Anton Blanchard <anton@au1.ibm.com>
11444 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11445 selection of field for vector splat in little endian mode.
11447 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11449 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11450 expanders to rs6000.md.
11451 (ior<mode>3): Likewise.
11452 (and<mode>3): Likewise.
11453 (one_cmpl<mode>2): Likewise.
11454 (nor<mode>3): Likewise.
11455 (andc<mode>3): Likewise.
11456 (eqv<mode>3): Likewise.
11457 (nand<mode>3): Likewise.
11458 (orc<mode>3): Likewise.
11460 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11463 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11464 to split multi-word logical operations.
11465 (rs6000_split_logical_di): Likewise.
11466 (rs6000_split_logical): Likewise.
11468 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11469 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11470 and allow TImode operations in 32-bit.
11471 (vsx_and<mode>3_64bit): Likewise.
11472 (vsx_ior<mode>3_32bit): Likewise.
11473 (vsx_ior<mode>3_64bit): Likewise.
11474 (vsx_xor<mode>3_32bit): Likewise.
11475 (vsx_xor<mode>3_64bit): Likewise.
11476 (vsx_one_cmpl<mode>2_32bit): Likewise.
11477 (vsx_one_cmpl<mode>2_64bit): Likewise.
11478 (vsx_nor<mode>3_32bit): Likewise.
11479 (vsx_nor<mode>3_64bit): Likewise.
11480 (vsx_andc<mode>3_32bit): Likewise.
11481 (vsx_andc<mode>3_64bit): Likewise.
11482 (vsx_eqv<mode>3_32bit): Likewise.
11483 (vsx_eqv<mode>3_64bit): Likewise.
11484 (vsx_nand<mode>3_32bit): Likewise.
11485 (vsx_nand<mode>3_64bit): Likewise.
11486 (vsx_orc<mode>3_32bit): Likewise.
11487 (vsx_orc<mode>3_64bit): Likewise.
11489 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11490 logical types in GPRs.
11492 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11493 logical insns to rs6000.md, and allow TImode operations in
11495 (altivec_ior<mode>3): Likewise.
11496 (altivec_xor<mode>3): Likewise.
11497 (altivec_one_cmpl<mode>2): Likewise.
11498 (altivec_nor<mode>3): Likewise.
11499 (altivec_andc<mode>3): Likewise.
11501 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11502 attributes for moving the 128-bit logical operations into
11504 (BOOL_REGS_OUTPUT): Likewise.
11505 (BOOL_REGS_OP1): Likewise.
11506 (BOOL_REGS_OP2): Likewise.
11507 (BOOL_REGS_UNARY): Likewise.
11508 (BOOL_REGS_AND_CR0): Likewise.
11509 (one_cmpl<mode>2): Add support for DI logical operations on
11510 32-bit, splitting the operations to 32-bit.
11511 (anddi3): Likewise.
11512 (iordi3): Likewise.
11513 (xordi3): Likewise.
11514 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11515 changes to combine the 32/64-bit code, allow logical operations on
11516 TI mode in 32-bit, and to use similar match_operator patterns like
11517 scalar mode uses. Combine the Altivec and VSX code for logical
11518 operations, and move it here.
11519 (ior<mode>3, 128-bit types): Likewise.
11520 (xor<mode>3, 128-bit types): Likewise.
11521 (one_cmpl<mode>3, 128-bit types): Likewise.
11522 (nor<mode>3, 128-bit types): Likewise.
11523 (andc<mode>3, 128-bit types): Likewise.
11524 (eqv<mode>3, 128-bit types): Likewise.
11525 (nand<mode>3, 128-bit types): Likewise.
11526 (orc<mode>3, 128-bit types): Likewise.
11527 (and<mode>3_internal): Likewise.
11528 (bool<mode>3_internal): Likewise.
11529 (boolc<mode>3_internal1): Likewise.
11530 (boolc<mode>3_internal2): Likewise.
11531 (boolcc<mode>3_internal1): Likewise.
11532 (boolcc<mode>3_internal2): Likewise.
11533 (eqv<mode>3_internal1): Likewise.
11534 (eqv<mode>3_internal2): Likewise.
11535 (one_cmpl1<mode>3_internal): Likewise.
11537 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11539 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11540 Rename flag_stack_usage to flag_stack_usage_info.
11542 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11544 * config/microblaze/sync.md: New file.
11545 * config/microblaze/microblaze.md: Include sync.md
11546 * config/microblaze/microblaze.c: Add print_operand 'y'.
11547 * config/microblaze/constraints.md: Add memory_contraint
11548 'Q' which is a single register.
11550 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11552 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11554 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11556 * reload.c (find_reloads): Exit loop once we find this operand
11557 cannot be reloaded somehow for this alternative.
11559 * reload.c (find_reloads): Exit loop once we find a hard register.
11561 * rtlanal.c (computed_jump_p): Exit loop once we find label
11564 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11566 * cfgloopmanip.c (remove_path): Exit loop after setting
11569 * gensupport.c (subst_dup): Avoid loop if code is not
11570 MATCH_DUP nor MATCH_OP_DUP.
11572 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11574 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11576 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11578 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11579 true for SP_REGNUM if mode == ptr_mode.
11580 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11581 with value R0_REGNUM + 31.
11583 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11585 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11586 pad pointer-typed argument downward.
11588 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11590 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11591 and __ILP32__ when the ILP32 model is in use.
11593 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11595 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11596 (aarch64_load_symref_appropriately): In the case of
11597 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11598 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11599 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11600 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11601 if 'mode' doesn't equal to 'ptr_mode'.
11602 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11603 'vcall_offset'; change to call aarch64_emit_move differently depending
11604 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11605 to calculate the upper bound of 'vcall_offset'.
11606 (aarch64_cannot_force_const_mem): Change to also return true if
11608 (aarch64_legitimize_reload_address): In the case of large
11609 displacements, add new local variable 'xmode' and an assertion
11610 based on it; change to use 'xmode' to generate the new rtx and
11612 (aarch64_asm_trampoline_template): Change to generate the template
11613 differently depending on TARGET_ILP32 or not; change to use
11614 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11615 (aarch64_trampoline_size): Removed.
11616 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11617 and replace immediate literals with it. Change to use 'ptr_mode'
11618 instead of 'DImode' and call convert_memory_address if the mode
11619 of 'fnaddr' doesn't equal to 'ptr_mode'.
11620 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11622 (aarch64_elf_asm_destructor): Likewise.
11623 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11624 on TARGET_ILP32 instead of aarch64_trampoline_size.
11625 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11626 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11627 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11628 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11629 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11630 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11631 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11632 depending on the value of 'mode'.
11633 (add_losym_<mode>): New.
11634 (ldr_got_small_<mode>): New, based on ldr_got_small.
11635 (ldr_got_small): Remove.
11636 (ldr_got_small_sidi): New.
11637 * config/aarch64/iterators.md (P): New.
11638 (PTR): Change to 'ptr_mode' in the condition.
11640 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11642 * config.gcc (aarch64*-*-*): Support --with-abi.
11643 (aarch64*-*-elf): Support --with-multilib-list.
11644 (aarch64*-*-linux*): Likewise.
11645 (supported_defaults): Add abi to aarch64*-*-*.
11646 * configure.ac: Mention AArch64 for --with-multilib-list.
11647 * configure: Re-generated.
11648 * config/aarch64/biarchilp32.h: New file.
11649 * config/aarch64/biarchlp64.h: New file.
11650 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11652 (MULTILIB_DEFAULTS): Ditto.
11653 (DRIVER_SELF_SPECS): Ditto.
11654 (ASM_SPEC): Update to also substitute -mabi.
11655 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11656 file whose name depends on -mabi= and -mbig-endian.
11657 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11659 (POINTER_SIZE): New define.
11660 (POINTERS_EXTEND_UNSIGNED): Ditto.
11661 (enum aarch64_abi_type): New enumeration tag.
11662 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11663 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11664 (TARGET_ILP32): New define.
11665 * config/aarch64/aarch64.opt (mabi): New.
11666 (aarch64_abi): New.
11667 (ilp32, lp64): New values for -mabi.
11668 * config/aarch64/t-aarch64 (comma): New define.
11669 (MULTILIB_OPTIONS): Ditto.
11670 (MULTILIB_DIRNAMES): Ditto.
11671 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11672 * doc/invoke.texi: Document -mabi for AArch64.
11674 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11676 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11678 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11679 Anton Blanchard <anton@au1.ibm.com>
11681 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11682 endianness when selecting field to splat.
11684 2013-07-22 Eric Christopher <echristo@gmail.com>
11686 * dwarf2out.c (die_odr_checksum): New function to use
11687 CHECKSUM_ macros and ULEB128 for DIE tag.
11688 (generate_type_signature): Use.
11690 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11692 * config.gcc (sparc*-*-*): Accept leon3 processor.
11693 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11694 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11695 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11696 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11697 (mfix-ut699): Adjust comment.
11698 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11699 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11700 (CPP_CPU_SPEC): Likewise.
11701 (ASM_CPU_SPEC): Likewise.
11702 * config/sparc/sparc.c (leon3_cost): New constant.
11703 (sparc_option_override): Add leon3 support.
11704 (mem_ref): New function.
11705 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11706 (sparc_do_work_around_errata): Look into the instruction in the delay
11707 slot and adjust accordingly. Add fix for the data cache nullify issues
11708 of the UT699. Change insertion position for the NOP.
11709 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11710 (leon3_load): New reservation.
11711 (leon_store): Bump latency to 2.
11712 (grfpu): New automaton.
11713 (grfpu_alu): New unit.
11714 (grfpu_ds): Likewise.
11715 (leon_fp_alu): Adjust.
11716 (leon_fp_mult): Delete.
11717 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11718 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11719 * config/sparc/sparc.md (cpu): Add leon3.
11720 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11721 (swapsi): Likewise.
11722 (atomic_test_and_set): Likewise.
11723 (ldstub): Likewise.
11725 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11727 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11728 default for R5900 targets.
11729 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11730 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11731 * config/mips/mips.c (mips_option_override): Report an error for
11732 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11733 for -march=r5900 -mhard-float.
11735 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11737 * df-problems.c (can_move_insns_across): Exit loop once we
11738 find a non-fixed, non-global register.
11740 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11743 * omega.c (omega_eliminate_red): Break after setting red_found.
11744 (omega_problem_has_red_equations): Similarly after setting found.
11745 (omega_query_variable): Similarly after setting coupled.
11747 2013-07-22 Marek Polacek <polacek@redhat.com>
11749 * gimplify.c: Don't include gimple.h twice.
11751 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11753 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11754 instead of TARGET_THUMB1.
11755 (Pz): New constraint.
11756 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11758 (compare_negsi_si): Likewise.
11759 (compare_addsi2_op0): Likewise.
11760 (compare_addsi2_op1): Likewise.
11761 (addsi3_carryin_<optab>): Likewise.
11762 (addsi3_carryin_alt2_<optab>): Likewise.
11763 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11764 for arm_restrict_it.
11765 (subsi3_carryin): Likewise.
11766 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11767 (minmax_arithsi): Disable for arm_restrict_it.
11768 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11769 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11770 (satsi_<SAT:code>_shift): Likewise.
11771 (arm_shiftsi3): Add alternative for 16-bit encoding.
11772 (arm32_movhf): Disable for arm_restrict_it.
11773 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11774 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11776 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11778 * config/arm/arm.md (attribute "insn"): Delete.
11779 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11780 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11781 (not_shiftsi): Update for attribute change.
11782 (not_shiftsi_compare0): Likewise.
11783 (not_shiftsi_compare0_scratch): Likewise.
11784 (arm_one_cmplsi2): Likewise.
11785 (thumb1_one_cmplsi2): Likewise.
11786 (notsi_compare0): Likewise.
11787 (notsi_compare0_scratch): Likewise.
11788 (thumb1_movdi_insn): Likewise.
11789 (arm_movsi_insn): Likewise.
11790 (movhi_insn_arch4): Likewise.
11791 (movhi_bytes): Likewise.
11792 (arm_movqi_insn): Likewise.
11793 (thumb1_movqi_insn): Likewise.
11794 (arm32_movhf): Likewise.
11795 (thumb1_movhf): Likewise.
11796 (arm_movsf_soft_insn): Likewise.
11797 (thumb1_movsf_insn): Likewise.
11798 (thumb_movdf_insn): Likewise.
11799 (movsicc_insn): Likewise.
11800 (movsfcc_soft_insn): Likewise.
11801 (and_scc): Likewise.
11802 (cond_move): Likewise.
11803 (if_move_not): Likewise.
11804 (if_not_move): Likewise.
11805 (if_shift_move): Likewise.
11806 (if_move_shift): Likewise.
11807 (if_shift_shift): Likewise.
11808 (if_not_arith): Likewise.
11809 (if_arith_not): Likewise.
11810 (cond_move_not): Likewise.
11811 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11812 (neon_mov<mode>): Likewise.
11813 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11814 (thumb2_movsi_vfp): Likewise.
11815 (movsf_vfp): Likewise.
11816 (thumb2_movsf_vfp): Likewise.
11817 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11819 (cortexa7_older_only): Likewise.
11820 (cortexa7_younger): Likewise.
11821 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11822 (1020alu_shift_op): Likewise.
11823 (1020alu_shift_reg_op): Likewise.
11824 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11825 (alu_shift_op): Likewise.
11826 (alu_shift_reg_op): Likewise.
11827 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11828 (11_alu_shift_op): Likewise.
11829 (11_alu_shift_reg_op): Likewise.
11830 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11831 (9_alu_shift_reg_op): Likewise.
11832 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11834 (cortex_a15_alu_shift): Likewise.
11835 (cortex_a15_alu_shift_reg): Likewise.
11836 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11837 (cortex_a5_alu_shift): Likewise.
11838 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11840 (cortex_a53_alu_shift): Likewise.
11841 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11843 (cortex_a7_alu_reg): Likewise.
11844 (cortex_a7_alu_shift): Likewise.
11845 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11846 (cortex_a8_alu_shift): Likewise.
11847 (cortex_a8_alu_shift_reg): Likewise.
11848 (cortex_a8_mov): Likewise.
11849 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11850 (cortex_a9_dp_shift): Likewise.
11851 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11852 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11853 (cortex_r4_mov): Likewise.
11854 (cortex_r4_alu_shift): Likewise.
11855 (cortex_r4_alu_shift_reg): Likewise.
11856 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11857 (526_alu_shift_op): Likewise.
11858 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11859 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11860 (626te_alu_shift_op): Likewise.
11861 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11862 (726te_alu_op): Likewise.
11863 (726te_alu_shift_op): Likewise.
11864 (726te_alu_shift_reg_op): Likewise.
11865 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11866 (mp626_alu_shift_op): Likewise.
11867 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11868 (pj4_alu_e1_conds): Likewise.
11869 (pj4_alu): Likewise.
11870 (pj4_alu_conds): Likewise.
11871 (pj4_shift): Likewise.
11872 (pj4_shift_conds): Likewise.
11873 (pj4_alu_shift): Likewise.
11874 (pj4_alu_shift_conds): Likewise.
11876 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11878 * config/arm/predicates.md (shiftable_operator_strict_it):
11880 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11881 Disable cond_exec version for arm_restrict_it.
11882 (thumb2_smaxsi3): Convert to generate cond_exec.
11883 (thumb2_sminsi3): Likewise.
11884 (thumb32_umaxsi3): Likewise.
11885 (thumb2_uminsi3): Likewise.
11886 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11887 (thumb2_neg_abssi2): Likewise.
11888 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11889 (thumb2_movsicc_insn): Adjust alternatives.
11890 (thumb2_mov_negscc): Disable for arm_restrict_it.
11891 (thumb2_mov_negscc_strict_it): New pattern.
11892 (thumb2_mov_notscc_strict_it): New pattern.
11893 (thumb2_mov_notscc): Disable for arm_restrict_it.
11894 (thumb2_ior_scc): Likewise.
11895 (thumb2_ior_scc_strict_it): New pattern.
11896 (thumb2_cond_move): Adjust for arm_restrict_it.
11897 (thumb2_cond_arith): Disable for arm_restrict_it.
11898 (thumb2_cond_arith_strict_it): New pattern.
11899 (thumb2_cond_sub): Adjust for arm_restrict_it.
11900 (thumb2_movcond): Likewise.
11901 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11902 (thumb2_zero_extendhisi2_v6): Likewise.
11903 (thumb2_zero_extendqisi2_v6): Likewise.
11904 (orsi_notsi_si): Likewise.
11905 (orsi_not_shiftsi_si): Likewise.
11907 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
11909 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11910 instruction sequence is 1 byte shorter.
11912 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
11914 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11915 it is not needed after split.
11917 2013-07-20 Iain Sandoe <iain@codesourcery.com>
11920 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11921 second label for nonlocal goto receivers. Don't output pic base labels
11922 unless we're producing PIC; mark that action unreachable().
11923 (ix86_save_reg): If the function contains a nonlocal label, save the
11925 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
11926 * config/darwin.c (emitted_pic_label_num): New GTY.
11927 (update_pic_label_number_if_needed): New.
11928 (machopic_output_function_base_name): Adjust for nonlocal receiver
11930 (machopic_should_output_picbase_label): New.
11931 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11932 (nonlocal_goto_receiver): New insn and split.
11934 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11936 * config/aarch64/aarch64-builtins.c
11937 (aarch64_fold_builtin): Fold abs in all modes.
11938 * config/aarch64/aarch64-simd-builtins.def
11939 (abs): Enable for all modes.
11940 * config/aarch64/arm_neon.h
11941 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11942 (vabs_f64): Add missing intrinsic.
11944 2013-07-19 Ian Bolton <ian.bolton@arm.com>
11946 * config/aarch64/arm_neon.h (vabs_s64): New function
11948 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
11951 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11952 * config/avr/avr.md (adjust_len): Add `round'.
11953 * config/avr/avr-protos.h (avr_out_round): New prototype.
11954 (avr_out_plus): Add `out_label' argument.
11955 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11956 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11957 Handle the case where `insn' is just a pattern.
11958 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11959 (avr_out_round): New function.
11960 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11962 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
11964 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11965 Add check for flag_stack_usage to handle -fstack-usage support
11967 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
11969 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11970 interaction for new Power8 flags and VSX.
11972 2013-07-18 Sriraman Tallam <tmsriram@google.com>
11974 PR middle-end/57698
11975 * tree-inline.c (expand_call_inline): Emit errors during
11976 early_inlining only if optimization is not turned on.
11978 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11982 * passes.c (init_optimization_passes): Move the construction of
11983 the pass hierarchy into a new passes.def file.
11985 * Makefile.in (passes.o): Add dependency on passes.def.
11987 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11989 * passes.c (init_optimization_passes): Introduce macros for
11990 constructing the tree of passes (INSERT_PASSES_AFTER,
11991 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
11992 TERMINATE_PASS_LIST).
11994 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
11995 Wei Mi <wmi@google.com>
11997 PR rtl-optimization/57878
11998 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12000 (reload_pseudo_compare_func): Check nregs first for reload
12003 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12005 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12007 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12009 * read-rtl.c (validate_const_int): Once an invalid character is
12010 seen, quit the loop.
12012 * gengtype.c (write_roots): Similarly once we find the "deletable"
12013 or "if_marked" option.
12015 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12017 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12018 "xtab" and "sat". Move value "clz" from here to ...
12019 (attriubte "type"): ... here.
12020 (satsi_<SAT:code>): Delete "insn" attribute.
12021 (satsi_<SAT:code>_shift): Likewise.
12022 (arm_zero_extendqisi2addsi): Likewise.
12023 (arm_extendqisi2addsi): Likewise.
12024 (clzsi2): Update for attribute changes.
12025 (rbitsi2): Likewise.
12026 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12028 (arm_usatsihi): Likewise.
12029 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12031 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12033 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12034 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12035 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12036 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12037 in alphabetical order.
12038 (attribute "core_cycles"): Update for attribute changes.
12039 (arm_addsi3): Likewise.
12040 (addsi3_compare0): Likewise.
12041 (addsi3_compare0_scratch): Likewise.
12042 (addsi3_compare_op1): Likewise.
12043 (addsi3_compare_op2): Likewise.
12044 (compare_addsi2_op0): Likewise.
12045 (compare_addsi2_op1): Likewise.
12046 (addsi3_carryin_shift_<optab>): Likewise.
12047 (subsi3_carryin_shift): Likewise.
12048 (rsbsi3_carryin_shift): Likewise.
12049 (arm_subsi3_insn): Likewise.
12050 (subsi3_compare0): Likewise.
12051 (subsi3_compare): Likewise.
12052 (arm_andsi3_insn): Likewise.
12053 (thumb1_andsi3_insn): Likewise.
12054 (andsi3_compare0): Likewise.
12055 (andsi3_compare0_scratch): Likewise.
12056 (zeroextractsi_compare0_scratch
12057 (andsi_not_shiftsi_si): Likewise.
12058 (iorsi3_insn): Likewise.
12059 (iorsi3_compare0): Likewise.
12060 (iorsi3_compare0_scratch): Likewise.
12061 (arm_xorsi3): Likewise.
12062 (thumb1_xorsi3_insn): Likewise.
12063 (xorsi3_compare0): Likewise.
12064 (xorsi3_compare0_scratch): Likewise.
12065 (satsi_<SAT:code>_shift): Likewise.
12067 (arm_shiftsi3): Likewise.
12068 (shiftsi3_compare0): Likewise.
12069 (not_shiftsi): Likewise.
12070 (not_shiftsi_compare0): Likewise.
12071 (not_shiftsi_compare0_scratch): Likewise.
12072 (arm_one_cmplsi2): Likewise.
12073 (thumb_one_complsi2): Likewise.
12074 (notsi_compare0): Likewise.
12075 (notsi_compare0_scratch): Likewise.
12076 (thumb1_zero_extendhisi2): Likewise.
12077 (arm_zero_extendhisi2): Likewise.
12078 (arm_zero_extendhisi2_v6): Likewise.
12079 (arm_zero_extendhisi2addsi): Likewise.
12080 (thumb1_zero_extendqisi2): Likewise.
12081 (thumb1_zero_extendqisi2_v6): Likewise.
12082 (arm_zero_extendqisi2): Likewise.
12083 (arm_zero_extendqisi2_v6): Likewise.
12084 (arm_zero_extendqisi2addsi): Likewise.
12085 (thumb1_extendhisi2): Likewise.
12086 (arm_extendhisi2): Likewise.
12087 (arm_extendhisi2_v6): Likewise.
12088 (arm_extendqisi): Likewise.
12089 (arm_extendqisi_v6): Likewise.
12090 (arm_extendqisi2addsi): Likewise.
12091 (thumb1_extendqisi2): Likewise.
12092 (thumb1_movdi_insn): Likewise.
12093 (arm_movsi_insn): Likewise.
12094 (movsi_compare0): Likewise.
12095 (movhi_insn_arch4): Likewise.
12096 (movhi_bytes): Likewise.
12097 (arm_movqi_insn): Likewise.
12098 (thumb1_movqi_insn): Likewise.
12099 (arm32_movhf): Likewise.
12100 (thumb1_movhf): Likewise.
12101 (arm_movsf_soft_insn): Likewise.
12102 (thumb1_movsf_insn): Likewise.
12103 (movdf_soft_insn): Likewise.
12104 (thumb_movdf_insn): Likewise.
12105 (arm_cmpsi_insn): Likewise.
12106 (cmpsi_shiftsi): Likewise.
12107 (cmpsi_shiftsi_swp): Likewise.
12108 (arm_cmpsi_negshiftsi_si): Likewise.
12109 (movsicc_insn): Likewise.
12110 (movsfcc_soft_insn): Likewise.
12111 (arith_shiftsi): Likewise.
12112 (arith_shiftsi_compare0
12113 (arith_shiftsi_compare0_scratch
12114 (sub_shiftsi): Likewise.
12115 (sub_shiftsi_compare0
12116 (sub_shiftsi_compare0_scratch
12117 (and_scc): Likewise.
12118 (cond_move): Likewise.
12119 (if_plus_move): Likewise.
12120 (if_move_plus): Likewise.
12121 (if_move_not): Likewise.
12122 (if_not_move): Likewise.
12123 (if_shift_move): Likewise.
12124 (if_move_shift): Likewise.
12125 (if_shift_shift): Likewise.
12126 (if_not_arith): Likewise.
12127 (if_arith_not): Likewise.
12128 (cond_move_not): Likewise.
12129 (thumb1_ashlsi3): Set type attribute.
12130 (thumb1_ashrsi3): Likewise.
12131 (thumb1_lshrsi3): Likewise.
12132 (thumb1_rotrsi3): Likewise.
12133 (shiftsi3_compare0_scratch): Likewise.
12134 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12135 (neon_mov<mode>): Likewise.
12136 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12138 (thumb2_movsi_insn): Likewise.
12139 (thumb2_cmpsi_neg_shiftsi): Likewise.
12140 (thumb2_extendqisi_v6): Likewise.
12141 (thumb2_zero_extendhisi2_v6): Likewise.
12142 (thumb2_zero_extendqisi2_v6): Likewise.
12143 (thumb2_shiftsi3_short): Likewise.
12144 (thumb2_addsi3_compare0_scratch): Likewise.
12145 (orsi_not_shiftsi_si): Likewise.
12146 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12147 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12149 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12150 (1020alu_shift_op): Likewise.
12151 (1020alu_shift_reg_op): Likewise.
12152 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12153 (alu_shift_op): Likewise.
12154 (alu_shift_reg_op): Likewise.
12155 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12156 (11_alu_shift_op): Likewise.
12157 (11_alu_shift_reg_op): Likewise.
12158 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12159 (9_alu_shift_reg_op): Likewise.
12160 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12162 (cortex_a15_alu_shift): Likewise.
12163 (cortex_a15_alu_shift_reg): Likewise.
12164 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12166 (cortex_a5_alu_shift): Likewise.
12167 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12169 (cortex_a53_alu_shift): Likewise.
12170 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12172 (cortex_a7_alu_reg): Likewise.
12173 (cortex_a7_alu_shift): Likewise.
12174 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12176 (cortex_a8_alu_shift): Likewise.
12177 (cortex_a8_alu_shift_reg): Likewise.
12178 (cortex_a8_mov): Likewise.
12179 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12180 (cortex_a9_dp_shift): Likewise.
12181 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12183 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12185 (cortex_r4_mov): Likewise.
12186 (cortex_r4_alu_shift): Likewise.
12187 (cortex_r4_alu_shift_reg): Likewise.
12188 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12189 (526_alu_shift_op): Likewise.
12190 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12191 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12192 (626te_alu_shift_op): Likewise.
12193 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12194 (726te_alu_op): Likewise.
12195 (726te_alu_shift_op): Likewise.
12196 (726te_alu_shift_reg_op): Likewise.
12197 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12198 (mp626_alu_shift_op): Likewise.
12199 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12200 (pj4_alu_e1_conds): Likewise.
12201 (pj4_alu): Likewise.
12202 (pj4_alu_conds): Likewise.
12203 (pj4_shift): Likewise.
12204 (pj4_shift_conds): Likewise.
12205 (pj4_alu_shift): Likewise.
12206 (pj4_alu_shift_conds): Likewise.
12207 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12209 (cortexa7_older_only): Likewise.
12210 (cortexa7_younger): Likewise.
12212 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12214 * ipa-pure-const.c (generate_summary): Rename to...
12215 (pure_const_generate_summary): ... this.
12217 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12219 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12221 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12224 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12227 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12229 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12230 enabled without -march=zEC12.
12231 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12234 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12236 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12237 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12238 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12239 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12240 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12241 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12242 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12244 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12245 and ISA_HAS_NMADD3_NMSUB3.
12246 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12247 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12248 (nmsub4<mode>, nmsub3<mode>): Likewise.
12249 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12251 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12253 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12254 TARGET_MIPS5400 checking.
12256 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12257 Peter Bergner <bergner@vnet.ibm.com>
12259 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12260 registers in the comment.
12261 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12262 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12263 rather than FIRST_PSEUDO_REGISTERS.
12265 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12267 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12268 enable extra ISA flags with TARGET_HTM.
12270 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12272 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12275 2013-07-15 Cong Hou <congh@google.com>
12277 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12278 in compare function for sorting.
12280 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12282 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12283 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12284 * config/rs6000/rs6000.opt: Add -mhtm option.
12285 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12286 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12287 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12288 __HTM__ if the HTM instructions are available.
12289 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12290 htm_spr_reg_operand): New define_predicates.
12291 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12292 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12294 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12295 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12296 HTM builtin functions.
12297 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12298 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12299 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12300 (rs6000_builtin_mask_calculate): Likewise.
12301 (rs6000_option_override_internal): Likewise.
12302 (bdesc_htm): Add new HTM builtin support.
12303 (htm_spr_num): New function.
12304 (htm_spr_regno): Likewise.
12305 (rs6000_htm_spr_icode): Likewise.
12306 (htm_expand_builtin): Likewise.
12307 (htm_init_builtins): Likewise.
12308 (rs6000_expand_builtin): Add support for HTM builtin functions.
12309 (rs6000_init_builtins): Likewise.
12310 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12312 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12313 (TARGET_HTM, MASK_HTM): Define macros.
12314 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12315 (FIXED_REGISTERS): Likewise.
12316 (CALL_USED_REGISTERS): Likewise.
12317 (CALL_REALLY_USED_REGISTERS): Likewise.
12318 (REG_ALLOC_ORDER): Likewise.
12319 (enum reg_class): Likewise.
12320 (REG_CLASS_NAMES): Likewise.
12321 (REG_CLASS_CONTENTS): Likewise.
12322 (REGISTER_NAMES): Likewise.
12323 (ADDITIONAL_REGISTER_NAMES): Likewise.
12324 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12325 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12326 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12327 * config/rs6000/htm.md: New file.
12328 * config/rs6000/htmintrin.h: New file.
12329 * config/rs6000/htmxlintrin.h: New file.
12331 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12333 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12334 Define SYMBOL_TINY_GOT, update comment.
12335 * config/aarch64/aarch64.c
12336 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12337 (aarch64_expand_mov_immediate): Likewise.
12338 (aarch64_print_operand): Likewise.
12339 (aarch64_classify_symbol): Likewise.
12340 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12341 (ldr_got_tiny): Define.
12343 2013-07-13 Tobias Grosser <tobias@grosser.es>
12345 PR tree-optimization/54094
12346 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12347 scheduling dimension for the parallelism check from the polyhedral
12348 information in the AST.
12349 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12352 2013-07-13 Jason Merrill <jason@redhat.com>
12354 * print-tree.c (debug_vec_tree): Use debug_raw.
12355 (debug_raw (vec<tree, va_gc> &)): New.
12356 (debug_raw (vec<tree, va_gc> *)): New.
12357 * tree.h: Declare them.
12359 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12361 * ifcvt.c (ifcvt_after_combine): New static variable.
12362 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12364 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12365 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12366 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12368 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12370 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12372 2013-07-12 Michael Matz <matz@suse.de>
12374 PR middle-end/55771
12375 * convert.c (convert_to_real): Reject non-float inner types.
12377 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12379 * config/aarch64/aarch64-protos.h
12380 (aarch64_simd_immediate_valid_for_move): Remove.
12381 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12382 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12384 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12386 2013-07-11 Steve Ellcey <sellcey@mips.com>
12388 * config/mips/mips.c (mips_conditional_register_usage): Do not
12389 use t[0-7] registers in MIPS16 mode when optimizing for size.
12391 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12393 * config/i386/i386.c (dispatch_function_versions): Fix array
12394 indexing of function_version_info to match actual_versions.
12396 2013-07-11 Teresa Johnson <tejohnson@google.com>
12398 * vec.h (struct va_gc): Move release out-of-line.
12399 (va_gc::release): Call ggc_free on released vec.
12401 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12403 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12404 Require GOT register as additional operand in UNSPEC.
12405 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12406 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12407 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12408 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12409 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12410 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12411 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12413 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12416 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12417 name seen by assembler/linker rather if available.
12419 2013-07-11 Andreas Schwab <schwab@suse.de>
12421 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12423 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12425 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12427 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12429 * doc/tm.texi.in: Move hook documentation to ....
12430 * target.def: ... here.
12432 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12433 text on @hook line.
12434 * doc/tm.texi: Regenerate.
12436 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12439 * doc/invoke.texi: Document -Wconditionally-supported.
12441 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12444 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12447 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12450 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12451 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12452 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12453 Remove duplicate devices.
12454 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12455 * config/avr/t-multilib: Regenerate.
12456 * config/avr/avr-tables.opt: Regenerate.
12457 * doc/avr-mmcu.texi: Regenerate.
12459 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12462 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12464 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12466 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12467 the cost of MULT when optimizing for size.
12469 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12471 * config/cr16/cr16-protos.h: Don't include target.h.
12473 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12475 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12476 adjust register size for TDmode and TFmode for VSX registers.
12478 2013-07-08 Kai Tietz <ktietz@redhat.com>
12481 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12482 hook_bool_const_tree_true.
12484 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12486 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12487 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12488 * config/s390/s390.md: Define FPR*_REGNUM constants.
12489 Fix FPR2_REGNUM constant (18 -> 17).
12490 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12491 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12492 ("trunc<BFP:mode><DFP_ALL:mode>2")
12493 ("trunc<DFP_ALL:mode><BFP:mode>2")
12494 ("*extend<BFP:mode><DFP_ALL:mode>2")
12495 ("*extend<DFP_ALL:mode><BFP:mode>2")
12496 ("extend<BFP:mode><DFP_ALL:mode>2")
12497 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12500 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12502 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12504 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12506 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12507 and cfun_fpr_bit_p to cfun_fpr_save_p.
12508 (s390_frame_area, s390_register_info, s390_frame_info)
12509 (s390_emit_prologue, s390_emit_epilogue)
12510 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12512 * config/s390/s390.h: Define *_REGNUM macros for floating point
12515 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12517 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12519 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12521 PR rtl-optimization/57786
12522 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12523 and break out of the loop when it is set to false.
12525 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12528 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12529 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12530 (const_int 63)) 0)).
12531 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12532 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12533 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12535 PR rtl-optimization/57829
12536 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12537 mask bits outside of mode are just sign-extension from mode to HWI.
12539 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12541 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12542 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12543 adjust_address instead of change_address to keep info about alignment.
12544 (emit_strmov): Remove.
12545 (emit_memmov): New function.
12546 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12547 (expand_movmem_epilogue): Likewise and return updated rtx for
12549 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12550 destination and source.
12551 (decide_alignment): Refactor, handle vector_loop.
12552 (ix86_expand_movmem): Likewise.
12553 (ix86_expand_setmem): Likewise.
12554 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12556 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12558 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12559 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12561 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12563 * config/i386/sse.md (sse_movlhps): Change alternative 3
12564 of operand 2 to "m".
12566 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12569 * config/i386/sse.md (iptr): New mode attribute.
12570 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12571 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12572 (<sse>_vmmul<mode>3): Ditto.
12573 (<sse>_vmdiv<mode>3): Ditto.
12574 (sse_vmrcpv4sf2): Ditto.
12575 (<sse>_vmsqrt<mode>2): Ditto.
12576 (sse_vmrsqrtv4sf2): Ditto.
12577 (<sse>_vm<code><mode>3): Ditto.
12578 (avx_vmcmp<mode>3): Ditto.
12579 (<sse>_vmmaskcmp<mode>3): Ditto.
12580 (<sse>_comi): Ditto.
12581 (<sse>_ucomi): Ditto.
12582 (*xop_vmfrcz_<mode>): Ditto.
12583 (*fmai_fmadd_<mode>): Ditto.
12584 (*fmai_fmsub_<mode>): Ditto.
12585 (*fmai_fnmadd_<mode>): Ditto.
12586 (*fmai_fnmsub_<mode>): Ditto.
12587 (*fma4i_vmfmadd_<mode>): Ditto.
12588 (*fma4i_vmfmsub_<mode>): Ditto.
12589 (*fma4i_vmfnmadd_<mode>): Ditto.
12590 (*fma4i_vmfnmsub_<mode>): Ditto.
12591 (*xop_vmfrcz_<mode>): Ditto.
12592 (sse_cvtps2pi): Ditto.
12593 (sse_cvttps2pi): Ditto.
12594 (sse_cvtss2si): Ditto.
12595 (sse_cvtss2si_2): Ditto.
12596 (sse_cvtss2siq_2): Ditto.
12597 (sse_cvttss2si): Ditto.
12598 (sse_cvttss2siq): Ditto.
12599 (sse_cvtsd2si): Ditto.
12600 (sse_cvtsd2si_2): Ditto.
12601 (sse_cvtsd2siq_2): Ditto.
12602 (sse_cvttsd2si): Ditto.
12603 (sse_cvttsd2siq): Ditto.
12604 (sse_cvtsd2ss): Ditto.
12605 (sse_cvtss2sd): Ditto.
12606 (avx2_pbroadcast<mode>): Ditto.
12607 (avx2_pbroadcast<mode>_1): Ditto.
12608 (*avx_vperm_broadcast_v4sf): Ditto.
12610 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12611 (sse_movlhps): Ditto.
12612 (sse_storehps): Ditto.
12613 (sse_loadhps): Ditto.
12614 (sse_storelps): Ditto.
12615 (sse_loadlps): Ditto.
12616 (*vec_concatv4sf): Ditto.
12617 (*vec_interleave_highv2df): Ditto.
12618 (*vec_interleave_lowv2df): Ditto.
12619 (*vec_extractv2df_1_sse): Ditto.
12620 (*vec_extractv2df_0_sse): Ditto.
12621 (sse2_storelpd): Ditto.
12622 (sse2_loadlpd): Ditto.
12623 (sse2_movsd): Ditto.
12624 (*vec_concatv4si): Ditto.
12625 (vec_concatv2di): Ditto.
12627 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12628 for Intel asm dialect.
12629 (mmx_punpcklwd): Ditto.
12630 (mmx_punpckldq): Ditto.
12632 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12633 for intel assembler dialect.
12635 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12638 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12639 for BUILT_IN_C{LZ,LRSB}*.
12640 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12641 * tree-vrp.c (extract_range_basic): Handle
12642 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12643 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12644 fall thru to code calling set_value*.
12645 * builtins.c (expand_builtin): Remove *IMAX cases.
12646 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12647 if width is bigger than 2*HWI.
12649 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12651 PR rtl-optimization/55342
12652 * lra-int.h (lra_subreg_reload_pseudos): New.
12653 * lra.c: Add undoing optional reloads to the block diagram.
12654 (lra_subreg_reload_pseudos): New.
12655 (lra_optional_reload_pseudos): Change comments.
12656 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12657 lra_optional_reload_pseudos after undo transformations.
12658 * lra-assigns.c (pseudo_prefix_title): New.
12659 (lra_setup_reg_renumber): Use it.
12660 (spill_for): Ditto. Check subreg reload pseudos too.
12661 (assign_by_spills): Consider subreg reload pseudos too.
12662 * lra-constraints.c (simplify_operand_subreg): Use
12663 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12664 (curr_insn_transform): Recognize and do optional reloads.
12665 (undo_optional_reloads): New.
12666 (lra_undo_inheritance): Call undo_optional_reloads.
12668 2013-07-05 Thomas Quinot <quinot@adacore.com>
12670 * tree-complex.c (expand_complex_operations_1): Fix typo.
12672 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12674 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12675 (tune_params): New member 'const vec_costs'.
12676 * config/aarch64/aarch64.c (generic_vector_cost): New.
12677 (generic_tunings): New member 'generic_vector_cost'.
12678 (aarch64_builtin_vectorization_cost): New.
12679 (aarch64_add_stmt_cost): New.
12680 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12681 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12683 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12686 * config/i386/predicates.md (vsib_address_operand): Disallow
12687 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12689 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12691 PR middle-end/55030
12692 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12693 expand_builtin_setjmp_receiver.
12694 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12695 with NULL for the label parameter.
12696 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12697 the frame-pointer. Adjust comments.
12698 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12699 only if LABEL is non-NULL.
12701 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12703 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12704 (ARM_ABI_AAPCS64): Ditto.
12706 (ARM_DEFAULT_ABI): Ditto.
12708 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12710 * config/aarch64/aarch64-builtins.c
12711 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12712 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12714 * config/aarch64/aarch64-simd.md
12715 (aarch64_ld1<VALL:mode>): New.
12716 (aarch64_st1<VALL:mode>): Likewise.
12717 * config/aarch64/arm_neon.h
12718 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12720 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12722 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12724 (ix86_option_override_internal):Turn on all -mavx target flags by
12725 default as they are dependent on AVX anyway.
12727 2013-07-02 Cary Coutant <ccoutant@google.com>
12729 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12730 deterministic hash.
12731 (loc_checksum_ordered): Likewise.
12732 (hash_loc_operands): Remove inline keyword.
12734 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12736 PR tree-optimization/57741
12737 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12738 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12739 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12740 Allow REAL_CST step_exprs if flag_associative_math.
12741 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12743 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12745 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12747 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12749 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12751 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12753 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12755 (iorsi3_insn): Likewise.
12756 (arm_xorsi3): Likewise.
12758 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12760 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12762 (attribute "type"): ... here, and prefix with "wmmx_".
12763 (attribute "core_cycles"): Update for attribute changes.
12764 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12765 (tbcstv4hi): Likewise.
12766 (tbcstv2si): Likewise.
12767 (iwmmxt_iordi3): Likewise.
12768 (iwmmxt_xordi3): Likewise.
12769 (iwmmxt_anddi3): Likewise.
12770 (iwmmxt_nanddi3): Likewise.
12771 (iwmmxt_arm_movdi): Likewise.
12772 (iwmmxt_movsi_insn): Likewise.
12773 (mov<mode>_internal): Likewise.
12774 (and<mode>3_iwmmxt): Likewise.
12775 (ior<mode>3_iwmmxt): Likewise.
12776 (xor<mode>3_iwmmxt): Likewise.
12777 (add<mode>3_iwmmxt): Likewise.
12778 (ssaddv8qi3): Likewise.
12779 (ssaddv4hi3): Likewise.
12780 (ssaddv2si3): Likewise.
12781 (usaddv8qi3): Likewise.
12782 (usaddv4hi3): Likewise.
12783 (usaddv2si3): Likewise.
12784 (sub<mode>3_iwmmxt): Likewise.
12785 (sssubv8qi3): Likewise.
12786 (sssubv4hi3): Likewise.
12787 (sssubv2si3): Likewise.
12788 (ussubv8qi3): Likewise.
12789 (ussubv4hi3): Likewise.
12790 (ussubv2si3): Likewise.
12791 (mulv4hi3_iwmmxt): Likewise.
12792 (smulv4hi3_highpart): Likewise.
12793 (umulv4hi3_highpart): Likewise.
12794 (iwmmxt_wmacs): Likewise.
12795 (iwmmxt_wmacsz): Likewise.
12796 (iwmmxt_wmacu): Likewise.
12797 (iwmmxt_wmacuz): Likewise.
12798 (iwmmxt_clrdi): Likewise.
12799 (iwmmxt_clrv8qi): Likewise.
12800 (iwmmxt_clr4hi): Likewise.
12801 (iwmmxt_clr2si): Likewise.
12802 (iwmmxt_uavgrndv8qi3): Likewise.
12803 (iwmmxt_uavgrndv4hi3): Likewise.
12804 (iwmmxt_uavgv8qi3): Likewise.
12805 (iwmmxt_uavgv4hi3): Likewise.
12806 (iwmmxt_tinsrb): Likewise.
12807 (iwmmxt_tinsrh): Likewise.
12808 (iwmmxt_tinsrw): Likewise.
12809 (iwmmxt_textrmub): Likewise.
12810 (iwmmxt_textrmsb): Likewise.
12811 (iwmmxt_textrmuh): Likewise.
12812 (iwmmxt_textrmsh): Likewise.
12813 (iwmmxt_textrmw): Likewise.
12814 (iwmxxt_wshufh): Likewise.
12815 (eqv8qi3): Likewise.
12816 (eqv4hi3): Likewise.
12817 (eqv2si3): Likewise.
12818 (gtuv8qi3): Likewise.
12819 (gtuv4hi3): Likewise.
12820 (gtuv2si3): Likewise.
12821 (gtv8qi3): Likewise.
12822 (gtv4hi3): Likewise.
12823 (gtv2si3): Likewise.
12824 (smax<mode>3_iwmmxt): Likewise.
12825 (umax<mode>3_iwmmxt): Likewise.
12826 (smin<mode>3_iwmmxt): Likewise.
12827 (umin<mode>3_iwmmxt): Likewise.
12828 (iwmmxt_wpackhss): Likewise.
12829 (iwmmxt_wpackwss): Likewise.
12830 (iwmmxt_wpackdss): Likewise.
12831 (iwmmxt_wpackhus): Likewise.
12832 (iwmmxt_wpackwus): Likewise.
12833 (iwmmxt_wpackdus): Likewise.
12834 (iwmmxt_wunpckihb): Likewise.
12835 (iwmmxt_wunpckihh): Likewise.
12836 (iwmmxt_wunpckihw): Likewise.
12837 (iwmmxt_wunpckilb): Likewise.
12838 (iwmmxt_wunpckilh): Likewise.
12839 (iwmmxt_wunpckilw): Likewise.
12840 (iwmmxt_wunpckehub): Likewise.
12841 (iwmmxt_wunpckehuh): Likewise.
12842 (iwmmxt_wunpckehuw): Likewise.
12843 (iwmmxt_wunpckehsb): Likewise.
12844 (iwmmxt_wunpckehsh): Likewise.
12845 (iwmmxt_wunpckehsw): Likewise.
12846 (iwmmxt_wunpckelub): Likewise.
12847 (iwmmxt_wunpckeluh): Likewise.
12848 (iwmmxt_wunpckeluw): Likewise.
12849 (iwmmxt_wunpckelsb): Likewise.
12850 (iwmmxt_wunpckelsh): Likewise.
12851 (iwmmxt_wunpckelsw): Likewise.
12852 (ror<mode>3): Likewise.
12853 (ashr<mode>3_iwmmxt): Likewise.
12854 (lshr<mode>3_iwmmxt): Likewise.
12855 (ashl<mode>3_iwmmxt): Likewise.
12856 (ror<mode>3_di): Likewise.
12857 (ashr<mode>3_di): Likewise.
12858 (lshr<mode>3_di): Likewise.
12859 (ashl<mode>3_di): Likewise.
12860 (iwmmxt_wmadds): Likewise.
12861 (iwmmxt_wmaddu): Likewise.
12862 (iwmmxt_tmia): Likewise.
12863 (iwmmxt_tmiaph): Likewise.
12864 (iwmmxt_tmiabb): Likewise.
12865 (iwmmxt_tmiatb): Likewise.
12866 (iwmmxt_tmiabt): Likewise.
12867 (iwmmxt_tmiatt): Likewise.
12868 (iwmmxt_tmovmskb): Likewise.
12869 (iwmmxt_tmovmskh): Likewise.
12870 (iwmmxt_tmovmskw): Likewise.
12871 (iwmmxt_waccb): Likewise.
12872 (iwmmxt_wacch): Likewise.
12873 (iwmmxt_waccw): Likewise.
12874 (iwmmxt_waligni): Likewise.
12875 (iwmmxt_walignr): Likewise.
12876 (iwmmxt_walignr0): Likewise.
12877 (iwmmxt_walignr1): Likewise.
12878 (iwmmxt_walignr2): Likewise.
12879 (iwmmxt_walignr3): Likewise.
12880 (iwmmxt_wsadb): Likewise.
12881 (iwmmxt_wsadh): Likewise.
12882 (iwmmxt_wsadbz): Likewise.
12883 (iwmmxt_wsadhz): Likewise.
12884 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12885 (iwmmxt_wabsdiffb): Likewise.
12886 (iwmmxt_wabsdiffh): Likewise.
12887 (iwmmxt_wabsdiffw): Likewise.
12888 (iwmmxt_waddsubhx): Likewise
12889 (iwmmxt_wsubaddhx): Likewise.
12890 (addc<mode>3): Likewise.
12891 (iwmmxt_avg4): Likewise.
12892 (iwmmxt_avg4r): Likewise.
12893 (iwmmxt_wmaddsx): Likewise.
12894 (iwmmxt_wmaddux): Likewise.
12895 (iwmmxt_wmaddsn): Likewise.
12896 (iwmmxt_wmaddun): Likewise.
12897 (iwmmxt_wmulwsm): Likewise.
12898 (iwmmxt_wmulwum): Likewise.
12899 (iwmmxt_wmulsmr): Likewise.
12900 (iwmmxt_wmulumr): Likewise.
12901 (iwmmxt_wmulwsmr): Likewise.
12902 (iwmmxt_wmulwumr): Likewise.
12903 (iwmmxt_wmulwl): Likewise.
12904 (iwmmxt_wqmulm): Likewise.
12905 (iwmmxt_wqmulwm): Likewise.
12906 (iwmmxt_wqmulmr): Likewise.
12907 (iwmmxt_wqmulwmr): Likewise.
12908 (iwmmxt_waddbhusm): Likewise.
12909 (iwmmxt_waddbhusl): Likewise.
12910 (iwmmxt_wqmiabb): Likewise.
12911 (iwmmxt_wqmiabt): Likewise.
12912 (iwmmxt_wqmiatb): Likewise.
12913 (iwmmxt_wqmiatt): Likewise.
12914 (iwmmxt_wqmiabbn): Likewise.
12915 (iwmmxt_wqmiabtn): Likewise.
12916 (iwmmxt_wqmiatbn): Likewise.
12917 (iwmmxt_wqmiattn): Likewise.
12918 (iwmmxt_wmiabb): Likewise.
12919 (iwmmxt_wmiabt): Likewise.
12920 (iwmmxt_wmiatb): Likewise.
12921 (iwmmxt_wmiatt): Likewise.
12922 (iwmmxt_wmiabbn): Likewise.
12923 (iwmmxt_wmiabtn): Likewise.
12924 (iwmmxt_wmiatbn): Likewise.
12925 (iwmmxt_wmiattn): Likewise.
12926 (iwmmxt_wmiawbb): Likewise.
12927 (iwmmxt_wmiawbt): Likewise.
12928 (iwmmxt_wmiawtb): Likewise.
12929 (iwmmxt_wmiawtt): Likewise.
12930 (iwmmxt_wmiawbbn): Likewise.
12931 (iwmmxt_wmiawbtn): Likewise.
12932 (iwmmxt_wmiawtbn): Likewise.
12933 (iwmmxt_wmiawttn): Likewise.
12934 (iwmmxt_wmerge): Likewise.
12935 (iwmmxt_tandc<mode>3): Likewise.
12936 (iwmmxt_torc<mode>3): Likewise.
12937 (iwmmxt_torvsc<mode>3): Likewise.
12938 (iwmmxt_textrc<mode>3): Likewise.
12939 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12940 (wmmxt_pack): Likewise.
12941 (wmmxt_mult_c1): Likewise.
12942 (wmmxt_mult_c2): Likewise.
12943 (wmmxt_alu_c1): Likewise.
12944 (wmmxt_alu_c2): Likewise.
12945 (wmmxt_alu_c3): Likewise.
12946 (wmmxt_transfer_c1): Likewise.
12947 (wmmxt_transfer_c2): Likewise.
12948 (wmmxt_transfer_c3): Likewise.
12949 (marvell_f_iwmmxt_wstr): Likewise.
12950 (marvell_f_iwmmxt_wldr): Likewise.
12952 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12954 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
12956 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12959 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12960 * lra-constraints.c (need_for_split_p): Check call used hard regs
12961 living through calls.
12963 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12964 call used regs for call insn.
12966 2013-06-28 Jakub Jelinek <jakub@redhat.com>
12969 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12970 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
12972 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12974 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
12976 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12978 * lra-constraints.c (need_for_split_p): Check call used hard regs
12979 living through calls.
12981 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
12984 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
12985 to tie with any other modes. Eliminate Altivec vector mode tests,
12986 since these are a subset of ALTIVEC or VSX vector modes. Simplify
12987 code, to return 0 if testing MODE2 for a condition, if we've
12988 already tested MODE1 for the same condition.
12990 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12992 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
12995 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12997 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12998 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13000 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13002 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13004 (aarch64_symbolic_constant_p): Remove.
13005 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13006 static. Fix line length and white space.
13007 (aarch64_symbolic_constant_p): Remove.
13008 * config/aarch64/predicates.md (aarch64_valid_symref):
13009 Use aarch64_classify_symbol_expression.
13011 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13013 * config/arm/constraints.md (Ts): New constraint.
13014 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13016 (compare_scc): Use "Ts" constraint for operand 0.
13017 (ior_scc_scc): Likewise.
13018 (and_scc_scc): Likewise.
13019 (and_scc_scc_nodom): Likewise.
13020 (ior_scc_scc_cmp): Likewise for operand 7.
13021 (and_scc_scc_cmp): Likewise.
13022 * config/arm/thumb2.md (thumb2_movsi_insn):
13023 Add alternatives for 16-bit encodings.
13024 (thumb2_movhi_insn): Likewise.
13025 (thumb2_movsicc_insn): Likewise.
13026 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13027 (thumb2_negscc): Use "Ts" constraint.
13028 Move mvn instruction outside cond_exec block.
13029 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13030 for 16-bit encodings.
13032 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13034 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13036 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13037 (mulsi3subsi): Likewise.
13038 (mulsidi3adddi): Likewise.
13039 (mulsidi3_v6): Likewise.
13040 (umulsidi3_v6): Likewise.
13041 (umulsidi3adddi_v6): Likewise.
13042 (smulsi3_highpart_v6): Likewise.
13043 (umulsi3_highpart_v6): Likewise.
13044 (mulhisi3tb): Likewise.
13045 (mulhisi3bt): Likewise.
13046 (mulhisi3tt): Likewise.
13047 (maddhisi4): Likewise.
13048 (maddhisi4tb): Likewise.
13049 (maddhisi4tt): Likewise.
13050 (maddhidi4): Likewise.
13051 (maddhidi4tb): Likewise.
13052 (maddhidi4tt): Likewise.
13053 (zeroextractsi_compare0_scratch): Likewise.
13054 (insv_zero): Likewise.
13055 (insv_t2): Likewise.
13056 (anddi_notzesidi_di): Likewise.
13057 (anddi_notsesidi_di): Likewise.
13058 (andsi_notsi_si): Likewise.
13059 (iordi_zesidi_di): Likewise.
13060 (xordi_zesidi_di): Likewise.
13061 (andsi_iorsi3_notsi): Likewise.
13062 (smax_0): Likewise.
13063 (smax_m1): Likewise.
13064 (smin_0): Likewise.
13065 (not_shiftsi): Likewise.
13066 (unaligned_loadsi): Likewise.
13067 (unaligned_loadhis): Likewise.
13068 (unaligned_loadhiu): Likewise.
13069 (unaligned_storesi): Likewise.
13070 (unaligned_storehi): Likewise.
13071 (extv_reg): Likewise.
13072 (extzv_t2): Likewise.
13073 (divsi3): Likewise.
13074 (udivsi3): Likewise.
13075 (arm_zero_extendhisi2addsi): Likewise.
13076 (arm_zero_extendqisi2addsi): Likewise.
13077 (compareqi_eq0): Likewise.
13078 (arm_extendhisi2_v6): Likewise.
13079 (arm_extendqisi2addsi): Likewise.
13080 (arm_movt): Likewise.
13081 (thumb2_ldrd): Likewise.
13082 (thumb2_ldrd_base): Likewise.
13083 (thumb2_ldrd_base_neg): Likewise.
13084 (thumb2_strd): Likewise.
13085 (thumb2_strd_base): Likewise.
13086 (thumb2_strd_base_neg): Likewise.
13087 (arm_negsi2): Add alternative for 16-bit encoding.
13088 (arm_one_cmplsi2): Likewise.
13090 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13092 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13093 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13094 (movdfcc): Likewise.
13095 * config/arm/vfp.md (*thumb2_movsf_vfp):
13096 Disable predication for arm_restrict_it.
13097 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13098 (*thumb2_movdfcc_vfp): Likewise.
13099 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13100 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13101 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13102 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13103 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13104 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13105 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13106 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13107 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13108 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13109 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13110 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13111 Disable predication for arm_restrict_it.
13113 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13115 * config/i386/bmiintrin.h (_bextr_u32): New.
13116 (_bextr_u64): Ditto.
13118 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13120 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13121 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13122 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13123 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13124 * config/mips/n32-elf.h: ...this new file.
13126 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13129 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13130 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13132 2013-06-27 Catherine Moore <clm@codesourcery.com>
13134 * config/mips/mips-tables.opt: Regenerate.
13135 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13136 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13137 * doc/invoke.texi: Add -m14kc.
13139 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13142 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13143 constraints of operand 1 and 2.
13146 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13147 to match RTL canonicalization. Swap predicates and
13148 constraints of operand 1 and 2.
13150 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13152 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13153 Process OP_INOUT regs for splitting too.
13155 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13157 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13158 decl before the loop, initialize to NULL.
13159 (vectorizable_load): Initialize ptr_incr to NULL.
13161 2013-06-27 Martin Jambor <mjambor@suse.cz>
13164 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13165 * ipa-ref.c (ipa_maybe_record_reference): New function.
13166 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13167 * ipa-cp.c (create_specialized_node): Record potential references from
13169 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13171 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13173 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13174 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13175 'mode' to force_reg.
13176 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13177 (aarch64_expand_mov_immediate): Likewise.
13179 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13181 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13182 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13184 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13186 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13187 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13188 (struct machine_function): Add tbegin_p.
13189 (s390_canonicalize_comparison): Fold CC mode compares to
13190 conditional jump if possible.
13191 (s390_emit_jump): Return the emitted jump.
13192 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13193 Handle CCRAWmode compares.
13194 (s390_option_override): Default to -mhtm if available.
13195 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13196 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13197 FPRs instead of df_regs_ever_live_p.
13198 (s390_optimize_nonescaping_tx): New function.
13199 (s390_init_frame_layout): Extend clobbered_regs array to cover
13201 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13202 (s390_expand_tbegin): New function.
13203 (enum s390_builtin): New enum definition.
13204 (code_for_builtin): New array definition.
13205 (s390_init_builtins): New function.
13206 (s390_expand_builtin): New function.
13207 (TARGET_INIT_BUILTINS): Define.
13208 (TARGET_EXPAND_BUILTIN): Define.
13209 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13210 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13211 (s390_alc_comparison): Likewise.
13212 * config/s390/s390-modes.def: Add CCRAWmode.
13213 * config/s390/s390.h (processor_flags): Add PF_TX.
13214 (TARGET_CPU_HTM): Define macro.
13215 (TARGET_HTM): Define macro.
13216 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13217 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13218 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13219 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13220 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13221 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13222 ("*cc_to_int"): Move up.
13223 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13224 constants other than 0.
13225 ("*ccraw_to_int"): New insn and splitter definition.
13226 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13227 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13228 ("tx_assist"): New expander.
13229 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13230 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13231 * config/s390/s390.opt: Add -mhtm option.
13232 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13233 * config/s390/htmxlintrin.h: New file.
13234 * config/s390/htmintrin.h: New file.
13235 * config/s390/s390intrin.h: New file.
13236 * doc/extend.texi: Document htm builtins.
13237 * config.gcc: Add the new header files to extra_headers.
13239 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13241 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13242 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13244 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13245 Pat Haugen <pthaugen@us.ibm.com>
13246 Peter Bergner <bergner@vnet.ibm.com>
13248 * config/rs6000/power8.md: New.
13249 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13250 setting for power8 entry.
13251 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13252 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13253 test for Power4/Power5 only.
13254 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13256 (force_new_group): Adjust comment.
13257 * config/rs6000/rs6000.md: Include power8.md.
13259 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13261 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13262 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13264 (tune_params): Update comment.
13265 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13266 (arm_max_conditional_execute): New function.
13267 (thumb2_final_prescan_insn): Use max_insn_skipped and
13268 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13270 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13272 PR tree-optimization/57705
13273 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13274 SSA_NAME step, provided that it is not defined inside the loop.
13275 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13276 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13278 2013-06-25 Martin Jambor <mjambor@suse.cz>
13280 PR middle-end/57670
13281 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13282 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13284 (ipa_note_param_call): Initialize member_ptr flag.
13285 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13286 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13287 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13288 (ipa_read_indirect_edge_info): Likewise.
13290 2013-06-25 Richard Biener <rguenther@suse.de>
13292 PR middle-end/56977
13293 * passes.c (init_optimization_passes): Move pass_fold_builtins
13294 and pass_dce earlier with -Og.
13296 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13298 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13299 <BIT_FIELD_REF>: Remove trailing TAB.
13300 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13303 2013-06-24 Martin Jambor <mjambor@suse.cz>
13305 PR tree-optimization/57358
13306 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13307 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13308 (ipa_analyze_params_uses): Generate pessimistic info when true.
13310 2013-06-24 Martin Jambor <mjambor@suse.cz>
13312 PR tree-optimization/57539
13313 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13314 global.inlined_to of the new node to it. All callers changed.
13315 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13316 inlining_into, pass it to cgraph_clone_node.
13317 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13318 ipa_free_edge_args_substructures.
13319 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13320 rdesc linked list. Do not assert rdesc edges have inlined caller.
13321 Assert we have found an rdesc in the rdesc list.
13323 2013-06-24 Richard Biener <rguenther@suse.de>
13325 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13326 (pointer_set_lookup): Declare.
13327 (class pointer_map): New template class implementing a
13328 generic pointer to T map.
13329 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13330 pointer_map<T>::contains, pointer_map<T>::insert,
13331 pointer_map<T>::traverse): New functions.
13332 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13333 (pointer_set_lookup): New function.
13334 (pointer_set_contains): Use pointer_set_lookup.
13335 (pointer_set_insert): Likewise.
13336 (insert_aux): Remove.
13337 (struct pointer_map_t): Embed a pointer_set_t.
13338 (pointer_map_create): Adjust.
13339 (pointer_map_destroy): Likewise.
13340 (pointer_map_contains): Likewise.
13341 (pointer_map_insert): Likewise.
13342 (pointer_map_traverse): Likewise.
13343 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13344 pointer_map<unsigned> instead of a pointer_map_t.
13345 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13346 (streamer_tree_cache_lookup): Likewise.
13347 (streamer_tree_cache_create): Likewise.
13348 (streamer_tree_cache_delete): Likewise.
13349 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13350 pointer_map<unsigned> instead of a pointer_map_t.
13351 (lto_init_tree_ref_encoder): Adjust.
13352 (lto_destroy_tree_ref_encoder): Likewise.
13353 * lto-section-out.c (lto_output_decl_index): Likewise.
13354 (lto_record_function_out_decl_state): Likewise.
13355 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13357 2013-06-24 Richard Biener <rguenther@suse.de>
13359 PR tree-optimization/57488
13360 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13362 2013-06-24 Alan Modra <amodra@gmail.com>
13364 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13365 (gen_easy_altivec_constant): Likewise.
13366 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13367 easy_vector_constant_msb): Likewise.
13369 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13372 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13373 add missing return true.
13375 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13378 * config/sh/predicates.md (general_extend_operand): Invoke
13379 general_movsrc_operand for memory operands.
13380 (general_movsrc_operand): Allow reg+reg addressing, do not use
13381 general_operand for memory operands.
13383 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13385 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13386 when current target options does not apply.
13387 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13388 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13389 * config/i386/bmiintrin.h: Pass appropriate target
13390 attributes to header.
13391 * config/i386/mmintrin.h: Ditto.
13392 * config/i386/nmmintrin.h: Ditto.
13393 * config/i386/avx2intrin.h: Ditto.
13394 * config/i386/fxsrintrin.h: Ditto.
13395 * config/i386/tbmintrin.h: Ditto.
13396 * config/i386/xsaveintrin.h: Ditto.
13397 * config/i386/f16cintrin.h: Ditto.
13398 * config/i386/xtestintrin.h: Ditto.
13399 * config/i386/xsaveoptintrin.h: Ditto.
13400 * config/i386/bmi2intrin.h: Ditto.
13401 * config/i386/lzcntintrin.h: Ditto.
13402 * config/i386/smmintrin.h: Ditto.
13403 * config/i386/wmmintrin.h: Ditto.
13404 * config/i386/x86intrin.h: Remove all header include guards.
13405 * config/i386/prfchwintrin.h: Ditto.
13406 * config/i386/pmmintrin.h: Ditto.
13407 * config/i386/tmmintrin.h: Ditto.
13408 * config/i386/xmmintrin.h: Ditto.
13409 * config/i386/popcntintrin.h: Ditto.
13410 * config/i386/rdseedintrin.h: Ditto.
13411 * config/i386/ammintrin.h: Ditto.
13412 * config/i386/emmintrin.h: Ditto.
13413 * config/i386/immintrin.h: Remove all header include guards.
13414 * config/i386/fma4intrin.h: Ditto.
13415 * config/i386/lwpintrin.h: Ditto.
13416 * config/i386/xopintrin.h: Ditto.
13417 * config/i386/ia32intrin.h: Ditto.
13418 * config/i386/avxintrin.h: Ditto.
13419 * config/i386/rtmintrin.h: Ditto.
13420 * config/i386/fmaintrin.h: Ditto.
13421 * config/i386/mm3dnow.h: Ditto.
13423 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13425 * common/config/i386/i386-common.c: Handle LZCNT.
13427 2013-06-22 Andi Kleen <ak@linux.intel.com>
13429 * doc/extend.texi: Use __atomic_store_n instead of
13430 __atomic_store in HLE example.
13432 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13434 * config/sh/sh.c: Remove <cstdlib> workaround.
13436 2013-06-21 Andi Kleen <ak@linux.intel.com>
13438 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13440 2013-06-21 Andi Kleen <ak@linux.intel.com>
13442 * doc/extend.texi: Document that __atomic_clear and
13443 __atomic_test_and_set should only be used with bool.
13445 2013-06-20 Jan Hubicka <jh@suse.cz>
13447 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13448 types_same_for_odr.
13449 * tree.c (decls_same_for_odr): New function.
13450 (same_for_edr): New function.
13451 (types_same_for_odr): New function.
13452 (get_binfo_at_offset): Use it.
13453 * tree.h (types_same_for_odr): Declare.
13455 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13456 Jason Merrill <jason@redhat.com>
13458 * system.h: Include <cstdlib> as well as <stdlib.h>.
13460 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13463 * config/i386/i386.c (construct_container): Report error if
13464 long double is used with disabled x87 float returns.
13466 2013-06-20 Jan Hubicka <jh@suse.cz>
13468 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13470 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13472 PR rtl-optimization/57425
13473 PR rtl-optimization/57569
13474 * alias.c (write_dependence_p): Remove parameters mem_mode and
13475 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13476 Changed all callers.
13477 (canon_anti_dependence): Get comments and semantics in sync.
13478 Add parameter mem_canonicalized. Changed all callers.
13479 * rtl.h (canon_anti_dependence): Update prototype.
13481 2013-06-20 Richard Biener <rguenther@suse.de>
13483 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13484 byte case, inline streamer_read_uchar and defer section
13487 2013-06-20 Richard Biener <rguenther@suse.de>
13489 PR tree-optimization/57584
13490 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13491 SSA names into the expanded expression that take part in
13492 abnormal coalescing.
13494 2013-06-19 Sharad Singhai <singhai@google.com>
13496 * gcov.c (print_usage): Handle new option.
13497 (process_args): Ditto.
13498 (get_gcov_intermediate_filename): New function.
13499 (output_intermediate_file): New function.
13500 (output_gcov_file): New function
13501 (generate_results): Handle new option.
13502 (release_function): Relase demangled name.
13503 (read_graph_file): Handle demangled name.
13504 (output_lines): Ditto.
13505 * doc/gcov.texi: Document gcov intermediate format.
13507 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13510 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13511 (lra_emit_add): Use the functions. Add comment about Y as an
13514 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13517 * collect2.c (collect_atexit): New.
13518 (collect_exit): Delete.
13519 (main): Register collect_atexit with atexit.
13520 (collect_wait): Change collect_exit to exit.
13522 * collect2.h (collect_exit): Delete.
13523 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13525 2013-06-19 Wei Mi <wmi@google.com>
13527 PR rtl-optimization/57518
13528 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13529 if regno is used in paradoxical subreg.
13530 (update_equiv_regs): Check pdx_subregs[regno] before
13531 set a reg to be equivalent with a mem.
13533 2013-06-19 Matthias Klose <doko@ubuntu.com>
13536 * file-find.h (find_a_file): Add a mode parameter.
13537 * file-find.c (find_a_file): Likewise.
13538 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13539 with X_OK for the executables.
13540 * collect2.c (main): Call find_a_file with X_OK.
13542 2013-06-19 Steve Ellcey <sellcey@mips.com>
13545 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13546 Use NEXT_INSN instead of next_real_insn.
13548 2013-06-19 Jan Hubicka <jh@suse.cz>
13550 * cgraph.h (const_value_known_p): Replace by ...
13551 (ctor_for_folding): .. this one.
13552 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13553 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13554 * expr.c (expand_expr_real_1): Likewise.
13555 (string_constant): Likewise.
13556 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13557 * ipa.c (process_references): Likewise.
13558 (symtab_remove_unreachable_nodes): Likewise.
13559 * ipa-inline-analysis.c (param_change_prob): Likewise.
13560 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13561 (get_base_constructor): Likwise.
13562 * varpool.c (varpool_remove_node): Likewise.
13563 (varpool_remove_initializer): LIkewise.
13564 (dump_varpool_node): LIkwise.
13565 (const_value_known_p): Rewrite to ...
13566 (ctor_for_folding): ... this one.
13568 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13571 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13572 PERSONALITY in $PATH derived prefixes.
13574 2013-06-19 Jeff Law <law@redhat.com>
13576 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13579 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13580 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13583 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13585 * config/arm/vfp.md: Move VFP instruction classification documentation
13587 * config/arm/arm.md: ... here. Update instruction classification
13590 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13592 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13593 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13594 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13596 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13599 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13600 with NEXT_INSN. Use tablejump_p to check for jump table data
13603 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13606 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13607 that now in C++ the value is correct per the C++ standards.
13609 2013-06-19 Richard Biener <rguenther@suse.de>
13611 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13612 for global context.
13614 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13617 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13620 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13621 with next_active_insn.
13623 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13625 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13626 functions are inlined during failures to flag an error.
13627 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13628 in early inline pass.
13630 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13632 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13635 2013-06-18 Julian Brown <julian@codesourcery.com>
13637 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13638 Permit virtual register pre-reload if !strict.
13639 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13641 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13643 * config/arm/neon.md (movmisalign<mode>): Use
13644 neon_perm_struct_or_reg_operand instead of
13645 neon_struct_or_register_operand.
13646 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13647 neon_permissive_struct_operand instead of neon_struct_operand.
13648 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13649 neon_vector_mem_operand.
13650 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13651 neon_vector_mem_operand.
13652 (neon_permissive_struct_operand): New.
13653 (neon_struct_or_register_operand): Rename to...
13654 (neon_perm_struct_or_reg_operand): This. Adjust call to
13655 neon_vector_mem_operand.
13657 2013-06-18 Richard Biener <rguenther@suse.de>
13659 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13660 * lto-streamer.h: Include pointer-set.h.
13661 (struct lto_decl_slot): Remove.
13662 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13663 Remove next_index entry.
13664 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13665 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13666 (lto_init_tree_ref_encoder): Adjust.
13667 (lto_destroy_tree_ref_encoder): Likewise.
13668 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13669 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13670 (lto_output_decl_index): Adjust.
13671 (lto_new_out_decl_state): Likewise.
13672 (lto_record_function_out_decl_state): Likewise.
13673 * lto-streamer-out.c (copy_function): Likewise.
13675 2013-06-18 Richard Biener <rguenther@suse.de>
13677 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13678 * cgraphunit.c: Include cfgloop.h.
13679 (init_lowered_empty_function): Initialize the loop tree.
13680 (assemble_thunk): Insert new BBs into loops.
13682 2013-06-18 Richard Biener <rguenther@suse.de>
13684 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13685 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13686 the map from cache entry to cache index optional.
13687 (streamer_tree_cache_replace_tree): Adjust accordingly.
13688 (streamer_tree_cache_append): Likewise.
13689 (streamer_tree_cache_delete): Likewise.
13690 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13691 streamer cache map from cache entry to cache index.
13692 * lto-streamer-out.c (create_output_block): Adjust.
13694 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13696 * config/arm/arm.md (attribute "insn"): Move multiplication and
13697 division attributes to...
13698 (attribute "type"): ... here. Remove mult.
13699 (attribute "mul32"): New attribute.
13700 (attribute "mul64"): Add umaal.
13701 (*arm_mulsi3): Update attributes.
13702 (*arm_mulsi3_v6): Likewise.
13703 (*thumb_mulsi3): Likewise.
13704 (*thumb_mulsi3_v6): Likewise.
13705 (*mulsi3_compare0): Likewise.
13706 (*mulsi3_compare0_v6): Likewise.
13707 (*mulsi_compare0_scratch): Likewise.
13708 (*mulsi_compare0_scratch_v6): Likewise.
13709 (*mulsi3addsi): Likewise.
13710 (*mulsi3addsi_v6): Likewise.
13711 (*mulsi3addsi_compare0): Likewise.
13712 (*mulsi3addsi_compare0_v6): Likewise.
13713 (*mulsi3addsi_compare0_scratch): Likewise.
13714 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13715 (*mulsi3subsi): Likewise.
13716 (*mulsidi3adddi): Likewise.
13717 (*mulsi3addsi_v6): Likewise.
13718 (*mulsidi3adddi_v6): Likewise.
13719 (*mulsidi3_nov6): Likewise.
13720 (*mulsidi3_v6): Likewise.
13721 (*umulsidi3_nov6): Likewise.
13722 (*umulsidi3_v6): Likewise.
13723 (*umulsidi3adddi): Likewise.
13724 (*umulsidi3adddi_v6): Likewise.
13725 (*smulsi3_highpart_nov6): Likewise.
13726 (*smulsi3_highpart_v6): Likewise.
13727 (*umulsi3_highpart_nov6): Likewise.
13728 (*umulsi3_highpart_v6): Likewise.
13729 (mulhisi3): Likewise.
13730 (*mulhisi3tb): Likewise.
13731 (*mulhisi3bt): Likewise.
13732 (*mulhisi3tt): Likewise.
13733 (maddhisi4): Likewise.
13734 (*maddhisi4tb): Likewise.
13735 (*maddhisi4tt): Likewise.
13736 (maddhidi4): Likewise.
13737 (*maddhidi4tb): Likewise.
13738 (*maddhidi4tt): Likewise.
13739 (divsi3): Likewise.
13740 (udivsi3): Likewise.
13741 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13742 (thumb2_mulsi_short_compare0): Likewise.
13743 (thumb2_mulsi_short_compare0_scratch): Likewise.
13744 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13745 (1020mult2): Likewise.
13746 (1020mult3): Likewise.
13747 (1020mult4): Likewise.
13748 (1020mult5): Likewise.
13749 (1020mult6): Likewise.
13750 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13752 (cortex_a15_mult64): Likewise.
13753 (cortex_a15_sdiv): Likewise.
13754 (cortex_a15_udiv): Likewise.
13755 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13761 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13762 (pj4_ir_div): Likewise.
13763 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13764 (11_mult2): Likewise.
13765 (11_mult3): Likewise.
13766 (11_mult4): Likewise.
13767 (11_mult5): Likewise.
13768 (11_mult6): Likewise.
13769 (11_mult7): Likewise.
13770 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13771 (cortex_a8_mla): Likewise.
13772 (cortex_a8_mull): Likewise.
13773 (cortex_a8_smulwy): Likewise.
13774 (cortex_a8_smlald): Likewise.
13775 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13776 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13777 (cortex_r4_mul_3): Likewise.
13778 (cortex_r4_mla_4): Likewise.
13779 (cortex_r4_mla_3): Likewise.
13780 (cortex_r4_smlald): Likewise.
13781 (cortex_r4_mull): Likewise.
13782 (cortex_r4_sdiv): Likewise.
13783 (cortex_r4_udiv): Likewise.
13784 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13785 (cortex_a7_idiv): Likewise.
13786 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13787 (9_mult2): Likewise.
13788 (9_mult3): Likewise.
13789 (9_mult4): Likewise.
13790 (9_mult5): Likewise.
13791 (9_mult6): Likewise.
13792 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13793 (cortex_a53_sdiv): Likewise.
13794 (cortex_a53_udiv): Likewise.
13795 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13796 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13797 (mp626_mult2): Likewise.
13798 (mp626_mult3): Likewise.
13799 (mp626_mult4): Likewise.
13800 * config/arm/fa526.md (526_mult1): Update attribute change.
13801 (526_mult2): Likewise.
13802 * config/arm/arm-generic.md (mult): Update attribute change.
13803 (mult_ldsched_strongarm): Likewise.
13804 (mult_ldsched): Likewise.
13805 (multi_cycle): Likewise.
13806 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13807 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13808 (606te_mult2): Likewise.
13809 (606te_mult3): Likewise.
13810 (606te_mult4): Likewise.
13811 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13812 (cortex_a9_mac16): Likewise.
13813 (cortex_a9_multiply): Likewise.
13814 (cortex_a9_mac): Likewise.
13815 (cortex_a9_multiply_long): Likewise.
13816 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13817 (626te_mult2): Likewise.
13818 (626te_mult3): Likewise.
13819 (626te_mult4): Likewise.
13821 2013-06-18 Richard Biener <rguenther@suse.de>
13824 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13826 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13829 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13830 with next_active_insn.
13832 2013-06-18 Alan Modra <amodra@gmail.com>
13834 * config/rs6000/rs6000.h (enum data_align): New.
13835 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13836 (DATA_ABI_ALIGNMENT): Define.
13837 (CONSTANT_ALIGNMENT): Correct comment.
13838 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13839 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13841 2013-06-17 David Malcolm <dmalcolm@redhat.com>
13843 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13844 ATTRIBUTE_UNUSED marking.
13846 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13848 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13849 alternative and update.
13850 (aarch64_dup_lanedi): Delete.
13851 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13852 * config/aarch64/aarch64-simd-builtins.def: Update.
13854 2013-06-17 Richard Biener <rguenther@suse.de>
13856 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13857 (lto_input_scc): Declare.
13858 (lto_input_tree_1): Likewise.
13859 (struct lto_stats_d): Add num_tree_bodies_output and
13860 num_pickle_refs_output.
13861 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13862 (lto_read_tree_1): Split out from ...
13863 (lto_read_tree): ... this.
13864 (lto_input_scc): New function.
13865 (lto_input_tree_1): Split out from ...
13866 (lto_input_tree): ... this. Handle LTO_tree_scc.
13867 (lto_data_in_create): Create the streamer cache without hashes.
13868 * lto-streamer-out.c (create_output_block): Create the streamer
13869 cache with hashes when not doing WPA.
13870 (lto_write_tree_1): Split out from ...
13871 (lto_write_tree): ... this.
13872 (get_symbol_initial_value): New function.
13873 (lto_output_tree_1): Split out from ...
13874 (lto_output_tree): ... this. Write trees as series of SCCs
13875 using a DFS walk via DFS_write_tree.
13876 (struct sccs, struct scc_entry): New types.
13877 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13878 (DFS_write_tree_body): New function.
13879 (DFS_write_tree): Likewise.
13880 (hash_tree): Likewise.
13881 (scc_entry_compare): Likewise.
13882 (hash_scc): Likewise.
13883 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13884 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13885 TREE_CHAIN as regular reference.
13886 (streamer_read_integer_cst): Remove.
13887 (streamer_get_pickled_tree): Adjust.
13888 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13889 of DECL_EXTERNALs in BLOCK_VARS for now.
13890 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13892 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13893 Add hash value argument and record that if hashes are recorded
13895 (streamer_tree_cache_insert_1): Adjust.
13896 (streamer_tree_cache_insert): Likewise.
13897 (streamer_tree_cache_insert_at): Rename to ...
13898 (streamer_tree_cache_replace_tree): ... this and adjust.
13899 (streamer_tree_cache_append): Adjust.
13900 (record_common_node): Likewise.
13901 (streamer_tree_cache_create): Add argument whether to
13902 record hash values together with trees.
13903 (streamer_tree_cache_delete): Adjust.
13904 * tree-streamer.h (struct streamer_tree_cache_d): Add
13906 (streamer_read_integer_cst): Remove.
13907 (streamer_tree_cache_insert): Adjust.
13908 (streamer_tree_cache_append): Likewise.
13909 (streamer_tree_cache_insert_at): Rename to ...
13910 (streamer_tree_cache_replace_tree): ... this and adjust.
13911 (streamer_tree_cache_create): Add argument whether to record hashes.
13912 (streamer_tree_cache_get): Rename to ...
13913 (streamer_tree_cache_get_tree): ... this.
13914 (streamer_tree_cache_get_hash): New function.
13915 * tree.c (cache_integer_cst): New function.
13916 * tree.h (cache_integer_cst): Declare.
13917 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13918 * lto-symtab.c (lto_varpool_replace_node): Only release
13919 DECL_INITIAL of non-prevailing decls.
13920 * varpool.c (varpool_remove_initializer): Do not release
13921 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13923 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13925 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13926 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13927 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13928 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13929 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13930 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13931 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13932 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13933 instead of TARGET_64BIT.
13934 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13935 Require ISA_HAS_<D>DIV.
13937 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13939 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13940 (mips*-*-linux*): Move default with_llsc setting to where other
13942 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13944 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13945 Likewise. Remove default with_tune setting. Move default float
13946 setting to its own block. Handle with_llsc in the same block as above.
13948 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
13950 PR rtl-optimization/57425
13951 PR rtl-optimization/57569
13952 * alias.c (write_dependence_p): Add new parameters mem_mode,
13953 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13954 Changed all callers.
13955 (canon_anti_dependence): New function.
13956 * cse.c (check_dependence): Use canon_anti_dependence.
13957 * cselib.c (cselib_invalidate_mem): Likewise.
13958 * rtl.h (canon_anti_dependence): Declare.
13960 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13962 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13963 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13964 ".set mips3" for 64-bit targets.
13966 2013-06-15 Dehao Chen <dehao@google.com>
13968 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13969 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13970 (gimple_check_call_args): Likewise.
13971 * value-prof.c (check_ic_target): Likewise.
13972 * ipa-inline.c (early_inliner): Likewise.
13973 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13974 * cgraph.c (cgraph_create_edge_1): Likewise.
13975 (cgraph_make_edge_direct): Likewise.
13977 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
13980 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
13981 rs6000_output_move_128bit to handle emitting quad memory
13982 operations. Set attribute length to 8 bytes.
13984 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
13986 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
13988 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
13989 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
13990 (aarch64_<su>mlsl<mode>): Likewise.
13992 2013-06-14 Mike Stump <mikestump@comcast.net>
13994 * Makefile.in (TARGET_H): Add insn-codes.h.
13996 2013-06-14 Alan Modra <amodra@gmail.com>
13998 PR middle-end/57134
13999 PR middle-end/57586
14000 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14001 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14002 bitfield expansion when EXPAND_MEMORY.
14003 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14005 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14007 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14008 test for clearing quad memory on 32-bit later.
14010 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14012 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14013 (fold_negate_expr): Likewise.
14014 (fold_real_zero_addition_p): Handle vectors.
14015 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14017 2013-06-14 Alan Modra <amodra@gmail.com>
14019 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14021 2013-06-13 Jan Hubicka <jh@suse.cz>
14023 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14024 Local comdats are not externally visible.
14025 * symtab.c (dump_symtab_base): Dump externally visible.
14026 (verify_symtab_base): Verify back links in the symtab hash.
14028 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14030 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14031 CONVERT_EXPR as equal nodes.
14033 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14035 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14037 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14039 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14040 Generalize to complex and vector.
14041 * tree.c (build_all_ones_cst): New function.
14042 * tree.h (build_all_ones_cst): Declare it.
14044 2013-06-13 Alan Modra <amodra@gmail.com>
14046 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14047 * config/rs6000/rs6000.md (signbittf2): New insn.
14048 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14049 (abstf2_internal, cmptf_internal2): Likewise.
14050 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14052 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14053 Pat Haugen <pthaugen@us.ibm.com>
14054 Peter Bergner <bergner@vnet.ibm.com>
14056 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14057 power8 byte, half-word, and quad-word atomic instructions.
14058 (emit_store_conditional): Likewise.
14059 (rs6000_expand_atomic_compare_and_swap): Likewise.
14060 (rs6000_expand_atomic_op): Likewise.
14062 * config/rs6000/sync.md (larx): Add new modes for power8.
14064 (AINT): New mode iterator to include TImode as well as normal
14065 integer modes on power8.
14066 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14067 that VSX registers are not considered. Use AINT mode iterator
14068 instead of INT1 to allow inclusion of quad word atomic operations
14070 (load_locked<mode>): Likewise.
14071 (store_conditional<mode>): Likewise.
14072 (atomic_compare_and_swap<mode>): Likewise.
14073 (atomic_exchange<mode>): Likewise.
14074 (atomic_nand<mode>): Likewise.
14075 (atomic_fetch_<fetchop_name><mode>): Likewise.
14076 (atomic_nand_fetch<mode>): Likewise.
14077 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14079 (ATOMIC): On power8, add QImode, HImode modes.
14080 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14081 modes that promote to SImode.
14082 (load_lockedti): Convert TImode arguments to PTImode, so that we
14083 get a guaranteed even/odd register pair.
14084 (load_lockedpti): Likewise.
14085 (store_conditionalti): Likewise.
14086 (store_conditionalpti): Likewise.
14088 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14089 atomic load/store instructions.
14092 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14094 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14096 (insn_count): New attribute, with most cases extracted from...
14097 (length): ...here. Redefine most cases in terms of insn_count.
14098 (single_insn): Delete.
14099 (can_delay): Use insn_count to check for single instructions.
14100 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14101 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14102 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14103 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14104 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14105 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14106 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14107 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14108 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14109 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14110 rather than "length".
14111 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14112 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14113 Use "insn_count" rather than "length".
14114 * config/mips/mips-dsp.md
14115 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14116 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14119 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14121 PR tree-optimization/57361
14122 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14124 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14126 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14128 (aarch64_simd_combine<mode>): New instruction expansion.
14129 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14130 function prototype.
14131 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14132 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14134 2013-06-12 Jan Hubicka <jh@suse.cz>
14136 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14137 decl has when in streaming stage.
14138 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14139 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14141 2013-06-12 Roland Stigge <stigge@antcom.de>
14144 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14146 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14148 PR tree-optimization/57537
14149 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14150 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14152 2013-06-12 Richard Biener <rguenther@suse.de>
14154 * data-streamer.h (streamer_write_char_stream): CSE
14155 obs->current_pointer.
14156 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14157 streamer_write_char_stream manually and optimize the resulting loop.
14158 (streamer_write_hwi_stream): Likewise.
14160 2013-06-12 Jan Hubicka <jh@suse.cz>
14162 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14163 * cgraph.h (varpool_create_empty_node): Declare.
14164 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14166 * symtab.c (symtab_unregister_node): Be lax about missin entries
14168 (symtab_get_node): Update comment.
14169 * varpool.c (varpool_create_empty_node): Break out from ...
14170 (varpool_node_for_decl): ... here.
14171 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14173 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14175 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14176 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14177 part. Use straight-line flow at the end.
14178 <COMPONENT_REF>: Remove superfluous else.
14179 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14181 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14184 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14185 target hook even for !TREE_PUBLIC decls. If no resolution info
14186 is available, return false for common and external decls.
14188 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14190 * config/rl78/constraints.md (U): New constraint.
14191 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14194 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14197 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14198 to allow returning address to AT_PLATFORM name.
14200 2013-06-11 Jan Hubicka <jh@suse.cz>
14202 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14203 * cgraph.h (symtab_node_base): Add weakref flag.
14204 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14205 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14206 (output_weakrefs): Use weakref flag.
14207 * fold-const.c (simple_operand_p): Handle WEAK.
14208 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14209 * ipa.c (varpool_externally_visible_p): Drop weakref.
14210 (function_and_variable_visibility): Update comment; fix weakref
14211 sanity checks; do not clear DECL_WEAK on them.
14212 * lto-cgraph.c (lto_output_node): update.
14213 (lto_output_varpool_node): Update.
14214 (input_overwrite_node): Update.
14215 (input_node): Update.
14216 (input_varpool_node): Update.
14217 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14218 (lto_symtab_merge_symbols): Add sanity check.
14219 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14220 * passes.c (rest_of_decl_compilation): Set static flag, too.
14221 * symtab.c (dump_symtab_base): Dump weakref.
14222 (verify_symtab_base): Sanity check weakrefs.
14223 (symtab_make_decl_local): Remove duplicated code.
14224 (symtab_alias_ultimate_target): Simplify.
14225 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14227 2013-06-11 Tom de Vries <tom@codesourcery.com>
14229 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14230 sequence_vect == NULL.
14232 2013-06-11 DJ Delorie <dj@redhat.com>
14234 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14235 (rl78_unwind_word_mode): New.
14237 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14239 * final.c (debug_prefix_maps): Make static.
14241 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14243 * function.c (initial_trampoline): Remove stray copy.
14245 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14247 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14249 2013-06-11 Martin Jambor <mjambor@suse.cz>
14251 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14252 within bounds at the beginning of the function.
14254 2013-06-11 Alan Modra <amodra@gmail.com>
14256 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14258 (get_named_section): Don't NULL !DECL_P decl.
14260 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14262 * doc/invoke.texi (core-avx2): Document.
14264 (atom): Updated with MOVBE.
14266 2013-06-11 Richard Biener <rguenther@suse.de>
14268 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14270 2013-06-11 Anton Blanchard <anton@samba.org>
14272 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14273 correct shift value in little-endian mode.
14275 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14278 * varasm.c (get_variable_align): Move #endif to the right place.
14280 2013-06-10 Cary Coutant <ccoutant@google.com>
14282 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14283 for hash so that hash table traversal order is deterministic.
14285 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14286 Pat Haugen <pthaugen@us.ibm.com>
14287 Peter Bergner <bergner@vnet.ibm.com>
14289 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14290 of vectors in GPRS if they are direct moves or quad word load or
14293 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14295 (direct_move_p): Likewise.
14296 (quad_load_store_p): Likewise.
14298 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14299 classes into bins based on the physical register type.
14300 (reg_class_to_reg_type): Likewise.
14301 (IS_STD_REG_TYPE): Likewise.
14302 (IS_FP_VECT_REG_TYPE): Likewise.
14303 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14304 use direct move instructions.
14305 (reload_gpr_vsx): Likewise.
14306 (reload_vsx_gpr): Likewise.
14307 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14308 information that is a simplification of register classes. Also
14309 precalculate direct move reload helpers.
14310 (direct_move_p): New function to return true if the operation can
14311 be done as a direct move instruciton.
14312 (quad_load_store_p): New function to return true if the operation
14313 is a quad memory operation.
14314 (rs6000_legitimize_address): If quad memory, only allow register
14315 indirect for TImode addresses.
14316 (rs6000_legitimate_address_p): Likewise.
14317 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14318 (rs6000_reload_register_type): Likewise.
14319 (register_to_reg_type): Return register type.
14320 (rs6000_secondary_reload_simple_move): New helper function for
14321 secondary reload and secondary memory needed to identify anything
14322 that is a simple move, and does not need reloading.
14323 (rs6000_secondary_reload_direct_move): New helper function for
14324 secondary reload to identify cases that can be done with several
14325 instructions via the direct move instructions.
14326 (rs6000_secondary_reload_move): New helper function for secondary
14327 reload to identify moves between register types that can be done.
14328 (rs6000_secondary_reload): Add support for quad memory operations
14329 and for direct move.
14330 (rs6000_secondary_memory_needed): Likewise.
14331 (rs6000_debug_secondary_memory_needed): Change argument names.
14332 (rs6000_output_move_128bit): New function to return the move to
14333 use for 128-bit moves, including knowing about the various
14334 limitations of quad memory operations.
14336 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14337 memory operations. call rs6000_output_move_128bit for the actual
14338 instruciton(s) to generate.
14339 (vsx_movti_64bit): Likewise.
14341 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14342 (UNSPEC_P8V_MTVSRWZ): Likewise.
14343 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14344 (UNSPEC_P8V_MTVSRD): Likewise.
14345 (UNSPEC_P8V_XXPERMDI): Likewise.
14346 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14347 (UNSPEC_FUSION_GPR): Likewise.
14348 (FMOVE128_GPR): New iterator for direct move.
14349 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14350 (f32_sv): Likewise.
14351 (f32_dm): Likewise.
14352 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14353 loads and direct move instructions.
14354 (zero_extendsidi2_lfiwzx): Likewise.
14355 (extendsidi2_lfiwax): Likewise.
14356 (extendsidi2_nocell): Likewise.
14357 (floatsi<mode>2_lfiwax): Likewise.
14358 (lfiwax): Likewise.
14359 (floatunssi<mode>2_lfiwzx): Likewise.
14360 (lfiwzx): Likewise.
14361 (fix_trunc<mode>_stfiwx): Likewise.
14362 (fixuns_trunc<mode>_stfiwx): Likewise.
14363 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14364 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14365 (parity<mode>2_cmpb): Set length/type attr.
14366 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14367 for 'mr.' to fast_compare.
14368 (bpermd_<mode>): Change type attr to popcnt.
14369 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14370 (p8_mtvsrwz_1): Likewise.
14371 (p8_mtvsrwz_2): Likewise.
14372 (reload_fpr_from_gpr<mode>): Likewise.
14373 (p8_mtvsrd_1): Likewise.
14374 (p8_mtvsrd_2): Likewise.
14375 (p8_xxpermdi_<mode>): Likewise.
14376 (reload_vsx_from_gpr<mode>): Likewise.
14377 (reload_vsx_from_gprsf): Likewise.
14378 (p8_mfvsrd_3_<mode>): LIkewise.
14379 (reload_gpr_from_vsx<mode>): Likewise.
14380 (reload_gpr_from_vsxsf): Likewise.
14381 (p8_mfvsrd_4_disf): Likewise.
14382 (multi-word GPR splits): Do not split direct moves or quad memory
14385 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14387 * tree-into-ssa.c (interesting_blocks): Make static.
14389 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14392 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14393 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14394 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14395 (get_variable_align): New function.
14396 (get_variable_section, emit_bss, emit_common,
14397 assemble_variable_contents, place_block_symbol): Use
14398 get_variable_align instead of DECL_ALIGN.
14399 (assemble_noswitch_variable): Add align argument, use it
14400 instead of DECL_ALIGN.
14401 (assemble_variable): Adjust caller. Use get_variable_align
14402 instead of DECL_ALIGN.
14403 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14405 (DATA_ABI_ALIGNMENT): Define.
14406 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14407 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14408 opt is false, only return the psABI mandated alignment increase.
14409 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14410 (DATA_ABI_ALIGNMENT): ... this.
14411 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14412 (DATA_ABI_ALIGNMENT): ... this.
14413 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14414 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14415 (DATA_ABI_ALIGNMENT): ... this.
14416 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14417 * doc/tm.texi: Regenerated.
14419 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14421 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14422 cmp_code to construct REG_EQUAL note.
14424 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14427 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14428 that operands[2] doesn't overlap with operands[0].
14430 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14431 Jan Hubicka <jh@suse.cz>
14433 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14434 hack to mark symbols as used.
14436 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14438 PR rtl-optimization/57559
14439 * lra-constraints.c (process_alt_operands): Don't discourage
14440 memory with known offset for offsetable memory constraint.
14441 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14443 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14445 * varasm.c (struct oc_local_state): Reorder fields.
14446 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14447 and adjust accordingly.
14448 (output_constructor): Reorder initialization code and adjust call to
14449 output_constructor_bitfield.
14451 2013-06-07 Jan Hubicka <jh@suse.cz>
14453 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14455 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14457 * tree-object-size.c (unknown): Make const.
14459 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14461 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14462 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14463 for last alternative in the cpu_facility attribute.
14465 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14468 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14469 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14470 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14471 * config/arm/constraints.md (Dg): New constraint.
14472 * config/arm/neon.md (xordi3_neon): Remove.
14473 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14474 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14476 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14478 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14479 Clean up alternatives.
14481 2013-06-07 Alan Modra <amodra@gmail.com>
14483 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14486 2013-06-07 Alan Modra <amodra@gmail.com>
14488 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14490 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14492 * config/arm/constraints.md (Df): New constraint.
14493 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14494 Correct length attribute for last two alternatives.
14496 2013-06-07 Alan Modra <amodra@gmail.com>
14498 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14499 override user -mfp-in-toc.
14500 (offsettable_ok_by_alignment): Consider just the current access
14501 rather than the whole object, unless BLKmode. Handle
14502 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14503 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14504 for -mcmodel=medium.
14505 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14506 override user -mfp-in-toc or -msum-in-toc. Default to
14507 -mno-fp-in-toc for -mcmodel=medium.
14509 2013-06-06 DJ Delorie <dj@redhat.com>
14511 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14512 TARGET_VALID_POINTER_MODE.
14514 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14515 Pat Haugen <pthaugen@us.ibm.com>
14516 Peter Bergner <bergner@vnet.ibm.com>
14518 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14519 Document new power8 builtins.
14521 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14522 condition code register, to allow 128-bit logical operations to be
14523 done in the VSX or GPR registers.
14524 (nor<mode>3): Use the canonical form for nor.
14525 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14526 vclz*, and vpopcnt* vector instructions.
14527 (nand<mode>3): Likewise.
14528 (orc<mode>3): Likewise.
14529 (clz<mode>2): LIkewise.
14530 (popcount<mode>2): Likewise.
14532 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14533 that only the GPRs are recognized.
14535 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14536 support for new power8 builtins.
14538 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14540 (xscvdpspn): Likewise.
14546 (vpopcnt): Likewise.
14547 (vpopcntb): Likewise.
14548 (vpopcnth): Likewise.
14549 (vpopcntw): Likewise.
14550 (vpopcntd): Likewise.
14552 (vmrgew): Likewise.
14553 (vmrgow): Likewise.
14555 (eqv_v16qi3): Likewise.
14556 (eqv_v8hi3): Likewise.
14557 (eqv_v4si3): Likewise.
14558 (eqv_v2di3): Likewise.
14559 (eqv_v4sf3): Likewise.
14560 (eqv_v2df3): Likewise.
14562 (nand_v16qi3): Likewise.
14563 (nand_v8hi3): Likewise.
14564 (nand_v4si3): Likewise.
14565 (nand_v2di3): Likewise.
14566 (nand_v4sf3): Likewise.
14567 (nand_v2df3): Likewise.
14569 (orc_v16qi3): Likewise.
14570 (orc_v8hi3): Likewise.
14571 (orc_v4si3): Likewise.
14572 (orc_v2di3): Likewise.
14573 (orc_v4sf3): Likewise.
14574 (orc_v2df3): Likewise.
14576 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14577 allow power8 quad mode in 64-bit.
14578 (rs6000_builtin_vectorized_function): Add support to vectorize
14579 ISA 2.07 count leading zeros, population count builtins.
14580 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14581 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14582 (builtin_function_type): Add vgbbd builtin function which takes an
14584 (altivec_expand_vec_perm_const): Add support for new power8 merge
14587 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14588 that does not include TImdoe for use with 32-bit.
14589 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14591 (UNSPEC_VSX_CVDPSPN): Likewise.
14592 (vsx_xscvdpspn): Likewise.
14593 (vsx_xscvspdpn): Likewise.
14594 (vsx_xscvdpspn_scalar): Likewise.
14595 (vsx_xscvspdpn_directmove): Likewise.
14596 (vsx_and<mode>3): Split logical operations into 32-bit and
14597 64-bit. Add support to do logical operations on TImode as well as
14598 VSX vector types. Allow logical operations to be done in either
14599 VSX registers or in general purpose registers in 64-bit mode. Add
14600 splitters if GPRs were used. For AND, add clobber of CCmode to
14601 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14603 (vsx_and<mode>3_32bit): Likewise.
14604 (vsx_and<mode>3_64bit): Likewise.
14605 (vsx_ior<mode>3): Likewise.
14606 (vsx_ior<mode>3_32bit): Likewise.
14607 (vsx_ior<mode>3_64bit): Likewise.
14608 (vsx_xor<mode>3): Likewise.
14609 (vsx_xor<mode>3_32bit): Likewise.
14610 (vsx_xor<mode>3_64bit): Likewise.
14611 (vsx_one_cmpl<mode>2): Likewise.
14612 (vsx_one_cmpl<mode>2_32bit): Likewise.
14613 (vsx_one_cmpl<mode>2_64bit): Likewise.
14614 (vsx_nor<mode>3): Likewise.
14615 (vsx_nor<mode>3_32bit): Likewise.
14616 (vsx_nor<mode>3_64bit): Likewise.
14617 (vsx_andc<mode>3): Likewise.
14618 (vsx_andc<mode>3_32bit): Likewise.
14619 (vsx_andc<mode>3_64bit): Likewise.
14620 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14621 and xxlorc instructions.
14622 (vsx_eqv<mode>3_64bit): Likewise.
14623 (vsx_nand<mode>3_32bit): Likewise.
14624 (vsx_nand<mode>3_64bit): Likewise.
14625 (vsx_orc<mode>3_32bit): Likewise.
14626 (vsx_orc<mode>3_64bit): Likewise.
14628 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14630 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14632 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14633 (p8_vmrgow): Likewise.
14634 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14635 GPRs to be split under VSX.
14636 (p8v_clz<mode>2): Add power8 count leading zero support.
14637 (p8v_popcount<mode>2): Add power8 population count support.
14638 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14641 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14644 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14646 (vec_nand): Likewise.
14647 (vec_vclz): Likewise.
14648 (vec_vclzb): Likewise.
14649 (vec_vclzd): Likewise.
14650 (vec_vclzh): Likewise.
14651 (vec_vclzw): Likewise.
14652 (vec_vgbbd): Likewise.
14653 (vec_vmrgew): Likewise.
14654 (vec_vmrgow): Likewise.
14655 (vec_vpopcnt): Likewise.
14656 (vec_vpopcntb): Likewise.
14657 (vec_vpopcntd): Likewise.
14658 (vec_vpopcnth): Likewise.
14659 (vec_vpopcntw): Likewise.
14661 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14663 PR rtl-optimization/57468
14664 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14667 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14669 PR rtl-optimization/57459
14670 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14671 type when setting live regs.
14673 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14675 * config/s390/s390.opt (mlra): New option.
14676 * config/s390/s390.c (s390_decompose_address): Check displacement
14677 for all registers for LRA.
14678 (s390_secondary_reload): Don't used secondary reloads for LRA.
14679 (s390_lra_p): New function.
14680 (TARGET_LRA_P): Define.
14681 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14682 of attribute cpu_facility to zarch for the last alternative.
14683 (*cmpmem_short): Ditto.
14685 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14687 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14688 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14689 (arm_expand_prologue): Likewise.
14691 2013-06-06 Teresa Johnson <tejohnson@google.com>
14694 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14695 as this is now done by redirect_edge_and_branch_force.
14696 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14697 barriers, and fix interaction with splitting.
14698 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14699 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14700 reflect changes made in the routine.
14701 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14702 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14703 since this is called in cfglayout mode, and replace partition fixup
14704 with assert as that is now done by force_nonfallthru_and_redirect.
14705 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14706 already be marked with region crossing note.
14707 (insert_section_boundary_note): Make non-static, gate on flag
14708 has_bb_partition, rewrite to also check for multiple partitions.
14709 (rest_of_handle_reorder_blocks): Remove call to
14710 insert_section_boundary_note, now done later during free_cfg.
14711 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14712 * bb-reorder.h (insert_section_boundary_note): Declare.
14713 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14714 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14715 invoke insert_section_boundary_note.
14716 (try_redirect_by_replacing_jump): Remove unnecessary
14717 check for region crossing note.
14718 (fixup_partition_crossing): New function.
14719 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14720 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14721 in non-cfglayout mode.
14722 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14723 remove old code that tried to do this. Emit barrier correctly
14724 when we are in cfglayout mode.
14725 (last_bb_in_partition): New function.
14726 (rtl_split_edge): Correctly fixup partition boundaries.
14727 (commit_one_edge_insertion): Remove old code that tried to
14728 fixup region crossing edge since this is now handled in
14729 split_block, and set up insertion point correctly since
14730 block may now end in a jump.
14731 (verify_hot_cold_block_grouping): Guard against checking when not in
14732 linearized RTL mode.
14733 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14735 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14736 rtl_verify_flow_info, so not called in cfglayout mode.
14737 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14738 (fixup_reorder_chain): Remove old code that attempted to fixup region
14739 crossing note as this is now handled in force_nonfallthru_and_redirect.
14740 (duplicate_insn_chain): Don't duplicate switch section notes.
14741 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14743 * basic-block.h (emit_barrier_after_bb): Declare.
14745 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14747 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14748 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14749 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14751 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14753 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14755 * config/arm/ldmstm.md: Regenerate.
14757 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14759 * config/arm/sync.md (atomic_loaddi_1):
14760 Disable predication for arm_restrict_it.
14761 (arm_load_exclusive<mode>): Likewise.
14762 (arm_load_exclusivesi): Likewise.
14763 (arm_load_exclusivedi): Likewise.
14764 (arm_load_acquire_exclusive<mode>): Likewise.
14765 (arm_load_acquire_exclusivesi): Likewise.
14766 (arm_load_acquire_exclusivedi): Likewise.
14767 (arm_store_exclusive<mode>): Likewise.
14768 (arm_store_exclusive<mode>): Likewise.
14769 (arm_store_release_exclusivedi): Likewise.
14770 (arm_store_release_exclusive<mode>): Likewise.
14772 2013-06-06 Richard Biener <rguenther@suse.de>
14774 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14776 (lto_tag_is_tree_code_p): Adjust.
14777 (lto_tag_is_gimple_code_p): Likewise.
14778 (lto_gimple_code_to_tag): Likewise.
14779 (lto_tag_to_gimple_code): Likewise.
14780 (lto_tree_code_to_tag): Likewise.
14781 (lto_tag_to_tree_code): Likewise.
14782 * data-streamer.h (streamer_write_hwi_in_range): Use
14783 uhwi streaming to stream the normalized range.
14784 (streamer_read_hwi_in_range): Likewise.
14786 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14788 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14789 (predicable_short_it): Likewise.
14790 (predicated): Likewise.
14791 (enabled): Handle above.
14792 (define_cond_exec): Set predicated attribute to yes.
14794 2013-06-05 Mike Stump <mikestump@comcast.net>
14796 * gdbinit.in (__FUNCTION__): Add.
14798 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14800 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14801 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14803 2013-06-05 Jan Hubicka <jh@suse.cz>
14805 * varasm.c (mark_decl_referenced): Revert the removal until targets
14808 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14810 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14811 instead of mark_decl_referenced.
14813 2013-06-05 Jan Hubicka <jh@suse.cz>
14815 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14816 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14817 and symtab_used_from_object_file_p.
14818 (cgraph_make_node_local_1): Clear forced_by_abi.
14819 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14820 * cgraph.h (symtab_node_base): Add forced_by_abi.
14821 (decide_is_variable_needed): Remove.
14822 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14823 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14824 (decide_is_symbol_needed): ... this one; handle symbols in general;
14825 always analyze virtuals; honnor forced_by_abi.
14826 (cgraph_finalize_function): Update.
14827 (varpool_finalize_decl): Update.
14828 (symbol_defined_and_needed): Remove.
14829 (analyze_functions): Update.
14830 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14831 output_refs, input_overwrite_node): Handle forced_by_abi.
14832 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14833 (address_taken_from_non_vtable_p): ... this one.
14834 (comdat_can_be_unshared_p_1): New function.
14835 (cgraph_comdat_can_be_unshared_p): Rename to ...
14836 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14837 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14838 (function_and_variable_visibility): Clear forced_by_abi as needed.
14839 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14840 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
14841 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14842 * varpool.c (decide_is_variable_needed): Remove.
14844 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14846 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14847 (arm_option_override): Override arm_restrict_it where appropriate.
14848 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14849 * config/arm/arm.opt (mrestrict-it): New command-line option.
14850 * doc/invoke.texi: Document -mrestrict-it.
14852 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14854 * tsan.c (tsan_atomic_table): Make const.
14856 2013-06-05 Richard Biener <rguenther@suse.de>
14858 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14859 index associated with the tree we are supposed to replace.
14860 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14861 TREE_ASM_WRITTEN as zero for everything but SSA names.
14863 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14865 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14867 2013-06-04 Jan Hubicka <jh@suse.cz>
14869 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14870 (input_node, input_varpool_node): Handle correctly external same
14872 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14873 nodes at ltrans stage.
14875 2013-06-04 Jan Hubicka <jh@suse.cz>
14877 * ipa-inline.c (update_caller_keys): Fix availability test.
14878 (update_callee_keys): Likewise.
14879 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14880 to follow ELF standard.
14882 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14884 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14885 (mips64r5900el-*-elf*): New configurations.
14886 * config/mips/mips-cpus.def (r5900): New processor.
14887 * config/mips/mips-tables.opt: Regenerate.
14888 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14889 (mips_issue_rate): Handle PROCESSOR_R5900.
14890 (mips_reorg_process_insns): Force reorder mode for the R5900.
14891 * config/mips/mips.h (TARGET_MIPS5900): Define.
14892 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14894 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14896 * config/mips/mips.md (processor): Add r5900.
14897 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14899 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14901 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14902 into function to generate MOVI instruction.
14903 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
14904 (aarch64_preferred_simd_mode): Turn into wrapper.
14905 (aarch64_output_scalar_simd_mov_immediate): New function.
14906 * config/aarch64/aarch64-protos.h: Add prototype for above.
14908 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14910 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14911 element_char member.
14912 (sizetochar): Return signed char.
14913 (aarch64_simd_valid_immediate): Remove elchar and other
14914 unnecessary variables.
14915 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14916 Calculate element_char as required.
14917 * config/aarch64/aarch64-protos.h: Update and move prototype
14918 for aarch64_output_simd_mov_immediate.
14919 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14922 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14924 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14925 information completed by aarch64_simd_valid_immediate.
14926 (aarch64_legitimate_constant_p): Update arguments.
14927 (aarch64_simd_valid_immediate): Work with struct rather than many
14929 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14930 (aarch64_simd_make_constant): Update arguments.
14931 (aarch64_output_simd_mov_immediate): Work with struct rather than
14932 many pointers. Output immediate directly rather than as operand.
14933 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14935 * config/aarch64/constraints.md (Dn): Update arguments.
14937 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14939 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14941 (aarch64_simd_immediate_valid_for_move): Remove.
14942 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14943 (aarch64_simd_make_constant): Update call.
14944 (aarch64_output_simd_mov_immediate): Update call.
14945 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14947 * config/aarch64/constraints.md (Dn): Update call.
14949 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14951 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14952 return type to bool for prototype.
14953 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14954 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14955 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14957 2013-06-04 Catherine Moore <clm@codesourcery.com>
14959 * config/mips/mips.opt (meva): New.
14960 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
14961 (ASM_SPEC): Handle -meva.
14962 * doc/invoke.texi (meva): Document.
14964 2013-06-04 Alan Modra <amodra@gmail.com>
14966 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14969 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14971 * rtl.def: Add extra fourth optional field to define_cond_exec.
14972 * gensupport.c (process_one_cond_exec): Process attributes from
14974 * doc/md.texi: Document fourth field in define_cond_exec.
14976 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14978 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14979 out the processing order as in store_bit_field_1.
14981 2013-06-04 Jan Hubicka <jh@suse.cz>
14983 PR middle-end/57500
14984 * cgraphunit.c (cgraph_process_same_body_aliases): Create
14985 non-VAR_DECL node if it does not exist yet.
14987 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
14989 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
14990 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
14991 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
14992 target_cpu_default setting.
14994 2013-06-03 Teresa Johnson <tejohnson@google.com>
14996 * dumpfile.c (opt_info_switch_p): Change -fopt-info
14997 default to -fopt-info=optimized instead of all.
14998 * doc/invoke.texi: Ditto.
14999 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15000 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15001 (execute_vect_slp): Emit BB vectorization success under
15002 MSG_OPTIMIZED_LOCATIONS.
15003 * tree-vect-slp.c (vect_slp_transform_bb): Change
15004 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15005 * tree-vect-loop.c (vect_transform_loop): Ditto.
15007 2013-06-03 Jason Merrill <jason@redhat.com>
15010 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15011 Use TARGET_EXPR for C++.
15013 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15015 PR rtl-optimization/57268
15016 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15017 if DEBUG_INSN_P (insn).
15020 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15022 PR rtl-optimization/57268
15023 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15024 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15026 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15028 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15029 (ix86_avoid_lea_for_addr): Likewise.
15030 (exact_dependency_1): Likewise.
15031 (ix86_adjust_cost): Likewise.
15032 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15034 (do_reorder_for_imul): Fix typo, formatting and
15035 !reload_completed check removed.
15036 (ix86_sched_reorder): Fix typo and formatting.
15037 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15040 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15042 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15044 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15046 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15047 <STRING_CST>: Likewise.
15048 <VECTOR_CST>: Likewise.
15050 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15051 Mikael Morin <mikael@gcc.gnu.org>
15053 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15054 * config.in: Regenerated.
15055 * configure: Regenerated.
15057 2013-06-01 Jan Hubicka <jh@suse.cz>
15059 PR middle-end/57366
15060 * cgraphunit.c (compile): When weakref is not supported,
15061 set up transparent aliases before final output pass.
15062 * varasm.c (assemble_alias): Do not try to do it here.
15064 2013-06-01 Jan Hubicka <jh@suse.cz>
15066 PR middle-end/57467
15067 * passes.c (for_per_function): Skip unanalyzed functions.
15069 2013-06-01 Jan Hubicka <jh@suse.cz>
15071 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15072 (lto_symtab_merge_symbols_1): ... this one.
15073 (lto_symtab_merge_cgraph_nodes): Rename to ...
15074 (lto_symtab_merge_symbols): ... this one; simplify.
15075 * cgraph.c (same_body_aliases_done): Rename to ...
15076 (cpp_implicit_aliases_done): ... this one.
15077 (cgraph_create_function_alias): Update.
15078 (cgraph_same_body_alias): Update.
15079 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15080 (verify_edge_corresponds_to_fndecl): Simplify.
15081 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15082 (cgraph_node): Remove same_body_alias.
15083 (varpool_node): Remove alias_of and extra_name_alias.
15084 (same_body_aliases_done): Rename to ..
15085 (cpp_implicit_aliases_done): ... this one.
15086 (symtab_alias_ultimate_target): Add default parameter.
15087 (symtab_resolve_alias): New function.
15088 (fixup_same_cpp_alias_visibility): Declare.
15089 (cgraph_function_node): Add default parameter.
15090 (cgraph_node_asm_name): Likewise.
15091 (cgraph_function_or_thunk_node): Add default parameter; do
15092 not ICE when it is NULL.
15093 (varpool_variable_node): Likewise.
15094 * tree-emutls.c (create_emultls_var): Update.
15095 (ipa_lower_emutls): Update.
15096 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15097 (cgraph_reset_node): Reset alias info.
15098 (cgraph_finalize_function): Update.
15099 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15100 (analyze_function): Simplify.
15101 (cgraph_process_same_body_aliases): Simplify.
15102 (analyze_functions): Fixup same body aliases.
15103 (handle_alias_pairs): Simplify.
15104 (assemble_thunk): Update.
15105 (assemble_thunks_and_aliases): Update.
15106 (output_weakrefs): Rewrite.
15107 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15108 (lto_output_varpool_node): Likewise.
15109 (compute_ltrans_boundary): Remve assert.
15110 (get_alias_symbol): New functoin.
15111 (input_node): Rewrite alias handling.
15112 (input_varpool_node): Likewise.
15113 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15114 * ipa.c (process_references): Handle weakrefs correctly.
15115 (symtab_remove_unreachable_nodes): Likewise.
15116 * trans-mem.c (get_cg_data): Update.
15117 (ipa_tm_create_version_alias): Update.
15118 (ipa_tm_execute): Update.
15119 * symtab.c (dump_symtab_base): Dump aliases.
15120 (verify_symtab_base): Verify aliases.
15121 (symtab_node_availability): New function.
15122 (symtab_alias_ultimate_target): Simplify.
15123 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15124 handle all the fixup cases.
15125 (symtab_resolve_alias): New function.
15126 * passes.c (ipa_write_summaries): Handle weakrefs.
15127 * varpool.c (varpool_analyze_node): Simplify.
15128 (assemble_aliases): Update.
15129 (varpool_create_variable_alias): Simplify.
15130 (varpool_extra_name_alias): Simplify.
15131 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15132 (lto_symtab_merge_symbols): ... this one.
15134 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15137 PR rtl-optimization/57268
15138 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15139 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15141 2013-06-01 Tobias Burnus <burnus@net-b.de>
15143 Partially reverted:
15144 2013-05-31 Tobias Burnus <burnus@net-b.de>
15146 PR middle-end/57073
15147 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15150 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15152 PR rtl-optimization/57268
15153 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15154 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15156 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15158 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15159 unordered comparison operators when -fno-trapping-math is in effect
15161 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15162 and implement unordered comparison operators properly on the e500.
15164 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15166 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15167 for constant scalar integers.
15168 (simplify_relational_operation_1): Likewise.
15170 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15172 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15173 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15176 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15177 Igor Zamyatin <igor.zamyatin@intel.com>
15179 Silvermont (SLM) architecture performance tuning.
15180 * config/i386/i386.h (enum ix86_tune_indices): Add
15181 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15182 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15184 * config/i386/i386.c (initial_ix86_tune_features)
15185 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15186 (ix86_lea_outperforms): Handle Silvermont tuning.
15187 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15189 (ix86_use_lea_for_mov): Likewise.
15190 (ix86_avoid_lea_for_addr): Likewise.
15191 (ix86_lea_for_add_ok): Likewise.
15192 (exact_dependency_1): New function.
15193 (exact_store_load_dependency): Likewise.
15194 (ix86_adjust_cost): Handle Silvermont tuning.
15195 (do_reoder_for_imul): Likewise.
15196 (swap_top_of_ready_list): New function.
15197 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15199 * config/i386/i386.md (peepholes that split memory operand in fp
15202 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15204 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15205 Remove un-necessary braces.
15207 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15209 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15210 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15212 2013-05-31 Tobias Burnus <burnus@net-b.de>
15214 PR middle-end/57073
15215 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15217 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15220 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15221 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15222 * config/arm/neon.md (iordi3_neon): Remove.
15223 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15224 * config/arm/predicates.md (imm_for_neon_logic_operand):
15225 Move to earlier in the file.
15226 (neon_logic_op2): Likewise.
15227 (arm_iordi_operand_neon): New predicate.
15229 2013-05-31 Richard Biener <rguenther@suse.de>
15231 PR tree-optimization/57478
15232 PR tree-optimization/57453
15233 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15236 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15238 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15239 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15241 2013-05-30 Tobias Burnus <burnus@net-b.de>
15242 Thomas Koenig <tkoenig@gcc.gnu.org>
15244 PR middle-end/57073
15245 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15246 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15248 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15250 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15252 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15254 * target.def (register_usage_leveling_p): New hook.
15255 * targhooks.c (default_register_usage_leveling_p): New.
15256 * targhooks.h (default_register_usage_leveling_p): New prototype.
15257 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15258 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15259 * doc/tm.texi: Update.
15260 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15262 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15264 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15265 (*insv_reg<mode>): New define_insn.
15267 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15269 PR rtl-optimization/57439
15270 * postreload.c (move2add_valid_value_p): Check that we have
15271 a zero subreg_regno_offset when accessing the register in
15272 the requested mode.
15274 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15275 Igor Zamyatin <igor.zamyatin@intel.com>
15277 Silvermont (SLM) architecture pipeline model, tuning and
15279 * config.gcc: Add slm config options and target.
15281 * config/i386/slm.md: New.
15283 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15285 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15287 (ix86_target_macros_internal): Likewise.
15289 * config/i386/i386.c (slm_cost): New cost.
15290 (m_SLM): New macro flag.
15291 (initial_ix86_tune_features): Set m_SLM.
15292 (x86_accumulate_outgoing_args): Likewise.
15293 (x86_arch_always_fancy_math_387): Likewise.
15294 (processor_target_table): Add slm cost.
15295 (cpu_names): Add slm cpu name.
15296 (x86_option_override_internal): Set SLM ISA.
15297 (ix86_issue_rate): New case PROCESSOR_SLM.
15298 (ia32_multipass_dfa_lookahead): Likewise.
15299 (fold_builtin_cpu): Add slm.
15301 * config/i386/i386.h (TARGET_SLM): New target macro.
15302 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15303 (processor_type): Add PROCESSOR_SLM.
15305 * config/i386/i386.md (cpu): Add new value "slm".
15306 (slm.md): Include slm.md.
15308 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15309 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15311 * config/arm/arm-protos.h: Add and update function protos.
15312 * config/arm/arm.c (use_simple_return_p): New added.
15313 (thumb2_expand_return): Check simple_return flag.
15314 * config/arm/arm.md: Add simple_return and conditional simple_return.
15315 * config/arm/iterators.md: Add iterator for return and simple_return.
15317 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15319 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15320 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15321 (arm_emit_vfp_multi_reg_pop): Likewise.
15322 (thumb2_emit_ldrd_pop): Likewise.
15323 (arm_expand_epilogue): Add misc REG_CFA notes.
15324 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15326 2013-05-29 Lawrence Crowl <crowl@google.com>
15328 * config/arm/t-arm: Update for below.
15330 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15331 Change type to hash_table. Update dependent calls and types.
15333 * config/i386/t-cygming: Update for below.
15335 * config/i386/t-interix: Update for below.
15337 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15338 Change type to hash_table. Update dependent calls and types.
15339 (i386_find_on_wrapper_list::wrappers): Likewise.
15341 * config/ia64/t-ia64: Update for below.
15343 * config/ia64/ia64.c (bundle_state_table):
15344 Change type to hash_table. Update dependent calls and types.
15346 * config/mips/mips.c (mips_reorg_process_insns::htab):
15347 Change type to hash_table. Update dependent calls and types.
15349 * config/sol2.c (solaris_comdat_htab):
15350 Change type to hash_table. Update dependent calls and types.
15352 * config/t-sol2: Update for above.
15354 2013-05-29 Teresa Johnson <tejohnson@google.com>
15356 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15357 functions are not yet marked as defined.
15359 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15360 Pat Haugen <pthaugen@us.ibm.com>
15361 Peter Bergner <bergner@vnet.ibm.com>
15363 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15367 (vrotl<mode>3): Likewise.
15368 (vashl<mode>3): Likewise.
15369 (vlshr<mode>3): Likewise.
15370 (vashr<mode>3): Likewise.
15372 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15373 support for power8 V2DI builtins.
15375 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15376 power8 V2DI builtins.
15377 (vupkhsw): Likewise.
15378 (vupklsw): Likewise.
15379 (vaddudm): Likewise.
15380 (vminsd): Likewise.
15381 (vmaxsd): Likewise.
15382 (vminud): Likewise.
15383 (vmaxud): Likewise.
15384 (vpkudum): Likewise.
15385 (vpksdss): Likewise.
15386 (vpkudus): Likewise.
15387 (vpksdus): Likewise.
15392 (vsubudm): Likewise.
15393 (vcmpequd): Likewise.
15394 (vcmpgtsd): Likewise.
15395 (vcmpgtud): Likewise.
15396 (vcmpequd_p): Likewise.
15397 (vcmpgtsd_p): Likewise.
15398 (vcmpgtud_p): Likewise.
15399 (vupkhsw): Likewise.
15400 (vupklsw): Likewise.
15401 (vaddudm): Likewise.
15402 (vmaxsd): Likewise.
15403 (vmaxud): Likewise.
15404 (vminsd): Likewise.
15405 (vminud): Likewise.
15406 (vpksdss): Likewise.
15407 (vpksdus): Likewise.
15408 (vpkudum): Likewise.
15409 (vpkudus): Likewise.
15414 (vsubudm): Likewise.
15416 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15417 support for power8 V2DI instructions.
15419 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15420 power8 V2DI instructions. Combine pack and unpack insns to use an
15421 iterator for each mode. Check whether a particular mode supports
15422 Altivec instructions instead of just checking TARGET_ALTIVEC.
15423 (UNSPEC_VPKUWUM): Likewise.
15424 (UNSPEC_VPKSHSS): Likewise.
15425 (UNSPEC_VPKSWSS): Likewise.
15426 (UNSPEC_VPKUHUS): Likewise.
15427 (UNSPEC_VPKSHUS): Likewise.
15428 (UNSPEC_VPKUWUS): Likewise.
15429 (UNSPEC_VPKSWUS): Likewise.
15430 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15431 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15432 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15433 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15434 (UNSPEC_VUPKHSB): Likewise.
15435 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15436 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15437 (UNSPEC_VUPKHSH): Likewise.
15438 (UNSPEC_VUPKLSB): Likewise.
15439 (UNSPEC_VUPKLSH): Likewise.
15441 (VI_char): Likewise.
15442 (VI_scalar): Likewise.
15443 (VI_unit): Likewise.
15445 (VP_small): Likewise.
15446 (VP_small_lc): Likewise.
15447 (VU_char): Likewise.
15448 (add<mode>3): Likewise.
15449 (altivec_vaddcuw): Likewise.
15450 (altivec_vaddu<VI_char>s): Likewise.
15451 (altivec_vadds<VI_char>s): Likewise.
15452 (sub<mode>3): Likewise.
15453 (altivec_vsubcuw): Likewise.
15454 (altivec_vsubu<VI_char>s): Likewise.
15455 (altivec_vsubs<VI_char>s): Likewise.
15456 (altivec_vavgs<VI_char>): Likewise.
15457 (altivec_vcmpbfp): Likewise.
15458 (altivec_eq<mode>): Likewise.
15459 (altivec_gt<mode>): Likewise.
15460 (altivec_gtu<mode>): Likewise.
15461 (umax<mode>3): Likewise.
15462 (smax<mode>3): Likewise.
15463 (umin<mode>3): Likewise.
15464 (smin<mode>3): Likewise.
15465 (altivec_vpkuhum): Likewise.
15466 (altivec_vpkuwum): Likewise.
15467 (altivec_vpkshss): Likewise.
15468 (altivec_vpkswss): Likewise.
15469 (altivec_vpkuhus): Likewise.
15470 (altivec_vpkshus): Likewise.
15471 (altivec_vpkuwus): Likewise.
15472 (altivec_vpkswus): Likewise.
15473 (altivec_vpks<VI_char>ss): Likewise.
15474 (altivec_vpks<VI_char>us): Likewise.
15475 (altivec_vpku<VI_char>us): Likewise.
15476 (altivec_vpku<VI_char>um): Likewise.
15477 (altivec_vrl<VI_char>): Likewise.
15478 (altivec_vsl<VI_char>): Likewise.
15479 (altivec_vsr<VI_char>): Likewise.
15480 (altivec_vsra<VI_char>): Likewise.
15481 (altivec_vsldoi_<mode>): Likewise.
15482 (altivec_vupkhsb): Likewise.
15483 (altivec_vupkhs<VU_char>): Likewise.
15484 (altivec_vupkls<VU_char>): Likewise.
15485 (altivec_vupkhsh): Likewise.
15486 (altivec_vupklsb): Likewise.
15487 (altivec_vupklsh): Likewise.
15488 (altivec_vcmpequ<VI_char>_p): Likewise.
15489 (altivec_vcmpgts<VI_char>_p): Likewise.
15490 (altivec_vcmpgtu<VI_char>_p): Likewise.
15491 (abs<mode>2): Likewise.
15492 (vec_unpacks_hi_v16qi): Likewise.
15493 (vec_unpacks_hi_v8hi): Likewise.
15494 (vec_unpacks_lo_v16qi): Likewise.
15495 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15496 (vec_unpacks_lo_v8hi): Likewise.
15497 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15498 (vec_pack_trunc_v8h): Likewise.
15499 (vec_pack_trunc_v4si): Likewise.
15500 (vec_pack_trunc_<mode>): Likewise.
15502 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15504 (vec_vmaxsd): Likewise.
15505 (vec_vmaxud): Likewise.
15506 (vec_vminsd): Likewise.
15507 (vec_vminud): Likewise.
15508 (vec_vpksdss): Likewise.
15509 (vec_vpksdus): Likewise.
15510 (vec_vpkudum): Likewise.
15511 (vec_vpkudus): Likewise.
15512 (vec_vrld): Likewise.
15513 (vec_vsld): Likewise.
15514 (vec_vsrad): Likewise.
15515 (vec_vsrd): Likewise.
15516 (vec_vsubudm): Likewise.
15517 (vec_vupkhsw): Likewise.
15518 (vec_vupklsw): Likewise.
15520 2013-05-29 Jan Hubicka <jh@suse.cz>
15522 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15523 flags; reorder rest of fields in more consistent way.
15524 (varpool_node): Remove analyzed, finalized and alias.
15525 (cgraph_ndoe): Likewise.
15526 (symtab_alias_ultimate_target): New function.
15527 (cgraph_function_node): Move offline.
15528 (cgraph_reset_node): Declare.
15529 (cgraph_comdat_can_be_unshared_p): Remove.
15530 (varpool_remove_initializer): Declare.
15531 (varpool_first_defined_variable, varpool_next_defined_variable
15532 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15533 (cgraph_function_with_gimple_body_p): Update.
15534 (varpool_all_refs_explicit_p): Update.
15535 (symtab_alias_target): New function.
15536 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15537 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15538 (cgraph_function_or_thunk_node): Simplify using
15539 symtab_alias_ultimate_target.
15540 (varpool_variable_node): Likewise.
15541 * cgraph.c (cgraph_create_function_alias): Update.
15542 (cgraph_add_thunk): Update.
15543 (cgraph_remove_node): Update.
15544 (dump_cgraph_node): Do not dump removed flags.
15545 (cgraph_function_body_availability): Update.
15546 (cgraph_propagate_frequency): Update.
15547 (verify_cgraph_node): Check sanity of local flag.
15548 (cgraph_function_node): Move here from cgraph.h; revamp for
15549 cgraph_function_or_thunk_node.
15550 * lto-symtab.c (lto_varpool_replace_node): Update.
15551 (lto_symtab_resolve_can_prevail_p): Update.
15552 (lto_symtab_merge_cgraph_nodes): Update.
15553 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15554 propagate_constants_accross_call, devirtualization_time_bonus,
15555 ipcp_propagate_stage): Update.
15556 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15557 * ipa-inline-transform.c (clone_inlined_nodes,
15558 preserve_function_body_p): Update.
15559 * ipa-reference.c (propagate): Update.
15560 (write_node_summary_p): Update.
15561 * toplev.c (wrapup_global_declaration_2): Update.
15562 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15563 (analyze_function) ... this one.
15564 (cgraph_process_new_functions): Update.
15565 (cgraph_reset_node): Export.
15566 (cgraph_finalize_function): Update.
15567 (cgraph_add_new_function): Update.
15568 (process_function_and_variable_attributes): Update.
15569 (varpool_finalize_decl): Update.
15570 (symbol_finalized): Remove.
15571 (symbol_finalized_and_needed): Rename to ...
15572 (symbol_defined_and_needed): ... update.
15573 (cgraph_analyze_functions): Update.
15574 (handle_alias_pairs): Update.
15575 (mark_functions_to_output): Update.
15576 (assemble_thunk): Update.
15577 (output_in_order): Update.
15578 (output_weakrefs): Update.
15579 (finalize_compilation_unit): Update.
15580 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15581 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15582 input_node, input_varpool_node): Update.
15583 * dbxout.c (dbxout_expand_expr): Update.
15584 * cgraphclones.c (cgraph_clone_node): Update.
15585 (cgraph_copy_node_for_versioning): Update.
15586 (cgraph_materialize_clone): Update.
15587 (cgraph_materialize_all_clones): Update.
15588 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15589 propagate_pure_const, propagate_nothrow): Update.
15590 * lto-streamer-out.c (lto_output, write_symbol): Update.
15591 * ipa-utils.c (ipa_reverse_postorder): Update.
15592 * ipa-inline.c (can_inline_edge_p): Update.
15593 (update_caller_keys, ipa_inline): Update.
15594 * dwarf2out.c (reference_to_unused,
15595 premark_types_used_by_global_vars_helper): Update.
15596 * tree-eh.c (tree_could_trap_p): Update.
15597 * ipa-split.c (consider_split, execute_split_functions): Update.
15598 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15599 has_addr_references_p): Update; move ahead in file for better
15601 (process_references): Simplify.
15602 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15603 bodies are removed.
15604 (cgraph_comdat_can_be_unshared_p): Make static.
15605 (cgraph_externally_visible_p): Update.
15606 (varpool_externally_visible_p): Update.
15607 (function_and_variable_visibility): Update.
15608 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15609 ipa_tm_mark_force_output_node): Update.
15610 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15611 estimate_edge_devirt_benefit, inline_generate_summary,
15612 inline_write_summary): Update.
15613 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15614 * ipa-prop.c (ipa_compute_jump_functions): Update.
15615 (ipa_print_node_params, ipa_prop_read_section,
15616 ipa_update_after_lto_read, read_replacements_section): Update.
15617 * varasm.c (mark_decl_referenced): Update.
15618 (assemble_alias, dump_tm_clone_pairs): Update.
15619 * tree-inline.c (copy_bb): Update.
15620 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15622 * symtab.c (dump_symtab_base): Print new flags.
15623 (verify_symtab_base): Verify new flags.
15624 (symtab_alias_ultimate_target): New function.
15625 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15626 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15628 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15630 * i386.c (ix86_get_function_versions_dispatcher,
15631 ix86_generate_version_dispatcher_body): Update.
15632 (fold_builtin_cpu): Use varpool_add_new_variable.
15633 * varpool.c (varpool_remove_initializer): Break out from ...
15634 (varpool_remove_node): ... this one.
15635 (dump_varpool_node, varpool_node_for_asm,
15636 cgraph_variable_initializer_availability, varpool_analyze_node,
15637 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15638 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15641 2013-05-29 Jan Hubicka <jh@suse.cz>
15643 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15645 2013-05-29 Easwaran Raman <eraman@google.com>
15647 PR tree-optimization/57442
15648 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15649 when control exits the main loop.
15651 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15653 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15655 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15656 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15657 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15660 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15662 PR tree-optimization/57441
15663 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15664 Don't limit size of incr_vec to number of candidates.
15666 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15668 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15669 and mips16 directories.
15670 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15671 (MULTILIB_DIRNAMES): Ditto.
15672 (MULTILIB_EXCEPTIONS): Add new exceptions.
15673 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15674 (MULTILIB_DIRNAMES): Ditto.
15675 (MULTILIB_EXCEPTIONS): Add new exceptions.
15677 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15678 Marcus Shawcroft <marcus.shawcroft@arm.com>
15680 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15681 SYMBOL_TINY_ABSOLUTE.
15682 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15683 SYMBOL_TINY_ABSOLUTE.
15684 (aarch64_expand_mov_immediate): Likewise.
15685 (aarch64_classify_symbol): Likewise.
15686 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15687 Permit SYMBOL_TINY_ABSOLUTE.
15688 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15690 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15691 Marcus Shawcroft <marcus.shawcroft@arm.com>
15693 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15694 Refactor if/switch. Replace gcc_assert with if.
15696 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15698 * config/i386/i386.c (initial_ix86_tune_features): Enable
15699 FP Reassociation for AMD bdver1 and bdver2.
15701 2013-05-29 Martin Jambor <mjambor@suse.cz>
15703 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15704 and IMAGPART_EXPR do not occur within other handled_components.
15706 2013-05-29 Richard Biener <rguenther@suse.de>
15708 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15709 access on whether the use is in the BB we currently try to
15711 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15712 vectorize to vect_bb_slp_scalar_cost.
15714 2013-05-29 Richard Biener <rguenther@suse.de>
15716 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15717 computing scalar cost offsetted by stmts that are kept live
15719 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15720 for computation of scalar cost.
15722 2013-05-28 Steve Ellcey <sellcey@mips.com>
15724 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15726 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15728 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15729 array notation built-in reduction functions.
15730 * doc/passes.texi (Passes): Added documentation about changes done
15732 * doc/invoke.texi (C Dialect Options): Added documentation about
15733 the -fcilkplus flag.
15734 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15735 (BUILTINS_DEF): Depend on cilkplus.def.
15736 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15737 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15738 * cilkplus.def: New file.
15740 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15742 PR rtl-optimization/57439
15743 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15745 2013-05-28 Easwaran Raman <eraman@google.com>
15747 PR tree-optimization/57337
15748 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15749 (find_insert_point): Correctly identify the insertion point
15750 when two statements with the same UID is compared.
15752 2013-05-28 Richard Biener <rguenther@suse.de>
15754 PR tree-optimization/56787
15755 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15756 from the list of data references.
15757 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15759 (vect_analyze_loop_operations): Likewise.
15760 (vect_transform_loop): Remove clobbers.
15762 2013-05-28 Martin Jambor <mjambor@suse.cz>
15764 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15765 and REALPART_EXPRs have scalar type.
15767 2013-05-28 Richard Biener <rguenther@suse.de>
15769 PR tree-optimization/57411
15770 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15772 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15773 virtual operand propagation.
15775 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15777 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15778 destination register for bmasksi_vis.
15779 (vector_init_bshuffle): Likewise.
15780 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15782 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15784 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15785 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15786 mode if the instruction isn't available in the original mode.
15787 * config/sparc/sparc.opt (mfix-ut699): New option.
15788 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15789 (divdf3): Turn into expander.
15790 (divdf3_nofix): New insn.
15791 (divdf3_fix): Likewise.
15792 (divsf3): Disable if -mfix-ut699.
15793 (sqrtdf2): Turn into expander.
15794 (sqrtdf2_nofix): New insn.
15795 (sqrtdf2_fix): Likewise.
15796 (sqrtsf2): Disable if -mfix-ut699.
15798 2013-05-27 Richard Biener <rguenther@suse.de>
15800 PR middle-end/57412
15801 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15802 block for the new loop.
15804 2013-05-27 Richard Biener <rguenther@suse.de>
15806 PR tree-optimization/57343
15807 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15808 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15809 (number_of_iterations_cond): Do not build the folded tree.
15811 2013-05-27 Richard Biener <rguenther@suse.de>
15814 PR middle-end/57381
15815 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15816 OEP_CONSTANT_ADDRESS_OF retained.
15818 PR tree-optimization/57417
15819 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15820 for unchanged base.
15821 (set_ssa_val_to): Compare addresses using
15822 get_addr_base_and_unit_offset.
15824 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15826 PR rtl-optimization/56833
15827 * postreload.c (move2add_record_mode): New function.
15828 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15829 (move2add_use_add2_insn): Use move2add_record_sym_value.
15830 (move2add_use_add3_insn): Likewise.
15831 (reload_cse_move2add): Use move2add_valid_value_p and
15832 move2add_record_mode. Invalidate call-clobbered and REG_INC
15833 affected regs by setting reg_mode to VOIDmode.
15834 (move2add_note_store): Don't pretend the inside of a SUBREG is
15835 the actual destination. Invalidate single/leading registers by
15836 setting reg_mode to VOIDmode.
15837 Use move2add_record_sym_value, move2add_valid_value_p and
15838 move2add_record_mode.
15840 2013-05-27 Richard Biener <rguenther@suse.de>
15842 PR tree-optimization/57396
15843 * tree-affine.c (double_int_constant_multiple_p): Properly
15844 return false for val == 0 and div != 0.
15846 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15848 * config/mips/mips.h: Use #elif in preprocessor conditions.
15850 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15853 * config/mips/constraints.md (kl): New constraint.
15854 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15855 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15856 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15857 from LO for MIPS16.
15858 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15860 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15863 * config/mips/mips.c (mips_can_inline_p): New function.
15864 (TARGET_CAN_INLINE_P): Define.
15866 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15868 * sched-int.h (ds_t, dw_t): Make unsigned int.
15869 Fix documentation that describes how all the ds_t bits are used.
15870 Reserve the last bit for delayed-branch scheduling.
15871 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15872 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15873 (gen_dep_weak_1): Remove prototype.
15874 * sched-deps.c (get_dep_weak_1): Make static.
15875 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15876 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15877 * doc/tm.texi: Regenerate.
15878 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15880 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15883 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15885 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15886 Sandra Loosemore <sandra@codesourcery.com>
15888 * config.gcc (powerpc-*): Allow native for with-cpu.
15890 2013-05-24 Jeff Law <law@redhat.com>
15892 PR tree-optimization/57124
15893 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15894 conversion feeding a condition if the range has an overflow
15895 if -fstrict-overflow. Add warnings for when we do make the
15898 2013-05-24 Dehao Chen <dehao@google.com>
15900 * tree-cfg.c (locus_discrim_map): Fix the typo.
15901 (locus_discrim_hasher): Likewise.
15902 (locus_discrim_hasher::hash): Likewise.
15903 (locus_discrim_hasher::equal): Likewise.
15905 2013-05-24 Martin Jambor <mjambor@suse.cz>
15907 PR tree-optimization/57294
15908 * cgraph.h (ipa_record_stmt_references): Declare.
15909 * cgraphbuild.c (ipa_record_stmt_references): New function.
15910 (build_cgraph_edges): Use ipa_record_stmt_references.
15911 (rebuild_cgraph_edges): Likewise.
15912 (cgraph_rebuild_references): Likewise.
15913 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15914 associated with the old statement and build references from the
15915 newly built statements.
15916 * ipa-ref.c (ipa_remove_stmt_references): New function.
15917 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15919 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15921 * lra-constraints.c (emit_spill_move): Use smaller mode for
15923 (check_and_process_move): Consider mem-reg moves for secondary
15925 (curr_insn_transform): Don't lose insns emitted before for
15926 secondary memory moves.
15927 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15928 reg set up in the current insn.
15930 2013-05-24 Dehao Chen <dehao@google.com>
15932 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
15934 (locus_descrim_hasher::equal): Likewise.
15935 (build_gimple_cfg): New discriminator assignment algorithm.
15936 (make_edges): Likewise.
15937 (next_discriminator_for_locus): Likewise.
15938 (same_line_p): Likewise.
15939 (assign_discriminators): Likewise.
15940 (make_cond_expr_edges): Likewise.
15941 (make_gimple_switch_edges): Likewise.
15942 (make_goto_expr_edges): Likewise.
15943 (make_gimple_asm_edges): Likewise.
15945 2013-05-24 Ian Bolton <ian.bolton@arm.com>
15947 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15948 X format specifier to only display bottom 16 bits.
15949 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15950 immediate to match for operand 2, since it will be masked.
15952 2013-05-24 Richard Biener <rguenther@suse.de>
15954 PR tree-optimization/57287
15955 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15956 all SSA names that occur in abnormal PHIs.
15958 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15961 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15962 that index is not negative.
15964 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15966 PR rtl-optimization/55177
15967 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15968 (simplify_byte_swapping_operation): New.
15969 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15970 (simplify_relational_operation_1): Deal with BSWAP.
15972 2013-05-23 Richard Henderson <rth@redhat.com>
15975 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15976 (ix86_reorg): Call it.
15978 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
15981 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
15982 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
15983 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
15985 2013-05-23 Christian Bruel <christian.bruel@st.com>
15988 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
15990 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15991 Marcus Shawcroft <marcus.shawcroft@arm.com>
15993 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
15994 * config/aarch64/constraints.md (Usa): Remove.
15995 * doc/md.texi (AArch64 Usa): Remove.
15997 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15998 Marcus Shawcroft <marcus.shawcroft@arm.com>
16000 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16001 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16002 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16003 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16005 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16007 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16008 instruction (AdvSIMD).
16009 * config/aarch64/aarch64-builtins.c
16010 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16011 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16013 2013-05-23 Martin Jambor <mjambor@suse.cz>
16015 PR middle-end/57347
16016 * tree.h (contains_bitfld_component_ref_p): Declare.
16017 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16018 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16020 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16021 not access a bit-field. Assert all final offsets are byte-aligned.
16023 2013-05-23 Richard Biener <rguenther@suse.de>
16025 PR tree-optimization/57380
16026 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16027 least one invariant or re-used load.
16028 * passes.c (init_optimization_passes): Move pass_phiprop before
16031 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16033 * config/aarch64/aarch64-simd.md
16034 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16036 2013-05-23 Richard Biener <rguenther@suse.de>
16038 PR middle-end/57381
16039 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16040 OEP_CONSTANT_ADDRESS_OF retained.
16042 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16044 PR middle-end/57344
16045 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16046 don't lower unit. Handle unit not being always BITS_PER_WORD.
16048 2013-05-23 Richard Biener <rguenther@suse.de>
16050 PR rtl-optimization/57341
16051 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16052 instead of true_dependence.
16054 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16056 * bb-reorder.c (branch_threshold): Make const.
16057 (exec_threshold): Ditto.
16059 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16060 Pat Haugen <pthaugen@us.ibm.com>
16061 Peter Bergner <bergner@vnet.ibm.com>
16063 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16064 documentation for the power8 crypto builtins.
16066 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16068 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16069 macros for defining power8 builtin functions.
16070 (BU_P8V_AV_2): Likewise.
16071 (BU_P8V_AV_P): Likewise.
16072 (BU_P8V_VSX_1): Likewise.
16073 (BU_P8V_OVERLOAD_1): Likewise.
16074 (BU_P8V_OVERLOAD_2): Likewise.
16075 (BU_CRYPTO_1): Likewise.
16076 (BU_CRYPTO_2): Likewise.
16077 (BU_CRYPTO_3): Likewise.
16078 (BU_CRYPTO_OVERLOAD_1): Likewise.
16079 (BU_CRYPTO_OVERLOAD_2): Likewise.
16080 (XSCVSPDP): Fix typo, point to the correct instruction.
16081 (VCIPHER): Add power8 crypto builtins.
16082 (VCIPHERLAST): Likewise.
16083 (VNCIPHER): Likewise.
16084 (VNCIPHERLAST): Likewise.
16085 (VPMSUMB): Likewise.
16086 (VPMSUMH): Likewise.
16087 (VPMSUMW): Likewise.
16088 (VPERMXOR_V2DI): Likewise.
16089 (VPERMXOR_V4SI: Likewise.
16090 (VPERMXOR_V8HI: Likewise.
16091 (VPERMXOR_V16QI: Likewise.
16092 (VSHASIGMAW): Likewise.
16093 (VSHASIGMAD): Likewise.
16094 (VPMSUM): Likewise.
16095 (VPERMXOR): Likewise.
16096 (VSHASIGMA): Likewise.
16098 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16099 __CRYPTO__ if the crypto instructions are available.
16100 (altivec_overloaded_builtins): Add support for overloaded power8
16103 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16104 support for power8 crypto builtins.
16105 (builtin_function_type): Likewise.
16106 (altivec_init_builtins): Add support for builtins that take vector
16107 long long (V2DI) arguments.
16109 * config/rs6000/crypto.md: New file, define power8 crypto
16112 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16113 Pat Haugen <pthaugen@us.ibm.com>
16114 Peter Bergner <bergner@vnet.ibm.com>
16116 * doc/invoke.texi (Option Summary): Add power8 options.
16117 (RS/6000 and PowerPC Options): Likewise.
16119 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16120 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16121 wm, wn, wr documentation.
16123 * config/rs6000/constraints.md (wm): New constraint for VSX
16124 registers if direct move instructions are enabled.
16125 (wn): New constraint for no registers.
16126 (wq): New constraint for quad word even GPR registers.
16127 (wr): New constraint if 64-bit instructions are enabled.
16128 (wv): New constraint if power8 vector instructions are enabled.
16129 (wQ): New constraint for quad word memory locations.
16131 * config/rs6000/predicates.md (const_0_to_15_operand): New
16132 constraint for 0..15 for crypto instructions.
16133 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16134 as GPR and floating point registers.
16135 (int_reg_operand): New predicate to match only GPR registers.
16136 (base_reg_operand): New predicate to match base registers.
16137 (quad_int_reg_operand): New predicate to match even GPR registers
16138 for quad memory operations.
16139 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16140 operations in both GPR and VSX registers.
16141 (quad_memory_operand): New predicate for quad memory operations.
16142 (reg_or_indexed_operand): New predicate for direct move support.
16144 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16145 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16146 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16147 (POWERPC_MASKS): Add power8 options.
16148 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16151 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16152 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16154 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16155 (-mpower8-fusion): New power8 options.
16156 (-mpower8-fusion-sign): Likewise.
16157 (-mpower8-vector): Likewise.
16158 (-mcrypto): Likewise.
16159 (-mdirect-move): Likewise.
16160 (-mquad-memory): Likewise.
16162 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16163 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16165 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16166 (rs6000_debug_vector_unit): Add p8_vector.
16167 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16168 definitions. Also print fusion state.
16169 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16170 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16171 (rs6000_option_override_internal): Add support for power8.
16172 (rs6000_common_init_builtins): Add debugging for skipped builtins
16173 if -mdebug=builtin.
16174 (rs6000_adjust_cost): Add power8 support.
16175 (rs6000_issue_rate): Likewise.
16176 (insn_must_be_first_in_group): Likewise.
16177 (insn_must_be_last_in_group): Likewise.
16178 (force_new_group): Likewise.
16179 (rs6000_register_move_cost): Likewise.
16180 (rs6000_opt_masks): Likewise.
16182 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16183 power8 capable assembler, default to power7 options.
16184 (TARGET_DIRECT_MOVE): Likewise.
16185 (TARGET_CRYPTO): Likewise.
16186 (TARGET_P8_VECTOR): Likewise.
16187 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16188 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16189 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16190 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16191 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16192 (TARGET_XSCVDPSPN): Likewise.
16193 (TARGET_XSCVSPDPN): Likewsie.
16194 (TARGET_SYNC_HI_QI): Likewise.
16195 (TARGET_SYNC_TI): Likewise.
16196 (MASK_CRYPTO): Likewise.
16197 (MASK_DIRECT_MOVE): Likewise.
16198 (MASK_P8_FUSION): Likewise.
16199 (MASK_P8_VECTOR): Likewise.
16200 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16201 temporary used by some of the direct move instructions to get two FP
16202 temporary registers does not force creation of a stack frame.
16203 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16204 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16205 that any VSX registers are tieable, even if they are also an
16206 Altivec vector mode.
16207 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16208 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16209 (RS6000_BTM_CRYPTO): Likewise.
16210 (RS6000_BTM_COMMON): Likewise.
16212 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16213 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16214 (enum rs6000_vector): Add power8 vector support.
16216 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16220 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16221 (any_sibcall_could_use_r3): this and handle indirect calls.
16222 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16224 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16226 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16228 2013-05-22 Richard Biener <rguenther@suse.de>
16230 PR middle-end/57349
16231 * profile.c (branch_prob): Do not split blocks that are
16232 abnormally receiving from ECF_RETURNS_TWICE functions.
16234 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16236 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16237 address mode. Move pointer mode initialization to the same place.
16239 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16241 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16242 while it has any effect.
16244 2013-05-21 Easwaran Raman <eraman@google.com>
16246 PR tree-optimization/57322
16247 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16248 UID of the statement added to the BB to be 1.
16250 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16252 PR tree-optimization/57331
16253 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16254 of conversion from pointer type to integral type with integer.
16256 2013-05-21 Martin Jambor <mjambor@suse.cz>
16259 * ipa-prop.c (ipa_read_node_info): Process param_used and
16260 controlled_uses in the same order as when writing.
16262 2013-05-21 Magnus Granberg <baldrick@free.fr>
16265 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16267 2013-05-21 Richard Biener <rguenther@suse.de>
16269 PR tree-optimization/57318
16270 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16271 estimate stmts with side-effects as likely eliminated.
16273 2013-05-21 Richard Biener <rguenther@suse.de>
16275 PR tree-optimization/57330
16276 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16277 preserve the call stmts fntype.
16279 2013-05-21 Richard Biener <rguenther@suse.de>
16281 PR tree-optimization/57303
16282 * tree-ssa-sink.c (statement_sink_location): Improve killing
16283 stmt detection and properly handle self-assignments.
16285 2013-05-21 Christian Bruel <christian.bruel@st.com>
16287 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16288 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16289 registers. Set register size out of the PARALLEL loop.
16291 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16294 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16295 (*fmasf4, *fmasf4_media): New insns.
16297 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16299 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16300 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16301 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16302 (mips_idiv_insns): Update the comments to say that the returned
16303 instruction counts are in units of BASE_INSN_LENGTH.
16304 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16305 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16306 using 2 rather than 4 as the length of indirect MIPS16 and
16307 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16308 length of a NOP. Don't divide MIPS16 lengths by 2.
16309 (mips16_split_long_branches): Assume a branch is long if the
16310 length is greater than 4 rather than 8.
16311 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16312 rather than multiplying them by 2. Multiply instruction counts
16313 by BASE_INSN_LENGTH rather than 4.
16314 (*jump_mips16, tls_get_tp_mips16_<mode>)
16315 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16317 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16319 * config/mips/mips.md (extended_mips16): Remove branch case.
16320 (length): Remove duplicated extended_mips16 test.
16322 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16324 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16326 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16328 * recog.h (Recog_data): Rename to...
16329 (recog_data_d): ...this.
16330 (recog_data): Update accordingly.
16331 * recog.c (recog_data): Likewise.
16332 * reload.c (save_recog_data): Likewise.
16333 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16334 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16336 2013-05-17 Julian Brown <julian@codesourcery.com>
16338 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16339 found in a REG_EQUAL note, invalidate it.
16341 2013-05-17 Easwaran Raman <eraman@google.com>
16343 * tree-ssa-reassoc.c (find_insert_point): New function.
16344 (insert_stmt_after): Likewise.
16345 (get_def_stmt): Likewise.
16346 (ensure_ops_are_available): Likewise.
16347 (not_dominated_by): Likewise.
16348 (rewrite_expr_tree): Do not move statements beyond what is
16349 necessary. Remove call to swap_ops_for_binary_stmt...
16350 (reassociate_bb): ... and move it here.
16351 (build_and_add_sum): Assign UIDs for new statements.
16352 (linearize_expr): Likewise.
16353 (do_reassoc): Renumber gimple statement UIDs.
16355 2013-05-17 Jan Hubicka <jh@suse.cz>
16357 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16359 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16360 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16362 (output_weakrefs): Update.
16364 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16365 Martin Jambor <mjambor@suse.cz>
16367 PR middle-end/57276
16368 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16369 value that corresponds to the given aggval is found in values vector.
16371 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16373 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16374 sse, sse2, sse3, ssse3 and sse4a flags to options.
16376 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16378 * gengtype-state.c: (s_expr_writer): New class, to handle
16379 prettifying of output layout of s-expressions.
16380 (state_writer): New class, to write out gtype.state.
16381 (state_written_type_count): Move this variable into member data of
16383 (s_expr_writer::s_expr_writer): New code: constructor for new class
16384 (state_writer::state_writer(): ditto
16385 (s_expr_writer::write_new_line): New function
16386 (s_expr_writer::write_any_indent): ditto
16387 (s_expr_writer::begin_s_expr): ditto
16388 (s_expr_writer::end_s_expr): ditto
16389 (write_state_fileloc): convert to method of state_writer...
16390 (state_writer:: write_state_fileloc): ...and use methods of
16391 s_expr_writer to write indentation into the gtype.state output file
16392 to visually represent the hierarchical structure of the list
16394 (write_state_fields): ditto, renaming to...
16395 (state_writer::write_state_fields)
16396 (write_state_a_string): ditto, renaming to...
16397 (state_writer::write_state_a_string)
16398 (write_state_string_option): ditto, renaming to...
16399 (state_writer::write_state_string_option)
16400 (write_state_type_option): ditto, renaming to...
16401 (state_writer::write_state_type_option)
16402 (write_state_nested_option): ditto, renaming to...
16403 (state_writer::write_state_nested_option)
16404 (write_state_option): ditto, renaming to...
16405 (state_writer::write_state_option)
16406 (write_state_options): ditto, renaming to...
16407 (state_writer::write_state_options)
16408 (write_state_lang_bitmap): ditto, renaming to...
16409 (state_writer::write_state_lang_bitmap)
16410 (write_state_version): ditto, renaming to...
16411 (state_writer::write_state_version)
16412 (write_state_scalar_type): ditto, renaming to...
16413 (state_writer::write_state_scalar_type)
16414 (write_state_string_type): ditto, renaming to...
16415 (state_writer::write_state_string_type)
16416 (write_state_undefined_type): ditto, renaming to...
16417 (state_writer::write_state_undefined_type)
16418 (write_state_struct_union_type): ditto, renaming to...
16419 (state_writer::write_state_struct_union_type)
16420 (write_state_struct_type): ditto, renaming to...
16421 (state_writer::write_state_struct_type)
16422 (write_state_user_struct_type): ditto, renaming to...
16423 (state_writer::write_state_user_struct_type)
16424 (write_state_lang_struct_type): ditto, renaming to...
16425 (state_writer::write_state_lang_struct_type)
16426 (write_state_param_struct_type): ditto, renaming to...
16427 (state_writer::write_state_param_struct_type)
16428 (write_state_pointer_type): ditto, renaming to...
16429 (state_writer::write_state_pointer_type)
16430 (write_state_array_type): ditto, renaming to...
16431 (state_writer::write_state_array_type)
16432 (write_state_gc_used): ditto, renaming to...
16433 (state_writer::write_state_gc_used)
16434 (write_state_common_type_content): ditto, renaming to...
16435 (state_writer::write_state_common_type_content)
16436 (write_state_type): ditto, renaming to...
16437 (state_writer::write_state_type)
16438 (write_state_pair_list): ditto, renaming to...
16439 (state_writer::write_state_pair_list)
16440 (write_state_pair): ditto, renaming to...
16441 (state_writer::write_state_pair)
16442 (write_state_typedefs): ditto, renaming to...
16443 (state_writer::write_state_typedefs)
16444 (write_state_structures): ditto, renaming to...
16445 (state_writer::write_state_structures)
16446 (write_state_param_structs): ditto, renaming to...
16447 (state_writer::write_state_param_structs)
16448 (write_state_variables): ditto, renaming to...
16449 (state_writer::write_state_variables)
16450 (write_state_srcdir): ditto, renaming to...
16451 (state_writer::write_state_srcdir)
16452 (write_state_files_list): ditto, renaming to...
16453 (state_writer::write_state_files_list)
16454 (write_state_languages): ditto, renaming to...
16455 (state_writer::write_state_languages)
16456 (write_state): create a state_writer instance and use it when
16457 writing out the state file
16459 2013-05-17 Mike Stump <mikestump@comcast.net>
16461 PR rtl-optimization/57304
16462 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16463 accessing DF_REF_REAL_LOC.
16465 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16467 PR rtl-optimization/57281
16468 PR rtl-optimization/57300
16469 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16470 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16471 what the other splitter did if the registers are dead.
16473 2013-05-17 Richard Biener <rguenther@suse.de>
16475 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16478 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16480 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16483 2013-05-17 Marek Polacek <polacek@redhat.com>
16485 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16486 length when doing non-zero store of storing '\0' to '\0'.
16488 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16490 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16491 vect_external_def oprnd1 with loop_vinfo, try to emit
16492 optional cast, negation and and stmts on the loop preheader
16493 edge instead of into the pattern def seq.
16495 PR tree-optimization/57051
16496 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16497 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16499 2013-05-16 Nick Clifton <nickc@redhat.com>
16501 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16502 (rl78_is_naked_func): New function.
16503 (rl78_expand_prologue): Skip prologue generation for naked functions.
16504 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16505 * doc/extend.texi (naked): Add RL78 to the list of processors
16506 that supports this attribute.
16508 2013-05-16 Jeff Law <law@redhat.com>
16510 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16512 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16514 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16515 cache parameters using detect_caches_amd also for CYRIX,
16516 NSC and TM2 signatures.
16518 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16519 Dzianis Kahanovich <mahatma@eu.by>
16523 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16524 VIA/Centaur processors and determine their cache parameters
16525 using detect_caches_amd.
16527 2013-05-16 Teresa Johnson <tejohnson@google.com>
16529 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16530 (rtl_verify_edges): New function.
16531 (rtl_verify_bb_insns): Ditto.
16532 (rtl_verify_bb_pointers): Ditto.
16533 (rtl_verify_bb_insn_chain): Ditto.
16534 (rtl_verify_fallthru): Ditto.
16535 (rtl_verify_bb_layout): Ditto.
16536 (rtl_verify_flow_info_1): Outline checks into new functions.
16537 (rtl_verify_flow_info): Ditto.
16539 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16541 * cfghooks.c (copy_bbs): Add update_dominance argument.
16542 * cfghooks.h (copy_bbs): Update prototype.
16543 * tree-cfg.c (gimple_duplicate_sese_region):
16544 Add update_dominance argument.
16545 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16546 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16547 gimple_duplicate_sese_region call.
16548 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16549 Update copy_bbs call.
16550 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16551 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16553 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16555 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16556 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16557 vect_recog_rotate_pattern.
16558 (vect_recog_rotate_pattern): New function.
16560 2013-05-16 Jason Merrill <jason@redhat.com>
16562 * Makefile.in (LLINKER): New variable.
16563 (mostlyclean): Remove link mutex.
16564 * configure.ac: Handle --enable-link-mutex.
16565 * lock-and-run.sh: New script.
16567 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16570 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16573 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16575 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16577 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16579 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16580 * config/arm/arm.c (next_consecutive_mem): New function.
16581 (gen_movmem_ldrd_strd): Likewise.
16582 * config/arm/arm.md (movmemqi): Update condition and code.
16583 (unaligned_loaddi, unaligned_storedi): New patterns.
16585 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16587 * config.gcc: Obsolete *-*-solaris2.9*.
16588 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16590 2013-05-16 Richard Biener <rguenther@suse.de>
16592 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16593 earlier, after GRAPHITE transforms and IV canonicalization.
16595 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16597 * omp-low.c (extract_omp_for_data): For collapsed loops,
16598 if at least one of the loops is known at compile time to
16599 iterate zero times, set count to 0.
16600 (expand_omp_regimplify_p): New function.
16601 (expand_omp_for_generic): For collapsed loops, if at least
16602 one of the loops isn't known to iterate at least once,
16603 add runtime check with setting count to 0.
16604 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16605 For unsigned types if it isn't known at compile time that
16606 the loop will iterate at least once, add runtime check to bypass
16607 the whole loop if initial condition isn't true.
16609 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16611 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16613 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16615 PR middle-end/57286
16616 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16617 transformations to avoid an infinite loop.
16619 2013-05-16 Marek Polacek <polacek@redhat.com>
16621 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16623 2013-05-15 Leif Ekblad <leif@rdos.net>
16625 * config/i386/i386.c (ix86_decompose_address): Use
16626 DEFAULT_TLS_SEG_REG to access TLS segment register.
16627 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16628 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16629 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16631 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16634 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16635 sibling calls to functions that would normally be lazily bound,
16636 unless $gp is call-clobbered.
16638 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16640 * config/i386/i386.c (ix86_option_override_internal): Update
16641 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16642 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16643 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16644 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16646 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16648 2013-05-15 Andreas Schwab <schwab@suse.de>
16650 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16651 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16654 2013-05-15 Teresa Johnson <tejohnson@google.com>
16656 * loop-unroll.c (report_unroll_peel): Check decision before
16657 emitting unroll/peel message.
16659 2013-05-15 Teresa Johnson <tejohnson@google.com>
16661 * function.h (has_bb_partition): New rtl_data flag.
16662 (bb_reorder_complete): Ditto.
16663 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16664 instead of flag_reorder_blocks_and_partition.
16665 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16666 with some enhancements.
16667 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16668 * bb-reorder.c (connect_traces): Check for has_bb_partition
16669 instead of flag_reorder_blocks_and_partition.
16670 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16671 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16672 verify_hot_cold_block_grouping.
16673 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16675 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16678 * config/arm/predicates.md (call_insn_operand): New predicate.
16679 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16680 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16681 if insn is not a tail call.
16682 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16684 * config/arm/arm.h (enum reg_class): New caller save register class.
16685 (REG_CLASS_NAMES): Likewise.
16686 (REG_CLASS_CONTENTS): Likewise.
16687 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16690 2013-05-15 Richard Biener <rguenther@suse.de>
16692 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16693 of MSG_OPTIMIZED_LOCATIONS.
16694 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16695 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16697 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16698 of MSG_OPTIMIZED_LOCATIONS.
16699 (execute_vect_slp): Likewise.
16700 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16701 (vect_create_cond_for_alias_checks): Likewise.
16702 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16703 (vect_recog_widen_mult_pattern): Likewise.
16704 (vect_recog_widen_sum_pattern): Likewise.
16705 (vect_recog_over_widening_pattern): Likewise.
16706 (vect_recog_widen_shift_pattern): Likewise.
16707 (vect_recog_vector_vector_shift_pattern): Likewise.
16708 (vect_recog_divmod_pattern): Likewise.
16709 (vect_recog_mixed_size_cond_pattern): Likewise.
16710 (vect_recog_bool_pattern): Likewise.
16711 (vect_pattern_recog_1): Likewise.
16713 2013-05-15 Martin Jambor <mjambor@suse.cz>
16715 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16716 non-functions to builtin_unreachable.
16717 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16718 correct when new direct edges were discovered.
16720 2013-05-15 Martin Jambor <mjambor@suse.cz>
16722 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16723 header, print symbol order instead of node uid, print more information
16724 about indirect edge targets.
16725 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16727 (ipa_make_edge_direct_to_target): Likewise.
16728 (remove_described_reference): Likewise.
16729 (propagate_controlled_uses): Likewise.
16730 (ipa_print_node_params): Also print symbol order.
16731 (ipcp_transform_function): Print symbol order instead of node uids.
16732 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16733 (cgraph_get_create_real_symbol_node): Likewise.
16734 * ipa-cp.c (print_lattice): Likewise.
16735 (print_all_lattices): Likewise.
16736 (determine_versionability): Likewise.
16737 (initialize_node_lattices): Likewise.
16738 (estimate_local_effects): Likewise.
16739 (update_profiling_info): Likewise.
16740 (create_specialized_node): Likewise.
16741 (perhaps_add_new_callers): Likewise.
16742 (decide_about_value): Likewise.
16743 (decide_whether_version_node): Likewise.
16744 (identify_dead_nodes): Likewise.
16745 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16746 (dump_inline_summary): Likewise.
16747 (estimate_node_size_and_time): Likewise.
16748 (inline_analyze_function): Likewise.
16749 * ipa-inline.c (report_inline_failed_reason): Likewise.
16750 (want_early_inline_function_p): Likewise.
16751 (edge_badness): Likewise.
16752 (update_edge_key): Likewise.
16753 (inline_small_functions): Likewise. Add dumping of order to two other
16755 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16756 instead of node uids.
16757 (propagate_pure_const): Likewise.
16758 (propagate_pure_const): Likewise.
16759 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16760 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16761 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16763 * tree-pretty-print.c (dump_function_header): Likewise.
16764 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16765 Print symbol order instead of node uids.
16767 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16769 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16770 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16772 2013-05-15 Richard Biener <rguenther@suse.de>
16774 PR tree-optimization/57275
16775 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16776 return value for fail to do runtime alias checks for gather loads.
16778 2013-05-15 Jan Hubicka <jh@suse.cz>
16782 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16783 weakrefs are not external.
16784 (lto_symtab_merge_decls): Fix thinko when dealing with
16785 non-lto_symtab decls.
16786 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16787 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16788 * varpool.c (dump_varpool_node): Dump more flags.
16790 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16792 * config/i386/i386.c (processor_alias_table): Add instruction
16793 FSGSBASE for AMD bdver3 architecture.
16795 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16797 * tree.c (warn_deprecated_use): Print file:line using locus color.
16798 * diagnostic.c (diagnostic_report_current_module): Print file:line
16799 and file:line:column using locus color.
16801 2013-05-14 Mike Stump <mikestump@comcast.net>
16803 * gdbinit.in: Add __null.
16805 2013-05-14 Mike Stump <mikestump@comcast.net>
16807 * recog.h: Rename struct recog_data to Recog_data.
16808 * recog.c: Likewise.
16809 * reload.c (can_reload_into): Likewise.
16810 * config/picochip/picochip.c: Likewise.
16812 2013-05-14 Mike Stump <mikestump@comcast.net>
16814 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16816 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16818 * resource.h (struct resources): Remove unch_memory member.
16819 (CLEAR_RESOURCE): Don't clear unch_memory.
16820 * resource.c (mark_referenced_resources): Don't set it.
16821 (mark_set_resources): Likewise.
16822 (mark_target_live_regs): Don't clear it.
16823 (init_resource_info): Likewise.
16824 * reorg.c (resource_conflicts_p): Don't compare it.
16825 (redundant_insn): Don't set it.
16827 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16829 * emit-rtl.c (next_label): Remove unused function.
16830 (skip_consecutive_labels, link_cc0_insns): Move to ...
16831 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16832 only place where these functions are used, and make them static.
16834 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16836 * fold-const.c (fold_negate_expr): Handle vectors.
16837 (fold_truth_not_expr): Make it static.
16838 (fold_invert_truthvalue): New static function.
16839 (invert_truthvalue_loc): Handle vectors. Do not call
16840 fold_truth_not_expr directly.
16841 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16842 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16843 (fold_comparison): Handle vector constants.
16844 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16845 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16846 * tree.h (fold_truth_not_expr): Remove declaration.
16848 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16850 * config/aarch64/aarch64-simd.md
16851 (aarch64_vcond_internal<mode>): Rename to...
16852 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16853 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16854 float modes. Clarify all iterator modes.
16855 (vcond<mode><mode>): Use new name for vcond expanders.
16856 (vcond<v_cmp_result><mode>): Likewise.
16857 (vcondu<mode><mode>: Likewise.
16858 * config/aarch64/iterators.md (VDQF_COND): New.
16860 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16863 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16864 variable for the shift amount. Check that we shift by non-negative
16867 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16870 * config/arm/arm.h (EPILOGUE_USES): Only return true
16871 for LR_REGNUM after epilogue_completed.
16873 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
16875 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
16876 is error_mark_node.
16878 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16881 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16882 and Solaris 11+/x86 with gld.
16883 * configure: Regenerate.
16885 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16887 * expmed.c (expand_shift_1): Canonicalize rotates by
16888 constant bitsize / 2 to bitsize - 1.
16889 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
16890 case ROTATERT>: Likewise.
16893 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16895 * config/i386/i386.md (rotateinv): New code attr.
16896 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16897 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16898 roll $31, %eax, etc.
16900 2013-05-14 Richard Biener <rguenther@suse.de>
16902 PR middle-end/57235
16903 * tree-eh.c (sink_clobbers): Give up for successors with
16904 multiple predecessors and no virtual uses.
16906 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16908 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16909 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16911 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16913 PR middle-end/57251
16914 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16915 the case when both op0 and op1 have VOIDmode.
16917 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16919 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16920 in multiply-accumulate mode.
16922 2013-05-13 Guozhi Wei <carrot@google.com>
16924 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16926 2013-05-13 Kai Tietz <ktietz@redhat.com>
16929 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
16930 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
16931 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
16932 * config/i386/i386.c (ix86_option_override_internal): Likewise.
16933 (ix86_expand_prologue): Likewise.
16934 (ix86_expand_split_stack_prologue): Likewise.
16935 (legitimate_pic_address_disp_p): Likewise.
16936 (legitimize_pic_address): Likewise.
16937 (legitimize_tls_address): Likewise.
16938 (legitimize_pe_coff_symbol): Likewise.
16939 (output_pic_addr_const): Likewise.
16940 (construct_plt_address): Likewise.
16941 (ix86_expand_call): Likewise.
16942 (x86_output_mi_thunk): Likewise.
16943 (x86_function_profiler): Likewise.
16945 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16947 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16948 similar switch cases.
16949 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16950 (aarch64_simd_mov_to_<mode>low): Delete.
16951 (aarch64_simd_mov_to_<mode>high): Delete.
16952 (move_lo_quad_<mode>): Add w<-r alternative.
16953 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16954 (aarch64_simd_mov_from_*): Update type attribute.
16955 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16958 2013-05-13 Jan Hubicka <jh@suse.cz>
16960 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16961 * config/i386/i386.c (ix86_compute_frame_layout,
16962 ix86_expand_epilogue, emit_i387_cw_initialization,
16963 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16964 ix86_local_alignment): Fix use of size/speed predicates.
16966 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16968 PR tree-optimization/45216
16969 PR tree-optimization/57157
16970 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16971 the (-Y) & (B - 1) variant if OP is |.
16972 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16973 return shifted. Use (-op1) & (prec - 1) as other_amount
16974 instead of prec - op1.
16976 2013-05-13 Martin Jambor <mjambor@suse.cz>
16978 PR middle-end/42371
16979 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
16980 (ipa_constant_data): New type.
16981 (ipa_jump_func): Use ipa_constant_data to hold information about
16982 constant jump functions.
16983 (ipa_get_jf_constant): Adjust to jump function type changes.
16984 (ipa_get_jf_constant_rdesc): New function.
16985 (ipa_param_descriptor): New field controlled_uses.
16986 (ipa_get_controlled_uses): New function.
16987 (ipa_set_controlled_uses): Likewise.
16988 * ipa-ref.h (ipa_find_reference): Declare.
16989 * ipa-prop.c (ipa_cst_ref_desc): New type.
16990 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
16992 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
16993 New parameter cs. Adjust all callers.
16994 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
16995 (remove_described_reference): New function.
16996 (jfunc_rdesc_usable): Likewise.
16997 (try_make_edge_direct_simple_call): Decrement controlled use count,
16998 attempt to remove reference if it hits zero.
16999 (combine_controlled_uses_counters): New function.
17000 (propagate_controlled_uses): Likewise.
17001 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17002 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17003 (ipa_print_node_params): Print described use counter.
17004 (ipa_write_jump_function): Adjust to jump function type changes.
17005 (ipa_read_jump_function): New parameter CS, pass it to
17006 ipa_set_jf_constant. Adjust caller.
17007 (ipa_write_node_info): Stream controlled use count
17008 (ipa_read_node_info): Likewise.
17009 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17011 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17012 count. Remove cloning-added reference if it reaches zero.
17013 * ipa-ref.c (ipa_find_reference): New function.
17015 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17017 * config/i386/i386.c (processor_target_table): Modified default
17018 alignment values for AMD BD and BT architectures.
17020 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17022 * tree-vect-generic.c (uniform_vector_p): Move ...
17023 * tree.c (uniform_vector_p): ... here.
17024 * tree.h (uniform_vector_p): Declare it.
17025 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17028 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17030 PR tree-optimization/57230
17031 * tree-ssa-strlen.c (handle_char_store): Record length for
17032 array store from STRING_CST.
17034 PR tree-optimization/57230
17035 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17038 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17040 * config/epiphany/epiphany.c (epiphany_init): Check size of
17041 NUM_MODES_FOR_MODE_SWITCHING.
17042 (epiphany_expand_prologue):
17043 Remove CONFIG_REGNUM initial value handling code.
17044 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17045 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17046 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17047 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17048 Don't return 1 for FP_MODE_NONE.
17049 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17050 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17051 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17052 * config/epiphany/epiphany.md (save_config): New pattern.
17054 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17056 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17058 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17060 * config/i386/i386.md (memory): Handle sseishft1.
17061 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17062 (*vec_extractv2di_1): Ditto.
17064 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17066 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17069 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17071 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17072 Add mthumb/march=armv7-a multilib.
17073 Add mthumb/march=armv7-r multilib.
17074 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17076 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17078 * config/v850/t-rtems: Add more multilibs.
17080 2013-05-10 Richard Biener <rguenther@suse.de>
17082 PR tree-optimization/57214
17083 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17084 not propagate from SSA names that occur in abnormal PHI nodes.
17086 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17088 * stor-layout.c (element_precision): New function.
17089 * machmode.h (element_precision): Declare it.
17090 * tree.c (build_minus_one_cst): New function.
17091 (element_precision): Likewise.
17092 * tree.h (build_minus_one_cst): Declare new function.
17093 (element_precision): Likewise.
17094 * fold-const.c (operand_equal_p): Use element_precision.
17095 (fold_binary_loc): Handle vector types.
17096 * convert.c (convert_to_integer): Use element_precision.
17097 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17100 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17102 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17103 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17104 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17105 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17106 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17107 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17108 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17109 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17110 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17111 (Uuw8): New constraints.
17112 (Usb4): Move into alphabetical order.
17113 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17114 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17115 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17116 previously unnamed patterns.
17117 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17118 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17119 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17120 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17121 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17122 of specific lengths.
17124 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17126 * config/i386/i386.md (rotateinv): New code attr.
17127 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17128 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17129 roll $31, %eax, etc.
17131 PR tree-optimization/45216
17132 PR tree-optimization/57157
17133 * tree-ssa-forwprop.c (simplify_rotate): New function.
17134 (ssa_forward_propagate_and_combine): Call it.
17136 2013-05-10 Richard Biener <rguenther@suse.de>
17138 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17139 disable peeling when we version for aliasing.
17140 (vector_alignment_reachable_p): Honor explicit user alignment.
17141 (vect_supportable_dr_alignment): Likewise.
17142 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17143 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17144 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17145 then peeling to arrange for the cost-model check to come first.
17147 2013-05-10 Alan Modra <amodra@gmail.com>
17149 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17150 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17151 * configure: Regenerate.
17153 2013-05-10 Alan Modra <amodra@gmail.com>
17156 * varasm.c (default_elf_select_section): Move !DECL_P check..
17157 (get_named_section): ..to here before calling get_section_name.
17159 (default_section_type_flags): Add DECL_P check.
17160 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17161 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17163 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17165 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17166 When using gen_stack_adjust_str with a register offset, add a
17167 REG_FRAME_RELATED_EXPR note.
17169 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17171 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17172 (*vec_extractv4si_zext_mem): Ditto.
17173 (*vec_extractv2di): Add 0->x and x->x alternatives.
17174 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17175 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17177 2013-05-09 Jason Merrill <jason@redhat.com>
17179 N3639 C++1y VLA support
17180 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17183 * tree.c (build_constructor_va): New.
17184 * tree.h: Declare it.
17186 2013-05-09 Martin Jambor <mjambor@suse.cz>
17189 * gimple-fold.c (canonicalize_constructor_val): Call
17190 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17192 2013-05-09 Jan Hubicka <jh@suse.cz>
17193 Richard Biener <rguenther@suse.de>
17196 * symtab.c (symtab_make_decl_local): Do not add private names.
17198 2013-05-09 Jan Hubicka <jh@suse.cz>
17201 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17202 (unlink_from_assembler_name_hash): Likewise.
17203 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17204 symtab_unregister_node, symtab_initialize_asm_name_hash,
17205 change_decl_assembler_name): Update.
17207 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17209 * config/aarch64/aarch64.md: New movtf split.
17210 (*movtf_aarch64): Update.
17211 (aarch64_movdi_tilow): Handle TF modes and rename to
17212 aarch64_movdi_<mode>low.
17213 (aarch64_movdi_tihigh): Handle TF modes and rename to
17214 aarch64_movdi_<mode>high
17215 (aarch64_movtihigh_di): Handle TF modes and rename to
17216 aarch64_mov<mode>high_di
17217 (aarch64_movtilow_di): Handle TF modes and rename to
17218 aarch64_mov<mode>low_di
17219 (aarch64_movtilow_tilow): Remove spurious whitespace.
17220 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17222 (aarch64_print_operand): Update.
17224 2013-05-09 Alan Modra <amodra@gmail.com>
17226 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17228 * configure: Regenerate.
17230 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17232 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17233 splitter preparation statements.
17234 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17235 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17238 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17240 * gimple-ssa-strength-reduction.c (count_candidates): Change
17241 return value to int.
17242 (analyze_candidates_and_replace): Change type of length to int.
17244 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17246 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17247 (*vec_extract<mode>): Use VI12_128 mode iterator.
17248 (*vec_extract<mode>_mem): Ditto.
17249 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17252 2013-05-08 Diego Novillo <dnovillo@google.com>
17257 2012-08-17 Diego Novillo <dnovillo@google.com>
17260 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17261 * config.in: Regenerate.
17262 * configure: Regenerate.
17263 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17265 2013-05-08 Jan Hubicka <jh@suse.cz>
17268 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17269 * cgraph.h (symtab_node_base): Add unique_name.
17270 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17271 input_overwrite_node, input_varpool_node): Stream unique_name.
17272 * cgraphclones.c (cgraph_create_virtual_clone,
17273 cgraph_function_versioning): Set unique_name.
17274 * ipa.c (function_and_variable_visibility): Set unique_name.
17276 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17278 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17279 (alloc_cand_and_find_basis): Restrict conditional candidate
17280 processing to CAND_MULTs.
17282 2013-05-08 Jan Hubicka <jh@suse.cz>
17285 lto-symtab.c (lto_symtab_symbol_p): New function.
17286 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17287 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17288 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17289 Skip static symbols.
17291 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17293 PR tree-optimization/57200
17294 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17295 Only call inform if the preceding warning_at returns true.
17297 2013-05-07 Han Shen <shenhan@google.com>
17299 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17300 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17301 * common.opt (fstack-protector-strong): New option.
17302 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17303 * doc/invoke.texi (Optimization Options): Document
17304 "-fstack-protector-strong".
17305 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17307 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17309 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17311 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17313 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17314 it is not lhs of the insn.
17315 (setup_reg_equiv): Remove insn having equiv memory which it is not
17317 * lra-constraints.c (process_address): Try to improve generation
17318 code for address base + disp.
17319 (lra_constraints): Make correct the code for checking insn setting
17320 up backward equivalence. Remove insn only if it is in the init
17322 * lra-eliminations.c (update_reg_eliminate): Change return value.
17323 (lra_eliminate): Use the result.
17325 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17327 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17328 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17329 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17330 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17331 register target operands.
17332 (*vec_extractv8hi_sse2): New pattern.
17333 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17334 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17335 (*vec_extract<mode>_mem): New insn and split pattern.
17337 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17339 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17340 (TARGET_ASAN_SHADOW_OFFSET): Define.
17341 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17342 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17344 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17346 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17347 (incr_vec_index): Return -1 if increment not found.
17348 (create_add_on_incoming_edge): Assert if increment not found.
17349 (record_increment): Limit number of increments recorded.
17350 (all_phi_incrs_profitable): Return false if an increment not found.
17351 (replace_profitable_candidates): Don't process increments that were
17353 (analyze_candidates_and_replace): Limit size of incr_vec.
17355 2013-05-07 Richard Biener <rguenther@suse.de>
17357 * calls.c (special_function_p): setjmp-like functions are leaf.
17358 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17359 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17361 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17363 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17364 (aarch64_simd_mov<mode>): New expander.
17365 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17366 (aarch64_simd_mov_to_<mode>high): Likewise.
17367 (aarch64_simd_mov_from_<mode>low): Likewise.
17368 (aarch64_simd_mov_from_<mode>high): Likewise.
17369 (aarch64_dup_lane<mode>): Update.
17370 (aarch64_dup_lanedi): New instruction pattern.
17371 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17372 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17374 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17376 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17377 (replace_mult_candidate): Remove unnecessary argument; remove
17378 unnecessary parameter from call to introduce_cast_before_cand.
17379 (replace_unconditional_candidate): Remove unnecessary parameter
17380 from call to replace_mult_candidate.
17381 (replace_conditional_candidate): Likewise.
17382 (insert_initializers): Use make_temp_ssa_name.
17383 (introduce_cast_before_cand): Remove unnecessary argument; use
17384 make_temp_ssa_name.
17385 (replace_one_candidate): Remove unnecessary argument; remove
17386 unnecessary parameter from calls to introduce_cast_before_cand.
17387 (replace_profitable_candidates): Remove unnecessary parameters
17388 from calls to replace_one_candidate.
17390 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17392 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17393 phi def as possibly hiding a basis for a CAND_ADD whose operands
17394 have been commuted in the analysis.
17395 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17397 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17399 * config/aarch64/aarch64.md
17400 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17401 shift value between 0-4.
17403 2013-05-07 Richard Biener <rguenther@suse.de>
17405 * double-int.h (rshift): New overload.
17406 * double-int.c (rshift): New function.
17407 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17408 (create_reference_ops_from_ref): Remove.
17409 (vn_reference_insert): Use shared ops for constructing the
17410 reference and copy it.
17412 2013-05-07 Richard Biener <rguenther@suse.de>
17414 PR middle-end/57190
17415 * tree-eh.c (sink_clobbers): Properly propagate
17416 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17418 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17420 PR tree-optimization/57149
17421 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17422 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17423 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17424 uninit_undefined_value_p instead of ssa_undefined_value_p.
17427 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17428 for modifier == EXPAND_INITIALIZER.
17430 2013-05-07 Anton Blanchard <anton@samba.org>
17432 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17433 for powerpc64 little endian.
17434 * configure: Regenerate.
17436 2013-05-06 Graham Stott <grahams@btinternet.com>
17438 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17439 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17440 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17441 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17443 2013-05-06 Graham Stott <grahams@btinternet.com>
17445 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17446 codes which allow non-lvalues.
17448 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17450 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17451 components are all 1s.
17452 (integer_minus_onep): New function.
17453 * tree.h (integer_minus_onep): Declare it.
17454 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17455 integer_minus_onep instead of integer_all_onesp.
17457 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17460 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17461 variations of these patterns.
17463 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17465 * config/i386/i386.md (isa): Add x64_sse4 member.
17466 (enabled): Handle x64_sse4.
17467 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17468 instruction for 64bit SSE4_1 targets. Update insn attributes.
17469 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17470 instruction for SSE4_1 targets. Update insn attributes.
17471 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17472 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17473 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17475 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17476 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17477 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17479 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17482 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17484 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17486 * final.c (do_assembler_dialects): Don't handle curly braces and
17487 vertical bar escaped by % as dialect delimiters.
17488 (output_asm_insn): Print curly braces and vertical bar if escaped
17489 by % and ASSEMBLER_DIALECT defined.
17490 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17491 * doc/tm.texi: Regenerated.
17493 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17495 * config/mips/mips.c: Include tree-pass.h.
17496 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17497 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17498 (pass_mips_machine_reorg2): New machine specific pass.
17499 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17500 (mips_option_override): Register the new pass.
17501 * rtl.h (cleanup_barriers): Remove prototype.
17502 (dbr_schedule): Likewise.
17503 * jump.c (cleanup_barriers): Make static.
17504 * reorg.c (dbr_schedule): Likewise.
17506 2013-05-06 Richard Biener <rguenther@suse.de>
17508 PR tree-optimization/57185
17509 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17510 SSA names properly.
17512 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17515 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17516 between "first_arg" and "insn", not "last" and "insn".
17518 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17520 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17521 (find_candidates_in_block): Re-enable slsr_process_phi.
17522 (create_phi_basis): Fix double counting of candidate adjustment.
17524 2013-05-06 Richard Biener <rguenther@suse.de>
17526 PR middle-end/57147
17527 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17528 the edge is also fallthru, preserve it and just clear the
17530 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17531 also complex, preserve that and just clear the fallthru flag.
17532 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17533 update virtual operands.
17535 2013-05-06 Alan Modra <amodra@gmail.com>
17537 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17538 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17539 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17540 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17541 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17542 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17544 2013-05-06 Alan Modra <amodra@gmail.com>
17546 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17547 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17548 (DEFAULT_ASM_ENDIAN): Define.
17549 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17550 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17551 Update -K PIC clause from sysv4.h.
17552 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17553 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17555 2013-05-06 Alan Modra <amodra@gmail.com>
17557 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17558 twice for little-endian.
17559 (ashrdi3_no_power, ashrdi3): Support little-endian.
17561 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17564 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17565 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17567 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17568 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17569 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17572 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17573 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17575 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17576 * config/pa/pa.opt: Make mbig-switch a no-op.
17577 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17578 (CASE_VECTOR_MODE): Always return SImode.
17579 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17580 for the !TARGET_BIG_SWITCH case.
17581 * config/pa/pa-linux.h: Likewise.
17582 * config/pa/pa-openbsd.h: Likewise.
17583 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17584 * config/pa/pa.md (short_jump): Remove define_insn.
17585 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17586 (casesi0): Remove define_insn.
17587 (type): Remove btable_branch.
17588 (pa_combine_type): Likewise.
17589 (in_nullified_branch_delay): Likewise.
17590 (in_call_delay): Likewise.
17591 (define_delay): Likewise.
17592 (define_insn_reservation "Z3"): Likewise.
17593 (define_insn_reservation "Z4"): Likewise.
17594 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17595 (pa_adjust_insn_length): Remove adjustment for btable branches.
17596 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17599 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17601 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17602 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17603 Add m->r,x alternatives.
17604 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17605 splitters using SWI48x mode iterator.
17606 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17607 TARGET_64BIT. Add m->x alternative.
17608 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17609 Add o->x alternative. Enable for TARGET_SSE.
17610 (sse_storeq): Remove expander.
17611 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17612 with memory input operand.
17613 (*vec_extractv2di_1 splitter): New.
17614 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17615 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17617 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17619 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17620 (INT_LOWPART): Delete.
17621 (extract_MB): Adjust.
17622 (extract_ME): Adjust.
17623 (print_operand): Adjust.
17625 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17627 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17628 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17629 (reg_or_logical_cint_operand, easy_fp_constant,
17630 logical_const_operand): Delete "CONST_DOUBLE" case.
17631 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17632 "HOST_BITS_PER_WIDE_INT == 64" test.
17633 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17634 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17635 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17636 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17638 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17639 CONST_DOUBLE DImode/VOIDmode case.
17640 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17641 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17642 CONST_DOUBLE VOIDmode case.
17643 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17644 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17645 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17646 Delete CONST_DOUBLE case.
17647 (splitters for mov FMOVE64 const_double): Delete
17648 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17649 "HOST_BITS_PER_WIDE_INT >= 64" test.
17650 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17652 (mov DI const_double): Delete.
17654 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17656 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17657 on op shows all bits zero in mode of a lowpart subreg, return zero.
17659 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17662 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17663 to save TFmode registers and DImode to save TImode registers for
17664 caller save operations.
17665 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17666 mark being partially clobbered since they only use the first
17669 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17670 and TDmode only use the upper 64-bits of each VSX register.
17672 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17674 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17675 (find_candidates_in_block): Disable slsr_process_phi.
17677 2013-05-03 Guozhi Wei <carrot@google.com>
17679 * coverage.c (coverage_obj_init): Move the construction of gcov
17681 (build_init_ctor): ... here.
17683 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17685 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17686 (slsr_cand_d): Redefine def_phi.
17687 (stride_status, phi_adjust_status, count_phis_status): New enums.
17688 (find_phi_def): New.
17689 (find_basis_for_base_expr): New.
17690 (find_basis_for_candidate): Handle hidden bases.
17691 (alloc_cand_and_find_basis): Handle phi candidates.
17692 (slsr_process_phi): New.
17693 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17694 (create_mul_imm_cand): Likewise.
17695 (create_add_ssa_cand): Exclude phi base candidates.
17696 (create_add_imm_cand): Likewise.
17697 (slsr_process_cast): Likewise.
17698 (slsr_process_copy): Likewise.
17699 (find_candidates_in_block): Handle phi candidates.
17700 (dump_candidate): Likewise.
17701 (unconditional_cands): Delete.
17702 (unconditional_cands_with_known_stride_p): Delete.
17703 (phi_dependent_cand_p): New.
17704 (cand_increment): Handle phi-dependent candidates.
17705 (replace_dependent): Delete.
17706 (replace_mult_candidate): New.
17707 (replace_unconditional_candidate): New.
17708 (incr_vec_index): Move to avoid forward reference.
17709 (create_add_on_incoming_edge): New.
17710 (create_phi_basis): New.
17711 (replace_dependents): Delete.
17712 (replace_conditional_candidate): New.
17713 (phi_add_costs): New.
17714 (replace_uncond_cands_and_profitable_phis): New.
17715 (record_increment): Handle phi adjustments.
17716 (record_phi_increments): New.
17717 (record_increments): Handle phi adjustments.
17718 (phi_incr_cost): New.
17719 (lowest_cost_path): Handle phis.
17720 (total_savings): Likewise.
17721 (analyze_increments): Likewise.
17722 (ncd_with_phi): New.
17723 (ncd_of_cand_and_phis): New.
17724 (nearest_common_dominator_for_cands): Handle phi increments.
17725 (all_phi_incrs_profitable): New.
17726 (replace_profitable_candidates): Handle phi-dependent candidates.
17727 (analyze_candidates_and_replace): Likewise.
17729 2013-05-03 Teresa Johnson <tejohnson@google.com>
17732 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17733 do not exceed REG_BR_PROB_BASE.
17735 2013-05-03 Jeff Law <law@redhat.com>
17737 PR tree-optimization/57144
17738 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17739 operand of the condition will bit into the new type when eliminating
17740 a cast feeding a condition.
17742 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17744 PR rtl-optimization/57130
17745 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17746 of COMPARE as in_code to the recursive call if needed.
17748 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17750 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17751 (enabled): Handle new members.
17752 * config/i386/sse.md (*vec_concatv2si): Merge from
17753 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17754 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17756 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17758 PR tree-optimization/57027
17759 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17760 for fnms opportunity, check we got the prerequisite kind
17761 of tree / gimple before using accessor functions.
17763 2013-05-03 Richard Biener <rguenther@suse.de>
17765 * double-int.h (lshift): New overload without precision
17766 and arith argument.
17767 (operator *=, operator +=, operator -=): Move ...
17768 * double-int.c (operator *=, operator +=, operator -=): ... here
17769 and implement more efficiently.
17770 (mul_double_with_sign): Remove.
17771 (lshift_double): Adjust to take unsinged shift argument, push
17772 dispatching code to callers.
17773 (mul_double_wide_with_sign): Add early out for callers that
17774 are not interested in high parts or overflow.
17775 (lshift): New function.
17776 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17777 dispatch code here.
17778 (lrotate, rrotate): Use logical shifts.
17779 * expr.c (get_inner_reference): Use lshift.
17780 * fixed-value.c (do_fixed_divide): Likewise.
17781 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17782 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17783 (indirect_refs_may_alias_p): Likewise.
17784 (stmt_kills_ref_p_1): Likewise.
17786 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17788 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17790 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17792 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17793 scalar form of FABD instruction.
17795 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17797 * lra-constraints.c (process_alt_operands): Add checking alt
17798 number to choose the best alternative.
17800 2013-05-02 Richard Biener <rguenther@suse.de>
17802 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17803 bitmap and its handling.
17804 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17806 2013-05-02 Richard Biener <rguenther@suse.de>
17808 PR middle-end/57140
17809 * tree-inline.c (copy_loops): Properly handle removed loops.
17810 (copy_cfg_body): Mark destination loops for fixup if source
17811 loops needed fixup.
17813 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17816 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17817 generating simple_return for naked functions.
17819 2013-05-02 Martin Jambor <mjambor@suse.cz>
17821 PR middle-end/56988
17822 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17823 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17825 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17826 ipa_agg_replacement_value structures.
17827 (known_aggs_to_agg_replacement_list): Likewise.
17828 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17829 (read_agg_replacement_chain): Likewise.
17830 (ipcp_transform_function): Also check that by_ref flags match.
17832 2013-05-02 Richard Biener <rguenther@suse.de>
17834 * graphds.h (struct graph): Add obstack member.
17835 * graphds.c (new_graph): Initialize obstack and allocate
17837 (add_edge): Allocate edge from the obstack.
17838 (free_graph): Free the obstack instead of all edges and vertices.
17840 2013-05-02 Teresa Johnson <tejohnson@google.com>
17842 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17844 * cfg.c (update_bb_profile_for_threading): Ditto.
17845 * tree-inline.c (copy_bb): Ditto.
17846 (copy_edges_for_bb): Ditto.
17847 (initialize_cfun): Ditto.
17848 (copy_cfg_body): Ditto.
17849 (expand_call_inline): Ditto.
17850 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17851 (estimate_node_size_and_time): Ditto.
17852 (inline_merge_summary): Ditto.
17853 * cgraphclones.c (cgraph_clone_edge): Ditto.
17854 (cgraph_clone_node): Ditto.
17855 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17856 (compute_trg_info): Ditto.
17858 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17860 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17861 S reg when fp attribute set.
17862 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17864 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17866 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17868 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17869 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17870 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17872 2013-05-02 Richard Biener <rguenther@suse.de>
17874 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17875 (struct instantiate_cache_entry): New type.
17876 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17877 (struct instantiate_cache_type): New type.
17878 (set_instantiated_value, get_instantiated_value): Remove.
17879 (get_instantiated_value_entry): New function.
17880 (instantiate_scev_name): Use the new cache and adjust.
17881 (instantiate_scev_poly): Adjust.
17882 (instantiate_scev_binary): Likewise.
17883 (instantiate_array_ref): Likewise.
17884 (instantiate_scev_convert): Likewise.
17885 (instantiate_scev_not): Likewise.
17886 (instantiate_scev_3): Likewise.
17887 (instantiate_scev_2): Likewise.
17888 (instantiate_scev_r): Likewise.
17889 (instantiate_scev): Likewise.
17890 (resolve_mixers): Likewise.
17892 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17895 * lra-constraints.c (best_small_class_operands_num): Remove.
17896 (process_alt_operands): Remove small_class_operands_num. Take
17897 small classes operands into losers and only if the operand is not
17898 matched. Modify debugging output.
17899 (curr_insn_transform): Remove best_small_class_operands_num.
17902 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17904 * config/aarch64/aarch64-builtins.c
17905 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17906 * config/aarch64/aarch64-simd-builtins.def
17907 (reduc_splus_): Add new modes.
17908 (reduc_uplus_): New.
17909 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17910 (reduc_uplus_v4sf): Likewise.
17911 (reduc_splus_v4sf): Likewise.
17912 (aarch64_addv<mode>): Likewise.
17913 (reduc_uplus_<mode>): Likewise.
17914 (reduc_splus_<mode>): Likewise.
17915 (aarch64_addvv2di): Likewise.
17916 (reduc_uplus_v2di): Likewise.
17917 (reduc_splus_v2di): Likewise.
17918 (aarch64_addvv2si): Likewise.
17919 (reduc_uplus_v2si): Likewise.
17920 (reduc_splus_v2si): Likewise.
17921 (reduc_<sur>plus_<mode>): New.
17922 (reduc_<sur>plus_v2di): Likewise.
17923 (reduc_<sur>plus_v2si): Likewise.
17924 (reduc_<sur>plus_v4sf): Likewise.
17925 (aarch64_addpv4sf): Likewise.
17926 * config/aarch64/arm_neon.h
17927 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17928 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17929 add UNSPEC_SADDV, UNSPEC_UADDV.
17931 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17933 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17935 * config/aarch64/arm_neon.h
17936 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17938 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17940 * config/aarch64/aarch64-builtins
17941 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17943 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17945 * config/aarch64/aarch64-simd-builtins.def
17946 (reduc_smax_): New.
17947 (reduc_smin_): Likewise.
17948 (reduc_umax_): Likewise.
17949 (reduc_umin_): Likewise.
17950 (reduc_smax_nan_): Likewise.
17951 (reduc_smin_nan_): Likewise.
17954 (smax): Update for V2SF, V4SF and V2DF modes.
17957 (smin_nan): Likewise.
17958 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17959 (<su><maxmin><mode>3): ...This, refactor.
17960 (s<maxmin><mode>3): New.
17961 (<maxmin_uns><mode>3): Likewise.
17962 (reduc_<maxmin_uns>_<mode>): Refactor.
17963 (reduc_<maxmin_uns>_v4sf): Likewise.
17964 (reduc_<maxmin_uns>_v2si): Likewise.
17965 (aarch64_<fmaxmin><mode>: Remove.
17966 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17968 (vmin<q>_f<32,64>): Likewise.
17969 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17971 (su): Add mappings for smax, smin, umax, umin.
17973 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17974 (FMAXMIN): Rename as...
17975 (FMAXMIN_UNS): ...This.
17977 (fmaxminv): Likewise.
17978 (fmaxmin): Likewise.
17980 (maxmin_uns_op): Likewise.
17982 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17984 * config/aarch64/arm_neon.h
17985 (vac<ge, gt><sd>_f<32, 64>): Rename to...
17986 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
17987 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
17989 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17991 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
17992 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
17994 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17996 * config/aarch64/aarch64-simd.md
17997 (vcond<mode>_internal): Handle special cases for constant masks.
17998 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
17999 (vcondu<mode><mode>): Likewise.
18000 (vcond<v_cmp_result><mode>): New.
18002 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18004 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18005 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18006 * config/aarch64/aarch64-simd-builtins.def
18007 (cmeq): Update to BUILTIN_VALLDI.
18012 * config/aarch64/arm_neon.h
18013 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18014 to builtins or C as appropriate.
18016 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18018 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18020 (cmhi): Rename to...
18022 * config/aarch64/aarch64-simd.md
18023 (simd_mode): Add SF.
18024 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18025 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18026 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18027 (cstore<mode>_neg): ...This.
18028 * config/aarch64/iterators.md
18030 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18031 (COMPARISONS): New.
18032 (UCOMPARISONS): Likewise.
18033 (optab): Add missing comparisons.
18039 (VCMP_S): Likewise.
18040 (VCMP_U): Likewise.
18041 (V_cmp_result): Add DF, SF modes.
18042 (v_cmp_result): Likewise.
18044 (vmtype): Likewise.
18045 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18047 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18049 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18050 define_insn to define_insn_and_split.
18051 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18052 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18053 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18054 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18055 (thumb2_negscc): Likewise.
18057 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18059 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18061 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18063 * config/arm/thumb2.md: Remove trailing whitespaces.
18065 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18067 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18068 Use gen_int_mode rather than GEN_INT.
18070 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18072 * value-prof.c (stream_in_histogram_value): Remove the strayed
18075 2013-04-30 Richard Biener <rguenther@suse.de>
18077 PR middle-end/57122
18078 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18080 2013-04-30 Richard Biener <rguenther@suse.de>
18082 PR middle-end/57107
18083 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18085 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18087 PR rtl-optimization/56957
18088 PR rtl-optimization/57105
18089 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18090 variable. Use just INSN_UID for determining whether an insn
18091 should be only disconnected from the insn stream.
18092 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18094 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18096 PR tree-optimization/57104
18097 * tsan.c (instrument_expr): Don't instrument accesses to
18098 DECL_HARD_REGISTER VAR_DECLs.
18100 2013-04-30 Richard Biener <rguenther@suse.de>
18102 * function.h (loops_for_fn): New inline function.
18103 (set_loops_for_fn): Likewise.
18104 * cfgloop.h (place_new_loop): Add struct function parameter.
18105 (get_loop): Likewise.
18106 (get_loops): Likewise.
18107 (number_of_loops): Likewise.
18108 (fel_next): Adjust.
18109 (fel_init): Likewise.
18110 * cfg.c (get_loop_copy): Adjust.
18111 * cfgloop.c (flow_loops_dump): Likewise.
18112 (record_loop_exits): Likewise.
18113 (verify_loop_structure): Likewise.
18114 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18115 (estimate_reg_pressure_cost): Likewise.
18116 (mark_loop_exit_edges): Likewise.
18117 * cfgloopmanip.c (place_new_loop): Likewise.
18118 (add_loop): Likewise.
18119 (duplicate_loop): Likewise.
18120 * graph.c (draw_cfg_nodes): Likewise.
18121 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18122 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18123 (extract_affine_chrec): Likewise.
18124 (build_scop_iteration_domain): Likewise.
18125 * graphite.c (graphite_initialize): Likewise.
18126 * ira-build.c (create_loop_tree_nodes): Likewise.
18127 (more_one_region_p): Likewise.
18128 (rebuild_regno_allocno_maps): Likewise.
18129 (mark_loops_for_removal): Likewise.
18130 (mark_all_loops_for_removal): Likewise.
18131 (remove_unnecessary_regions): Likewise.
18132 (ira_build): Likewise.
18133 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18134 * loop-init.c (fix_loop_structure): Likewise.
18135 (gate_rtl_move_loop_invariants): Likewise.
18136 (gate_rtl_unswitch): Likewise.
18137 (gate_rtl_unroll_and_peel_loops): Likewise.
18138 (rtl_doloop): Likewise.
18139 * lto-streamer-in.c (input_cfg): Likewise.
18140 * lto-streamer-out.c (output_cfg): Likewise.
18141 * modulo-sched.c (sms_schedule): Likewise.
18142 * predict.c (tree_estimate_probability): Likewise.
18143 (tree_estimate_probability_driver): Likewise.
18144 (estimate_loops): Likewise.
18145 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18146 (move_sese_region_to_fn): Likewise.
18147 (debug_loop_num): Likewise.
18148 * tree-chrec.c (chrec_evaluate): Likewise.
18149 (hide_evolution_in_other_loops_than_loop): Likewise.
18150 (chrec_component_in_loop_num): Likewise.
18151 (reset_evolution_in_loop): Likewise.
18152 (evolution_function_is_invariant_rec_p): Likewise.
18153 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18154 * tree-inline.c (copy_loops): Likewise.
18155 (copy_cfg_body): Likewise.
18156 (tree_function_versioning): Likewise.
18157 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18158 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18160 (add_to_evolution_1): Likewise.
18161 (scev_const_prop): Likewise.
18162 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18163 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18164 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18165 (tree_ssa_lim_initialize): Likewise.
18166 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18167 (verify_loop_closed_ssa): Likewise.
18168 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18169 (tree_ssa_loop_im): Likewise.
18170 (tree_ssa_loop_unswitch): Likewise.
18171 (tree_vectorize): Likewise.
18172 (check_data_deps): Likewise.
18173 (tree_ssa_loop_ivcanon): Likewise.
18174 (tree_ssa_loop_bounds): Likewise.
18175 (tree_complete_unroll): Likewise.
18176 (tree_complete_unroll_inner): Likewise.
18177 (tree_parallelize_loops): Likewise.
18178 (tree_ssa_loop_prefetch): Likewise.
18179 (tree_ssa_loop_ivopts): Likewise.
18180 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18181 * tree-vectorizer.c (vectorize_loops): Likewise.
18183 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18185 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18186 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18187 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18188 with EABI_LINK_SPEC.
18190 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18193 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18196 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18199 * lra-constraints.c (process_alt_operands): Discourage a bit more
18200 using memory for pseudos. Print cost dump for alternatives.
18201 Modify cost values for conflicts with early clobbers.
18202 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18204 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18207 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18209 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18211 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18212 from/to S register.
18213 (movdi_aarch64): Support LDR/STR from/to D register.
18215 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18217 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18218 or higher by default.
18220 2013-04-29 Richard Biener <rguenther@suse.de>
18222 PR middle-end/57075
18223 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18224 even if not adding abnormal edges for calls that can make
18227 2013-04-29 Richard Biener <rguenther@suse.de>
18229 PR middle-end/57103
18230 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18232 (move_stmt_r): Remove redundant checking.
18234 2013-04-29 Teresa Johnson <tejohnson@google.com>
18237 * basic-block.h (apply_scale): New function.
18238 (apply_probability): Use apply_scale.
18239 * gimple-streamer-in.c (input_bb): Ditto.
18240 * lto-streamer-in.c (input_cfg): Ditto.
18241 * lto-cgraph.c (merge_profile_summaries): Ditto.
18242 * tree-optimize.c (execute_fixup_cfg): Ditto.
18243 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18244 (copy_edges_for_bb): Ditto.
18245 (copy_cfg_body): Ditto.
18247 2013-04-29 Tom de Vries <tom@codesourcery.com>
18249 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18250 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18251 (tail_merge_optimize): Handle current_loops == NULL.
18253 2013-04-26 Jeff Law <law@redhat.com>
18255 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18256 (simplify_cond_using_ranges): Generalize code to simplify
18257 COND_EXPRs where one argument is a constant and the other
18258 is an SSA_NAME created by an integral type conversion.
18260 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18262 * config/arm/arm.md (store_minmaxsi): Use only when
18263 optimize_insn_for_size_p.
18265 2013-04-29 Christian Bruel <christian.bruel@st.com>
18268 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18270 2013-04-29 Richard Biener <rguenther@suse.de>
18272 PR middle-end/57089
18273 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18274 loop tree make sure to schedule a fixup for the child as well.
18275 (expand_omp_for_generic): Properly add loops.
18276 (expand_omp_for_static_nochunk): Likewise.
18277 (expand_omp_for_static_chunk): Likewise.
18278 (expand_omp_for): For the degenerate case fixup loops.
18279 (expand_omp_sections): Fix default bb placement in loops.
18280 (expand_omp_atomic_pipeline): Properly add loops.
18282 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18284 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18286 2013-04-29 Tom de Vries <tom@codesourcery.com>
18288 * tree-ssa-tail-merge.c: Update header comment.
18290 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18292 * config/aarch64/arm_neon.h
18293 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18294 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18295 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18296 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18297 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18298 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18299 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18300 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18302 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18304 * config/aarch64/aarch64-simd.md
18305 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18306 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18307 fix_trunc, fixuns_trunc.
18308 (ftrunc<VDQF:mode>2): New.
18309 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18310 (fix_trunc_optab): New.
18312 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18314 * config/aarch64/aarch64-builtins.c
18315 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18316 iceilf, lround, iroundf.
18318 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18321 * config/i386/i386.h (enum ix86_tune_indices)
18322 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18323 New, split from X86_TUNE_INTER_UNIT_MOVES.
18324 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18325 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18326 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18327 (TARGET_INTER_UNIT_MOVES): Remove.
18328 * config/i386/i386.c (initial_ix86_tune_features): Update.
18329 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18330 (ix86_expand_convert_uns_didf_sse): Use
18331 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18332 (ix86_expand_vector_init_one_nonzero): Ditto.
18333 (ix86_expand_vector_init_interleave): Ditto.
18334 (inline_secondary_memory_needed): Return true for moves from SSE class
18335 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18336 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18337 * config/i386/constraints.md (Yi, Ym): Depend on
18338 TARGET_INTER_UNIT_MOVES_TO_VEC.
18339 (Yj, Yn): New constraints.
18340 * config/i386/i386.md (*movdi_internal): Change constraints of
18341 operand 1 from Yi to Yj and from Ym to Yn.
18342 (*movsi_internal): Ditto.
18343 (*movdf_internal): Ditto.
18344 (*movsf_internal): Ditto.
18345 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18346 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18347 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18348 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18349 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18350 * config/i386/sse.md (movdi_to_sse): Ditto.
18351 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18352 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18353 TARGET_INTER_UNIT_MOVES.
18354 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18355 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18356 instead of TARGET_INTER_UNIT_MOVES.
18357 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18358 operand 1 from Yi to Yj and from Ym to Yn.
18360 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18362 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18363 (float_truncate_hi_): Likewise.
18364 (float_extend_lo_): Likewise.
18365 (float_truncate_lo_): Likewise.
18366 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18367 (aarch64_float_extend_lo_v2df): Likewise.
18368 (vec_unpacks_hi_v4sf): Likewise.
18369 (aarch64_float_truncate_lo_v2sf): Likewise.
18370 (aarch64_float_truncate_hi_v4sf): Likewise.
18371 (vec_pack_trunc_v2df): Likewise.
18372 (vec_pack_trunc_df): Likewise.
18374 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18376 * config/aarch64/aarch64-builtins.c
18377 (aarch64_fold_builtin): Fold float conversions.
18378 * config/aarch64/aarch64-simd-builtins.def
18379 (floatv2si, floatv4si, floatv2di): New.
18380 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18381 * config/aarch64/aarch64-simd.md
18382 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18383 * config/aarch64/iterators.md (FLOATUORS): New.
18384 (optab): Add float, floatuns.
18385 (su_optab): Likewise.
18387 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18389 * config/aarch64/aarch64-builtins.c
18390 (aarch64_builtin_vectorized_function): Use new names for
18392 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18393 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18394 (fcvtzu): Split as...
18395 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18396 (fcvtas): Split as...
18397 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18398 (fcvtau): Split as...
18399 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18400 (fcvtps): Split as...
18401 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18402 (fcvtpu): Split as...
18403 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18404 (fcvtms): Split as...
18405 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18406 (fcvtmu): Split as...
18407 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18408 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18409 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18410 (lfrintnusf, lfrintnudf): Likewise.
18411 * config/aarch64/aarch64-simd.md
18412 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18414 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18415 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18416 (fcvt_pattern): Likewise.
18418 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18420 * config/aarch64/aarch64-simd.md
18421 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18422 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18424 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18426 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18427 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18428 (vrnd<a,m,n,p>_f32): Implement using builtins.
18429 (vrnd<i,x><q>_f<32, 64>): New.
18431 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18433 * config/aarch64/aarch64-builtins.c
18434 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18435 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18436 (frintz): Rename to...
18438 (frintp): Rename to...
18440 (frintm): Rename to...
18442 (frinti): Rename to...
18443 (nearbyint): ...this.
18444 (frintx): Rename to...
18446 (frinta): Rename to...
18448 * config/aarch64/aarch64-simd.md
18449 (aarch64_frint<frint_suffix><mode>): Delete.
18450 (<frint_pattern><mode>2): Convert to insn.
18451 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18452 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18453 (frint_pattern): Likewise.
18454 (frint_suffix): Likewise.
18456 2013-04-29 Richard Biener <rguenther@suse.de>
18458 PR tree-optimization/57081
18459 * loop-init.c: Include tree-flow.h.
18460 (loop_optimizer_finalize): Free number of iteration estimates.
18461 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18463 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18465 PR tree-optimization/57083
18466 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18467 non-singleton shift count range, zero extend low_bound for uns case.
18469 * config/i386/predicates.md (general_vector_operand): New predicate.
18470 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18471 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18472 if they aren't nonimmediate operands. If their original values
18473 satisfy const_vector_equal_evenodd_p, don't shift them.
18474 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18475 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18476 if not nonimmediate_operand.
18477 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18478 instead of register_operand.
18479 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18481 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18483 * stor-layout.c (finalize_size_functions): Allocate a structure and
18484 reset cfun before dumping the functions.
18486 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18488 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18491 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18492 use xop_pmacsdqh if uns_p.
18493 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18494 the immediate rotate count.
18496 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18498 * rtl.h (struct rtx_def): Add comment for field jump.
18499 (LRA_SUBREG_P): New macro.
18500 * recog.c (register_operand): Check LRA_SUBREG_P.
18501 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18503 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18504 (lra_final_code_change): Skip subreg change for operators.
18505 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18506 if there are no operand changes.
18507 * lra-constraints.c (curr_insn_set): New.
18508 (match_reload): Set LRA_SUBREG_P.
18509 (emit_spill_move): Ditto.
18510 (check_and_process_move): Use curr_insn_set. Process only single
18511 set insns. Don't initialize sec_mem_p and change_p.
18512 (simplify_operand_subreg): Use LRA_SUBREG_P.
18513 (reg_in_class_p): New function.
18514 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18515 of #ifdef. Add code to remove cycling.
18516 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18517 non-null disp. Reload inner instead of disp when base and index
18518 are null. Try to put lo_sum into register.
18519 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18520 (check_and_process_move): Move code for move cost check to
18521 simple_move_p. Remove equiv_substitution.
18522 (simple_move_p): New function.
18523 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18524 curr_insn_set. Call check_and_process_move only for single set
18525 insns. Use the new function. Move call of check_and_process_move
18526 after operand equiv substitution and address process.
18528 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18531 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18532 with nonlocal goto receivers or returns twice calls, ignore
18533 unininitialized values from abnormal edges to nl goto receiver
18534 or returns twice call.
18536 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18538 PR tree-optimization/57051
18539 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18540 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18543 2013-04-26 Richard Biener <rguenther@suse.de>
18545 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18546 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18547 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18548 (fixup_loop_arrays_after_move): New function.
18549 (move_sese_region_to_fn): Properly outline the loop tree parts
18550 of the SESE region.
18552 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18554 * config/i386/i386.md (type, unit): Fix long lines.
18556 2013-04-26 Richard Biener <rguenther@suse.de>
18558 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18559 (lto-streamer-out.o): Likewise.
18560 * cfgloop.c (init_loops_structure): Export, add struct function
18561 argument and adjust.
18562 (flow_loops_find): Adjust.
18563 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18564 (init_loops_structure): Declare.
18565 * lto-streamer-in.c: Include cfgloop.h.
18566 (input_cfg): Input the loop tree.
18567 * lto-streamer-out.c: Include cfgloop.h.
18568 (output_cfg): Output the loop tree.
18569 (output_struct_function_base): Do not drop PROP_loops.
18571 2013-03-26 Richard Biener <rguenther@suse.de>
18573 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18574 (pass_build_cfg): Provide PROP_loops.
18575 (move_sese_region_to_fn): Remove loops that are outlined into fn
18577 * tree-inline.c: Include cfgloop.h.
18578 (initialize_cfun): Do not drop PROP_loops.
18579 (copy_loops): New function.
18580 (copy_cfg_body): Copy loop structure.
18581 (tree_function_versioning): Initialize destination loop tree.
18582 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18583 (pass_parallelize_loops): Do IL verification.
18584 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18585 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18586 the CFG make sure we fixup loops as well.
18587 * tree-ssa-tail-merge.c: Include cfgloop.h.
18588 (replace_block_by): When merging loop latches mark loops for fixup.
18589 * lto-streamer-out.c (output_struct_function_base): Drop
18590 PROP_loops for now.
18591 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18592 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18593 * ipa-split.c: Include cfgloop.h.
18594 (split_function): Add the new return block to the loop tree root.
18595 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18596 whether we have removed the forwarder block.
18597 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18598 * cfgloop.h (place_new_loop): Declare.
18599 * cfgloopmanip.c (place_new_loop): Export.
18600 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18601 (tree-switch-conversion.o): Likewise.
18602 (tree-complex.o): Likewise.
18603 (tree-inline.o): Likewise.
18604 (tree-ssa-tailmerge.o): Likewise.
18605 (ipa-split.o): Likewise.
18606 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18607 (tree-ssa-copy.o): Likewise.
18608 * tree-switch-conversion.c: Include cfgloop.h
18609 (process_switch): If we emit a bit-test cascade, schedule loops
18611 * tree-complex.c: Include cfgloop.h.
18612 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18613 * asan.c: Include cfgloop.h.
18614 (create_cond_insert_point): Properly add new basic-blocks to
18615 loops, schedule loop fixup.
18616 * cfgloop.c (verify_loop_structure): Check that looks are not
18618 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18620 (expand_omp_for_generic): Likewise.
18621 (expand_omp_sections): Likewise.
18622 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18623 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18624 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18625 is initialized, not when loops are present.
18626 * tree-parloops.c (parallelize_loops): Remove checking here.
18627 * passes.c (init_optimization_passes): Schedule a copy-propagation
18628 pass before complete unrolling of inner loops.
18630 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18632 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18633 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18634 (colorize_init): Add argument to _WIN32 version.
18635 * toplev.c: Include diagnostic-color.h.
18636 (process_options): Default to -fdiagnostics-color=auto if
18637 GCC_COLORS env var is in the environment.
18638 * common.opt (fdiagnostics-color=): Add Var and Init.
18639 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18640 env var is in the environment, the default is auto rather than never.
18642 * diagnostic.h (file_name_as_prefix): Add context argument.
18643 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18644 the string as locus.
18645 * langhooks.c (lhd_print_error_function): Adjust caller.
18647 2013-04-25 Lawrence Crowl <crowl@google.com>
18649 * var-tracking.c (shared_hash_def::htab):
18650 Change type to hash_table. Update dependent calls and types.
18652 2013-04-25 Lawrence Crowl <crowl@google.com>
18654 * Makefile.in: Update as needed below.
18656 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18657 Move declaration to after the type's method definitons.
18659 * attribs.c (htab_t scoped_attributes::attribute_hash):
18660 Change type to hash_table. Update dependent calls and types.
18662 * bitmap.c (htab_t bitmap_desc_hash):
18663 Change type to hash_table. Update dependent calls and types.
18665 * cselib.c (htab_t cselib_hash_table):
18666 Change type to hash_table. Update dependent calls and types.
18668 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18669 (hash_string_slot_node): Move implementation into lto-streamer.h
18670 struct string_slot_hasher.
18671 (eq_string_slot_node): Likewise.
18673 * data-streamer-out.c: Update output_block::string_hash_table
18674 dependent calls and types.
18676 * dwarf2cfi.c (htab_t trace_index):
18677 Change type to hash_table. Update dependent calls and types.
18679 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18680 Change type to hash_table. Update dependent calls and types.
18681 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18682 (htab_t optimize_external_refs::map): Likewise.
18683 (htab_t output_comp_unit::extern_map): Likewise.
18684 (htab_t output_comdat_type_unit::extern_map): Likewise.
18685 (htab_t output_macinfo::macinfo_htab): Likewise.
18686 (htab_t optimize_location_lists::htab): Likewise.
18687 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18689 * except.c (htab_t ehspec_hash_type):
18690 Change type to hash_table. Update dependent calls and types.
18691 (assign_filter_values::ttypes): Likewise.
18692 (assign_filter_values::ehspec): Likewise.
18693 (sjlj_assign_call_site_values::ar_hash): Likewise.
18694 (convert_to_eh_region_ranges::ar_hash): Likewise.
18696 * gcse.c (htab_t pre_ldst_table):
18697 Change type to hash_table. Update dependent calls and types.
18699 * ggc-common.c (htab_t saving_htab):
18700 Change type to hash_table. Update dependent calls and types.
18701 (htab_t loc_hash): Likewise.
18702 (htab_t ptr_hash): Likewise.
18703 (call_count): Rename ggc_call_count.
18704 (call_alloc): Rename ggc_call_alloc.
18705 (loc_descriptor): Rename make_loc_descriptor.
18706 (add_statistics): Rename ggc_add_statistics.
18708 * ggc-common.c (saving_htab):
18709 Change type to hash_table. Update dependent calls and types.
18711 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18712 (push_gimplify_context): Likewise.
18713 (pop_gimplify_context): Likewise.
18714 (struct gimple_temp_hash_elt): Added.
18715 (struct gimplify_hasher): Likewise.
18716 (struct gimplify_ctx.temp_htab):
18717 Change type to hash_table. Update dependent calls and types.
18719 * gimple-fold.c: Include gimplify-ctx.h.
18721 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18722 Change type to hash_table. Update dependent calls and types.
18723 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18724 avoid potential global name collision.
18726 * gimplify.c: Include gimplify-ctx.h.
18727 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18728 (htab_t gimplify_ctx::temp_htab):
18729 Update dependent calls and types for new type hash_table.
18730 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18731 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18733 * gimplify-ctx.h: New.
18734 (struct gimple_temp_hash_elt): Move from gimplify.c.
18735 (class gimplify_hasher): New.
18736 (struct gimplify_ctx): Move from gimple.h.
18737 (htab_t gimplify_ctx::temp_htab):
18738 Change type to hash_table. Update dependent calls and types.
18740 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18741 (htab_t ivs_params::newivs_index):
18742 Change type to hash_table. Update dependent calls and types.
18743 (htab_t ivs_params::params_index): Likewise.
18744 (htab_t print_generated_program::params_index): Likewise.
18745 (htab_t gloog::newivs_index): Likewise.
18746 (htab_t gloog::params_index): Likewise.
18748 * graphite.c: Include graphite-htab.h.
18749 4htab_t graphite_transform_loops::bb_pbb_mapping):
18750 Change type to hash_table. Update dependent calls and types.
18752 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18753 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18754 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18756 * graphite-dependences.c: Include graphite-htab.h.
18757 (loop_is_parallel_p): Change hash table type of parameter.
18759 * graphite-htab.h: New.
18760 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18761 (extern find_pbb_via_hash): Move from graphite-poly.h.
18762 (extern loop_is_parallel_p): Move from graphite-poly.h.
18763 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18765 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18766 (extern loop_is_parallel_p): Move to graphite-htab.h.
18767 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18769 * haifa-sched.c (htab_t delay_htab):
18770 Change type to hash_table. Update dependent calls and types.
18771 (htab_t delay_htab_i2): Likewise.
18773 * ira-color.c (htab_t allocno_hard_regs_htab):
18774 Change type to hash_table. Update dependent calls and types.
18776 * ira-costs.c (htab_t cost_classes_htab):
18777 Change type to hash_table. Update dependent calls and types.
18779 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18780 Change type to hash_table. Update dependent calls and types.
18782 * loop-iv.c (htab_t bivs):
18783 Change type to hash_table. Update dependent calls and types.
18785 * loop-unroll.c (htab_t opt_info::insns_to_split):
18786 Change type to hash_table. Update dependent calls and types.
18787 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18789 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18790 (struct string_slot_hasher): New.
18791 (htab_t output_block::string_hash_table):
18792 Change type to hash_table. Update dependent calls and types.
18794 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18795 (htab_t file_name_hash_table):
18796 Change type to hash_table. Update dependent calls and types.
18798 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18801 * lto-streamer.c (htab_t tree_htab):
18802 Change type to hash_table. Update dependent calls and types.
18804 * omp-low.c: Include gimplify-ctx.h.
18806 * passes.c (htab_t name_to_pass_map):
18807 Change type to hash_table. Update dependent calls and types.
18808 (pass_traverse): Rename to passes_pass_traverse.
18810 * plugin.c (htab_t event_tab):
18811 Change type to hash_table. Update dependent calls and types.
18813 * postreload-gcse.c (htab_t expr_table):
18814 Change type to hash_table. Update dependent calls and types.
18815 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18817 * sese.c (debug_rename_map_1): Make extern.
18818 (htab_t copy_bb_and_scalar_dependences::rename_map):
18819 Change type to hash_table. Update dependent calls and types.
18821 * sese.h (extern debug_rename_map): Move to .c file.
18823 * store-motion.c (htab_t store_motion_mems_table):
18824 Change type to hash_table. Update dependent calls and types.
18826 * trans-mem.c (htab_t tm_new_mem_hash):
18827 Change type to hash_table. Update dependent calls and types.
18829 * tree-browser.c (htab_t TB_up_ht):
18830 Change type to hash_table. Update dependent calls and types.
18832 * tree-cfg.c (htab_t discriminator_per_locus):
18833 Change type to hash_table. Update dependent calls and types.
18835 * tree-complex.c: Include tree-hasher.h
18836 (htab_t complex_variable_components):
18837 Change type to hash_table. Update dependent calls and types.
18839 * tree-eh.c (htab_t finally_tree):
18840 Change type to hash_table. Update dependent calls and types.
18842 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18843 struct int_tree_hasher.
18844 (extern int_tree_map_eq): Likewise.
18845 (uid_decl_map_hash): Removed.
18846 (extern decl_tree_map_eq): Likewise.
18848 * tree-hasher.h: New.
18849 (struct int_tree_hasher): New.
18850 (typedef int_tree_htab_type): New.
18852 * tree-inline.c: Include gimplify-ctx.h.
18854 * tree-mudflap.c: Include gimplify-ctx.h.
18856 * tree-parloops.c: Include tree-hasher.h.
18857 (htab_t eliminate_local_variables_stmt::decl_address):
18858 Change type to hash_table. Update dependent calls and types.
18859 (htab_t separate_decls_in_region::decl_copies): Likewise.
18861 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18862 Change type to hash_table. Update dependent calls and types.
18864 * tree-sra.c (candidates):
18865 Change type to hash_table. Update dependent calls and types.
18867 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18869 (int_tree_map_hash): Likewise.
18871 * tree-ssa-dom.c (htab_t avail_exprs):
18872 Change type to hash_table. Update dependent calls and types.
18874 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18875 Change type to hash_table. Update dependent calls and types.
18877 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18878 Change type to hash_table. Update dependent calls and types.
18880 * tree-ssa-phiopt.c (seen_ssa_names):
18881 Change type to hash_table. Update dependent calls and types.
18883 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18884 Change type to hash_table. Update dependent calls and types.
18886 * tree-ssa-uncprop.c (equiv):
18887 Change type to hash_table. Update dependent calls and types.
18889 2013-04-25 Jakub Jelinek <jakub@redhat.com>
18891 PR rtl-optimization/57003
18892 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18893 call note_stores with kill_clobbered_value callback again after
18894 killing regs_invalidated_by_call.
18896 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18898 * config/aarch64/aarch64-simd.md
18899 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18900 (aarch64_simd_bsl<mode>): Likewise.
18901 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18903 2013-04-25 Marek Polacek <polacek@redhat.com>
18905 PR tree-optimization/57066
18906 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
18908 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18910 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18912 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18914 * config/aarch64/aarch64-builtins.c
18915 (aarch64_fold_builtin): New.
18916 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18917 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18918 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18919 * config/aarch64/arm_neon.h
18920 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18922 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18923 Tejas Belagod <tejas.belagod@arm.com>
18925 * config/aarch64/aarch64-builtins.c
18926 (aarch64_gimple_fold_builtin): New.
18927 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18928 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18929 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18930 (addvv4sf): Update.
18931 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18933 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18935 * config/aarch64/aarch64.md
18936 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18938 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18940 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18941 (*ngcsi_uxtw): New pattern.
18943 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18944 Julian Brown <julian@codesourcery.com>
18946 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18947 (TB_DREG): Add T_V4HF.
18948 (v4hf_UP): New macro.
18949 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18950 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18951 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18953 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18954 (arm_vector_mode_supported_p): Handle V4HF.
18955 (arm_mangle_map): Handle V4HFmode.
18956 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18957 * config/arm/arm_neon_builtins.def: Add entries for
18958 vcvtv4hfv4sf, vcvtv4sfv4hf.
18959 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18960 (neon_vcvtv4hfv4sf): Likewise.
18961 * config/arm/neon-gen.ml: Handle half-precision floating point
18963 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18964 * config/arm/arm_neon.h: Regenerate.
18965 * config/arm/neon.ml (type elts): Add F16.
18966 (type vectype): Add T_float16x4, T_floatHF.
18967 (type vecmode): Add V4HF.
18968 (type features): Add Requires_FP_bit feature.
18969 (elt_width): Handle F16.
18970 (elt_class): Likewise.
18971 (elt_of_class_width): Likewise.
18972 (mode_of_elt): Refactor.
18973 (type_for_elt): Handle F16, fix error messages.
18974 (vectype_size): Handle T_float16x4.
18975 (vcvt_sh): New function.
18976 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18977 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18978 (string_of_mode): Handle V4HF.
18979 * doc/arm-neon-intrinsics.texi: Regenerate.
18981 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18983 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
18984 format specifier in 'X' case.
18986 2013-04-25 Alan Modra <amodra@gmail.com>
18989 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
18990 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
18991 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
18992 Repeat for many other rotate/shift and mask patterns using subregs.
18993 Name lshiftrt insns.
18994 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
18995 on WORDS_BIG_ENDIAN.
18997 2013-04-25 Alan Modra <amodra@gmail.com>
18999 * config.gcc: Support little-endian powerpc-linux targets.
19000 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19001 (LINK_OS_LINUX_SPEC): Define.
19002 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19003 Preserve MASK_LITTLE_ENDIAN.
19004 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19005 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19006 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19007 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19008 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19009 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19010 Correct fp word order for little-endian. Don't shift toc entries
19011 smaller than a word for little-endian.
19012 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19013 (bswapdi2 splits): Correct low-part subreg for little-endian.
19014 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19015 low/high where such is correct only for be.
19016 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19017 little-endian for -mcall-aixdesc.
19019 2013-04-25 Alan Modra <amodra@gmail.com>
19021 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19022 replace_equiv_address_nv.
19024 2013-04-25 Alan Modra <amodra@gmail.com>
19026 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19028 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19031 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19032 * rtl.h (struct rtx_def): ...
19034 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19036 PR rtl-optimizations/57046
19037 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19038 for multi-reg splits.
19040 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19042 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19044 2013-04-24 Sterling Augustine <saugustine@google.com>
19046 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19047 (comp_dir_string, debug_str_dwo_section): New.
19048 (DEBUG_STR_DWO_SECTION): Rename to ...
19049 (DEBUG_DWO_STR_SECTION): ... this.
19050 (DEBUG_NORM_STR_SECTION): Delete.
19051 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19052 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19053 (find_AT_string): Move most logic to ...
19054 (find_AT_string_in_table): ... here. New.
19055 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19056 add_skeleton_AT_string. Delete logic.
19057 (output_skeleton_debug_sections): Remove call to
19058 add_top_level_skeleton_die_attrs.
19059 (add_comp_dir_attribute): Move logic to comp_dir_string.
19060 (dwarf2out_init): Initialize debug_str_dwo_section.
19061 (output_indirect_string): Call find_string_form.
19062 (output_indirect_strings): Rewrite.
19063 (prune_unused_types): Empty skeleton_debug_str_hash.
19064 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19065 (dwarf2out_finish): Call output_indirect_strings.
19067 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19069 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19071 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19073 * rtl.h (struct rtx_def): Add comment for field jump.
19074 (LRA_SUBREG_P): New macro.
19075 * recog.c (register_operand): Check LRA_SUBREG_P.
19076 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19078 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19079 (lra_final_code_change): Skip subreg change for operators.
19080 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19081 if there are no operand changes.
19082 * lra-constraints.c (curr_insn_set): New.
19083 (match_reload): Set LRA_SUBREG_P.
19084 (emit_spill_move): Ditto.
19085 (check_and_process_move): Use curr_insn_set. Process only single
19086 set insns. Don't initialize sec_mem_p and change_p.
19087 (simplify_operand_subreg): Use LRA_SUBREG_P.
19088 (reg_in_class_p): New function.
19089 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19090 of #ifdef. Add code to remove cycling.
19091 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19092 non-null disp. Reload inner instead of disp when base and index
19093 are null. Try to put lo_sum into register.
19094 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19095 (check_and_process_move): Move code for move cost check to
19096 simple_move_p. Remove equiv_substitution.
19097 (simple_move_p): New function.
19098 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19099 curr_insn_set. Call check_and_process_move only for single set
19100 insns. Use the new function. Move call of check_and_process_move
19101 after operand equiv substitution and address process.
19103 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19105 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19106 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19107 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19109 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19111 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19113 2013-04-24 Marek Polacek <polacek@redhat.com>
19115 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19116 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19117 (select_loops_exit_conditions): Likewise.
19118 (number_of_iterations_for_all_loops): Likewise.
19119 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19120 (scev_analysis): Likewise.
19122 2013-04-02 Catherine Moore <clm@codesourcery.com>
19123 Chao-ying Fu <fu@mips.com>
19125 * config/mips/micromips.md (jraddiusp): New pattern.
19126 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19127 instruction if possible.
19129 2013-04-24 Alan Modra <amodra@gmail.com>
19131 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19133 2013-04-24 Julian Brown <julian@codesourcery.com>
19134 Chung-Lin Tang <cltang@codesourcery.com>
19136 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19137 dependency behavior in enumeration type DIE generation. Add TODO note
19138 to comments about future DW_FORM_sdata/udata re-work of related code.
19140 2013-04-23 Lawrence Crowl <crowl@google.com>
19142 * Makefile.in: Update as needed below.
19144 * hash-table.h (class hash_table):
19145 Correct many methods with parameter types compare_type to the correct
19146 value_type. (Correct code was unlikely to notice the change.)
19147 (hash_table::elements_with_deleted) New.
19148 (class hashtable::iterator): New.
19149 (hashtable::begin()): New.
19150 (hashtable::end()): New.
19151 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19153 * statistics.c (statistics_hashes):
19154 Change type to hash_table. Update dependent calls and types.
19156 * tree-into-ssa.c (var_infos):
19157 Change type to hash_table. Update dependent calls and types.
19159 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19160 Change type to hash_table. Update dependent calls and types.
19162 * tree-ssa-loop-im.c (struct mem_ref.refs):
19163 Change type to hash_table. Update dependent calls and types.
19165 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19166 Change type to hash_table. Update dependent calls and types.
19168 * tree-ssa-sccvn.c (vn_tables_s::nary):
19169 Change type to hash_table. Update dependent calls and types.
19170 (vn_tables_s::phis): Likewise.
19171 (vn_tables_s::references): Likewise.
19173 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19174 (vn_reference_eq): Update parameter and return types.
19176 * tree-ssa-structalias.c (pointer_equiv_class_table):
19177 Change type to hash_table. Update dependent calls and types.
19178 (location_equiv_class_table): Likewise.
19180 * tree-vect-data-refs.c: Consequential changes for making
19181 peeling a hash_table.
19183 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19184 (destroy_loop_vec_info): Dependent hash_table update.
19186 * tree-vectorizer.h (peeling_htab):
19187 Change type to hash_table. Update dependent calls and types.
19189 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19191 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19192 to check the register content is equal or not.
19193 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19194 to assign register content record.
19195 * lra-eliminations.c (update_reg_eliminate): Use
19196 lra_update_reg_val_offset to update register content offset.
19197 * lra-int.h (struct lra_reg): Add offset member.
19198 (lra_reg_val_equal_p): New static inline function.
19199 (lra_update_reg_val_offset): New static inline function.
19200 (lra_assign_reg_val): New static inline function.
19201 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19202 to assign register content record.
19203 (initialize_lra_reg_info_element): Initial offset to zero.
19205 2013-04-23 Catherine Moore <clm@codesourcery.com>
19207 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19208 operands. Record compression.
19210 2013-04-23 Xinliang David Li <davidxl@google.com>
19212 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19214 2013-04-23 Richard Biener <rguenther@suse.de>
19216 PR middle-end/57036
19217 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19218 parameter, only add abnormal goto edges from the copied body
19219 if the call could perform abnormal gotos.
19220 (copy_cfg_body): Adjust.
19222 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19224 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19226 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19228 * coretypes.h (gimple_stmt_iterator): Add struct to make
19231 2013-04-23 Richard Biener <rguenther@suse.de>
19233 PR tree-optimization/57026
19234 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19235 from SSA names occuring in abnormal PHI nodes.
19237 2013-04-22 Andi Kleen <ak@linux.intel.com>
19239 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19241 2013-04-22 Andi Kleen <ak@linux.intel.com>
19243 * lto/lto.c (print_lto_report_1): Declare early.
19244 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19246 2013-04-22 Andi Kleen <ak@linux.intel.com>
19248 * common.opt (-flto-report-wpa): Add.
19249 * doc/invoke.texi (-flto-report-wpa): Add.
19250 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19253 2013-04-22 Xinliang David Li <davidxl@google.com>
19255 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19256 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19257 * Makefile.in: New dependency
19259 David Daney <ddaney.cavm@gmail.com>
19261 * configure.ac (gcc_cv_as_micromips_support): Use the
19262 --fatal-warnings option.
19263 * configure: Regenerate.
19265 2013-04-22 Marek Polacek <polacek@redhat.com>
19268 * tsan.c (instrument_expr): Don't instrument expression
19269 in case its size is zero.
19271 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19275 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19277 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19279 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19281 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19282 (gimple_stmt_iterator): New typedef.
19283 * gimple.h (gimple_stmt_iterator): Rename to...
19284 (gimple_stmt_iterator_d): ... This.
19285 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19286 trees be valid for GIMPLE and GENERIC.
19287 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19288 * gimple-fold.c (gimple_fold_call): Call target hook
19289 gimple_fold_builtin.
19290 * hooks.c (hook_bool_gsiptr_false): New.
19291 * hooks.h (hook_bool_gsiptr_false): New.
19292 * target.def (fold_stmt): New.
19293 * doc/tm.texi: Regenerate.
19295 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19298 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19299 a set sp if no stack realignment.
19301 2013-04-22 Nick Clifton <nickc@redhat.com>
19303 * config.gcc (tilegx-linux): Extend extra_objs rather than
19305 (tilepro-linux): Likewise.
19307 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19309 * config/aarch64/aarch64-builtins.c
19311 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19312 (VAR<1-12>): Add MAP parameter.
19313 (BUILTIN_*): Likewise.
19314 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19315 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19316 (aarch64_ushl_n<mode>): Likewise.
19317 (aarch64_sshr_n<mode>): Likewise.
19318 (aarch64_ushr_n<mode>): Likewise.
19319 (aarch64_<maxmin><mode>): Likewise.
19320 (aarch64_sqrt<mode>): Likewise.
19321 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19322 (vshr<q>_n_*): Likewise.
19324 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19326 * config/aarch64/aarch64-builtins.c
19327 (aarch64_simd_builtin_type_mode): Handle SF types.
19329 (BUILTIN_GPF): Define.
19330 (aarch64_init_simd_builtins): Handle SF types.
19331 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19332 (frecps): Likewise.
19333 (frecpx): Likewise.
19334 * config/aarch64/aarch64-simd.md
19335 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19336 (aarch64_frecpe<mode>): New.
19337 (aarch64_frecps<mode>): Likewise.
19338 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19339 (v8type): Add frecp<esx>.
19340 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19341 (aarch64_frecps<mode>): Likewise.
19342 * config/aarch64/iterators.md (FRECP): New.
19343 (frecp_suffix): Likewise.
19344 * config/aarch64/arm_neon.h
19345 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19347 2013-04-22 Christian Bruel <christian.bruel@st.com>
19350 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19351 (REG_CLASS_NAMES): Idem.
19352 (REG_CLASS_CONTENTS): Idem.
19353 (REGCLASS_HAS_FP_REG): Idem.
19354 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19355 (sh_conditional_register_usage): Idem.
19357 2013-04-21 Jeff Law <law@redhat.com>
19359 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19360 (ssa_forward_propagate_and_combine): Use it.
19362 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19364 * lra.c: Update the flow chart diagram.
19366 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19368 PR rtl-optimization/56847
19369 * lra-constraints.c (process_alt_operands): Discourage alternative
19370 with non-matche doffsettable memory constraint fro memory with
19373 2013-04-19 Richard Biener <rguenther@suse.de>
19375 PR tree-optimization/56982
19376 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19378 * gimplify.c (gimplify_call_expr): Notice special calls.
19379 (gimplify_modify_expr): Likewise.
19380 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19381 abnormal control flow receivers.
19382 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19383 in the same way as cfun->has_nonlocal_labels.
19384 (gimple_purge_dead_abnormal_call_edges): Likewise.
19385 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19386 receivers start a basic-block.
19388 2013-04-19 Richard Biener <rguenther@suse.de>
19390 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19392 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19393 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19394 (SLP_TREE_LOAD_PERMUTATION): Add.
19395 (vect_transform_slp_perm_load): Adjust prototype.
19396 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19397 (vect_free_slp_instance): Likewise.
19398 (vect_create_new_slp_node): Likewise.
19399 (vect_supported_slp_permutation_p): Remove.
19400 (vect_slp_rearrange_stmts): Adjust.
19401 (vect_supported_load_permutation_p): Likewise. Inline
19402 vect_supported_slp_permutation_p here.
19403 (vect_analyze_slp_instance): Compute load permutations per
19404 slp node instead of per instance.
19405 (vect_get_slp_defs): Adjust.
19406 (vect_transform_slp_perm_load): Likewise.
19407 (vect_schedule_slp_instance): Remove redundant code.
19408 (vect_schedule_slp): Remove hack for PR56270, add it ...
19409 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19410 CSE loads for SLP. Adjust.
19412 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19414 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19415 spelling in two comments.
19417 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19420 * config/arm/arm.c (load_multiple_sequence): Require SP
19421 as base register for loads if SP is in the register list.
19423 2013-04-19 Martin Jambor <mjambor@suse.cz>
19425 PR tree-optimization/56718
19426 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19427 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19428 and made public. Adjusted all callers.
19429 (ipa_intraprocedural_devirtualization): New function.
19430 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19431 (ipa_intraprocedural_devirtualization): Likewise.
19432 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19434 2013-04-19 Richard Biener <rguenther@suse.de>
19436 PR tree-optimization/57000
19437 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19439 2013-04-19 Terry Guo <terry.guo@arm.com>
19441 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19443 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19444 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19445 (cortex_m4_fmacs): Use new reservations.
19446 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19448 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19450 PR rtl-optimization/56999
19451 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19453 (lra_coalesce): Remove split_origin_bitmap and related code.
19454 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19455 ranges if necessary.
19457 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19459 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19461 (ix86_expand_call): Remove clobbered_registers array and use
19462 x86_64_ms_sysv_extra_clobbered_registers instead.
19463 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19465 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19467 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19468 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19469 (*call_value_rex64_ms_sysv): Ditto.
19471 2013-04-18 Cary Coutant <ccoutant@google.com>
19473 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19474 parent instead of die_mark.
19476 2013-04-18 Diego Novillo <dnovillo@google.com>
19478 * gimple.c (create_gimple_tmp): New.
19479 (get_expr_type): New.
19480 (build_assign): New.
19481 (build_type_cast): New.
19482 * gimple.h (enum ssa_mode): Define.
19483 (gimple_seq_set_location): New.
19484 * asan.c (build_check_stmt): Change some gimple_build_* calls
19485 to use build_assign and build_type_cast.
19487 2013-04-18 Richard Biener <rguenther@suse.de>
19489 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19490 handle negative step. Remove redundant checks.
19491 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19492 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19493 for negative step and grouped loads fail to vectorize.
19495 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19497 * emit-rtl.c (reset_insn_used_flags): New function.
19498 (reset_all_used_flags): Use it.
19499 (verify_insn_sharing): New function.
19500 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19502 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19504 PR tree-optimization/56984
19505 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19506 and (x >> M) >= N don't register any assertion if N << M is the
19509 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19511 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19512 do not delete_insn insns that have not yet been emitted, only
19513 unlink them with remove_insn.
19514 * df-scan.c (df_insn_delete): Revert r197492.
19516 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19518 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19519 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19521 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19523 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19524 define_insn_and_split.
19525 (and_scc,ior_scc,negscc): Likewise.
19526 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19528 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19530 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19531 can benefit from using a sequence of LDRD instructions in epilogue
19532 instead of a single LDM instruction.
19534 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19537 * doc/extend.texi: Fix typo.
19539 2013-04-17 Richard Biener <rguenther@suse.de>
19541 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19542 (vect_build_slp_tree): ... here.
19543 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19544 match. Remove special-casing of mismatched complex loads.
19545 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19546 re-try the match with swapped commutative operands.
19547 (vect_supported_load_permutation_p): Remove special-casing of
19548 mismatched complex loads.
19549 (vect_analyze_slp_instance): Adjust.
19551 2013-04-17 Richard Biener <rguenther@suse.de>
19553 PR rtl-optimization/56921
19554 * cfgloop.h (struct loop): Add simple_loop_desc member.
19555 (struct niter_desc): Mark with GTY(()).
19556 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19557 * loop-iv.c (get_simple_loop_desc): Likewise.
19558 (free_simple_loop_desc): Likewise.
19561 2013-04-16 Richard Biener <rguenther@suse.de>
19563 PR rtl-optimization/56921
19564 * loop-init.c (pass_rtl_move_loop_invariants): Add
19565 TODO_do_not_ggc_collect to todo_flags_finish.
19566 (pass_rtl_unswitch): Same.
19567 (pass_rtl_unroll_and_peel_loops): Same.
19568 (pass_rtl_doloop): Same.
19570 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19572 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19573 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19574 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19576 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19577 * tree-streamer.c (record_common_node): Adjust reference in comment.
19579 2013-04-17 Terry Guo <terry.guo@arm.com>
19581 * config/arm/cortex-m4.md: Add a new bypass.
19583 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19585 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19587 (*subs_<optab><mode>_multp2): New pattern.
19588 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19589 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19591 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19593 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19594 (*subs_mul_imm_<mode>): New pattern.
19596 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19599 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19600 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19601 (vsx_movti_32bit): Change j->wa to O->wa.
19603 2013-04-16 Richard Biener <rguenther@suse.de>
19605 PR rtl-optimization/56921
19606 * loop-init.c (pass_rtl_move_loop_invariants): Add
19607 TODO_do_not_ggc_collect to todo_flags_finish.
19608 (pass_rtl_unswitch): Same.
19609 (pass_rtl_unroll_and_peel_loops): Same.
19610 (pass_rtl_doloop): Same.
19612 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19614 * config/arm/arm.c (emit_multi_reg_push): New declaration
19615 for an existing function.
19616 (arm_emit_strd_push): New function.
19617 (arm_expand_prologue): Used here.
19618 (arm_emit_ldrd_pop): New function.
19619 (arm_expand_epilogue): Used here.
19620 (arm_get_frame_offsets): Update condition.
19621 (arm_emit_multi_reg_pop): Add a special case for load of a single
19622 register with writeback.
19624 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19626 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19629 2013-04-16 Richard Biener <rguenther@suse.de>
19631 PR tree-optimization/56756
19632 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19633 (first_mem_ref_loc): New.
19634 (execute_sm): Place the load temporarily before a previous
19635 access instead of in the latch edge to ensure its SSA dependencies
19636 are defined at points dominating the load.
19638 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19640 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19641 correct fix by moving header and footer insn to the footer of
19642 the merged basic block. Clear BB_END of the merged-away block.
19644 PR middle-end/43631
19645 * emit-rtl.c (make_note_raw): New function.
19646 (link_insn_into_chain): New static inline function.
19647 (add_insn): Use it.
19648 (add_insn_before, add_insn_after): Factor insn chain linking code...
19649 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19650 using link_insn_into_chain.
19651 (note_outside_basic_block_p): New helper function for emit_note_after
19652 and emit_note_before.
19653 (emit_note_after): Use nobb variant of add_insn_after if the note
19654 should not be contained in a basic block.
19655 (emit_note_before): Use nobb variant of add_insn_before if the note
19656 should not be contained in a basic block.
19657 (emit_note_copy): Use make_note_raw.
19658 (emit_note): Likewise.
19659 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19660 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19661 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19662 the moved barrier the tail of the basic block it follows.
19663 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19665 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19667 PR tree-optimization/56962
19668 * gimple-ssa-strength-reduction.c (record_increment): Only set
19669 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19670 either rhs1 or rhs2 is equal to c->base_expr.
19672 2013-04-15 Richard Biener <rguenther@suse.de>
19674 PR tree-optimization/56933
19675 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19677 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19678 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19679 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19680 dependence check ...
19681 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19683 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19684 GROUP_READ_WRITE_DEPENDENCE.
19686 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19688 * emit-rtl.c (reset_all_used_flags): New function.
19689 (verify_rtl_sharing): Call reset_all_used_flags before and after
19690 performing the checks.
19692 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19694 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19695 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19696 * config/arm/constraints.md (De): New constraint.
19697 * config/arm/neon.md (anddi3_neon): Delete.
19698 (neon_vand<mode>): Expand to standard anddi3 pattern.
19699 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19700 Move earlier in the file.
19701 (neon_inv_logic_op2): Likewise.
19702 (arm_anddi_operand_neon): New predicate.
19704 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19706 * configure.ac (gcc_cv_ld_as_needed): Set
19707 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19708 Use -z ignore, -z record on *-*-solaris2*.
19709 (HAVE_LD_AS_NEEDED): Update comment.
19710 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19711 * configure: Regenerate.
19712 * config.in: Regenerate.
19713 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19714 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19715 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19716 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19717 equivalents. Fix markup.
19718 * doc/tm.texi: Regenerate.
19720 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19722 * config/i386/i386.opt: New option mstack-protector-guard=.
19723 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19724 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19725 TARGET_SSP_TLS_GUARD.
19726 * config/i386/i386.c (ix86_option_override_internal): Set
19727 ix86_stack_protector_guard.
19728 * config/i386/i386.md (stack_protect_set): Enable for
19729 TARGET_SSP_TLS_GUARD only.
19730 (stack_protect_set_<mode>): Ditto.
19731 (stack_protect_test): Ditto.
19732 (stack_protect_test_<mode>): Ditto.
19733 * doc/invoke.texi (i386 Option): Document.
19735 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19738 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19739 (S_MODES): Set H_MODE bit.
19740 (SF_MODES): Set only S_MODE and SF_MODE bits.
19741 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19742 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19743 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19744 <MODE_FLOAT>: Likewise.
19746 2013-04-15 Joey Ye <joey.ye@arm.com>
19748 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19750 2013-04-15 Joey Ye <joey.ye@arm.com>
19752 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19754 (thumb_far_jump_used_p): Count instruction size and set
19757 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19759 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19760 * resource.c (mark_target_live_regs): Reformat conditional block.
19762 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19764 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19765 notes, they are emitted only just before final.
19766 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19768 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19770 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19771 * cfgrtl.c (delete_insn): Call it here instead.
19772 * lra-spills.c (lra_final_code_change): Use delete_insn.
19773 * haifa-sched.c (sched_remove_insn): Likewise.
19774 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19775 returning to the nop pool.
19776 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19777 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19779 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19781 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19782 * doc/tm.texi: Regenerated.
19784 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19786 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19789 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19791 * df-core.c (df_find_def): Compare register numbers.
19792 (df_find_use): Likewise.
19794 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19797 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19798 lra_in_progress for return.
19800 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19802 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19803 define_insn into define_insn_and_split and emit movsicc patterns.
19805 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19807 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19809 2013-04-12 Richard Biener <rguenther@suse.de>
19811 * tree-pass.h (TODO_do_not_ggc_collect): New.
19812 * passes.c (execute_one_ipa_transform_pass): Honor
19813 TODO_do_not_ggc_collect.
19814 (execute_one_pass): Likewise.
19817 2013-04-10 Richard Biener <rguenther@suse.de>
19819 * passes.c (init_optimization_passes): Remove reload pass.
19820 * ira.c (do_reload): Merge into ...
19822 (rest_of_handle_reload): Remove.
19823 (pass_reload): Likewise.
19824 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19825 of reload for vzeroupper pass placement.
19827 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19829 PR tree-optimization/56918
19830 PR tree-optimization/56920
19831 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19832 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19833 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19834 use wide_mul_with_sign method.
19836 2013-04-12 Richard Biener <rguenther@suse.de>
19838 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19839 not be considered a gimple constant.
19841 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
19843 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19844 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19847 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19848 Jakub Jelinek <jakub@redhat.com>
19850 * opts.c: Include diagnostic-color.h.
19851 (common_handle_option): Handle OPT_fdiagnostics_color_.
19852 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19853 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19854 (diagnostic-color.o): New.
19855 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19856 (diagnostic_color_rule): New enum.
19857 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19858 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19859 the location string.
19860 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19861 either NULL, or color kind.
19862 * diagnostic-color.c: New file.
19863 * diagnostic-color.h: New file.
19864 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19866 * doc/invoke.texi (-fdiagnostics-color): Document.
19867 * pretty-print.h (pp_show_color): Define.
19868 (struct pretty_print_info): Add show_color field.
19869 * diagnostic.c: Include diagnostic-color.h.
19870 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19871 macros. Colorize error:, warning: etc. strings and also the location
19873 (diagnostic_show_locus): Colorize the caret line.
19874 * pretty-print.c: Include diagnostic-color.h.
19875 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19876 inside of %< %> quotes or quoted through q format modifier.
19878 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19880 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
19882 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19884 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19885 code in CC_NZ mode.
19886 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19889 2013-04-11 Marek Polacek <polacek@redhat.com>
19891 PR tree-optimization/48184
19892 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
19894 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19896 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19897 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19898 (skip_simple_arithmetic): Tidy up.
19899 * tree.h (skip_simple_constant_arithmetic): Declare.
19901 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19903 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19905 2013-04-11 Richard Biener <rguenther@suse.de>
19907 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19908 generate vector constants.
19910 2013-04-11 Richard Biener <rguenther@suse.de>
19912 PR tree-optimization/56878
19913 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19914 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19916 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19917 Prefer to align the DR with the most invariant base address.
19919 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19921 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19924 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19926 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19927 floating-point vector comparisons against 0.
19929 2013-04-11 Jakub Jelinek <jakub@redhat.com>
19931 PR tree-optimization/56899
19932 * fold-const.c (extract_muldiv_1): Apply distributive law
19933 only if TYPE_OVERFLOW_WRAPS (ctype).
19935 2013-04-11 Bin Cheng <bin.cheng@arm.com>
19938 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19939 loading has side effect.
19941 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19943 * config/sparc/sparc.c: Include tree-pass.h.
19944 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19945 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19946 head of file. Change return type. Split off gate function.
19947 (sparc_gate_work_around_errata): New function.
19948 (pass_work_around_errata): New pass definition.
19949 (insert_pass_work_around_errata) New pass insert definition to
19950 insert pass_work_around_errata just after delayed-branch scheduling.
19951 (sparc_option_override): Insert the pass.
19952 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19954 2013-04-10 David S. Miller <davem@davemloft.net>
19956 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19957 or -mcpu=hypersparc.
19959 * target.def (cstore_mode): New hook.
19960 * target.h: Include insn-codes.h
19961 * targhooks.c: Likewise.
19962 (default_cstore_mode): New function.
19963 * targhooks.h: Declare it.
19964 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19965 * doc/tm.texi: Rebuild.
19966 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19967 target hook, rather than inspecting the insn_data.
19968 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19969 (TARGET_CSTORE_MODE): Redefine.
19970 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19972 * config/sparc/predicates.md (cstore_result_operand): New special
19974 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19975 Use it for operand 0.
19976 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19977 (*snesi_special): Likewise.
19978 (*snesi_zero): Likewise.
19979 (*seqsi_zero): Likewise.
19980 (*sltu_insn): Likewise.
19981 (*sgeu_insn): Likewise.
19982 (*seqdi_special): Make operand 0 and comparison operation be of
19984 (*snedi_special): Likewise.
19985 (*snedi_special_vis3): Likewise.
19986 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
19987 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
19988 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
19989 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
19990 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
19991 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
19992 (*sltu_extend_sp64): Likewise.
19993 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
19994 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
19995 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
19996 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
19997 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
19999 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20001 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20002 (aarch64_start_file): Use the new function.
20004 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20005 Jason Merrill <jason@redhat.com>
20007 * common.opt: Add -gdwarf.
20008 * opts.c (common_handle_option): Handle it.
20009 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20011 2013-04-10 Richard Biener <rguenther@suse.de>
20013 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20014 (execute_one_ipa_transform_pass): But unconditionally here.
20015 (execute_one_pass): And here.
20016 (init_optimization_passes): Remove reload pass.
20017 * tree-pass.h (TODO_ggc_collect): Remove.
20018 (pass_reload): Likewise.
20019 * ira.c (do_reload): Merge into ...
20021 (rest_of_handle_reload): Remove.
20022 (pass_reload): Likewise.
20023 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20024 of reload for vzeroupper pass placement.
20025 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20026 and todo_flags_finish of all passes.
20028 2013-04-10 Richard Biener <rguenther@suse.de>
20030 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20031 first_const_oprnd field, rename first_def_type to first_op_type.
20032 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20033 (vect_get_and_check_slp_defs): Always use the type of the
20034 operand. Allow mixed vect_external_def, vect_constant_def types.
20035 (vect_get_constant_vectors): Handle mixed vect_external_def,
20036 vect_constant_def types.
20038 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20040 PR tree-optimization/55524
20041 * tree-ssa-math-opts.c
20042 (convert_mult_to_fma): Don't use an fms construct
20043 when we don't have an fms operation, but fnma, and it looks
20044 likely that we'll be able to use the latter.
20046 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20048 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20050 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20051 inline fail caused by overwritable functions.
20053 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20055 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20056 unnecessary bits in the constant power of two case.
20058 2013-04-10 Richard Biener <rguenther@suse.de>
20060 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20061 broken code swapping operands.
20062 (vect_build_slp_tree): Do not compute load permutations here.
20063 (vect_analyze_slp_instance): Compute load permutations here,
20064 after building the SLP tree.
20066 2013-04-09 Christian Bruel <christian.bruel@st.com>
20068 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20069 of next/prev_real_insn.
20071 2013-04-09 Jan Hubicka <jh@suse.cz>
20073 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20074 Drop aliased parameter.
20075 (function_and_variable_visibility): Do not handle alias pairs.
20076 * cgraph.c (varpool_externally_visible_p): Update prototype.
20077 * varpool.c (varpool_add_new_variable): Update.
20079 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20081 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20083 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20085 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20087 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20089 2013-04-09 Marek Polacek <polacek@redhat.com>
20091 PR tree-optimization/48762
20092 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20094 2013-04-09 Richard Biener <rguenther@suse.de>
20096 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20098 (vect_build_slp_tree): Likewise.
20099 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20100 calculating the cost of a SLP instance.
20101 (vect_analyze_slp_instance): Use it from here, after building
20104 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20106 PR middle-end/56883
20107 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20108 expand_omp_for_static_chunk): Use simple_p = true in
20109 force_gimple_operand_gsi calls when assigning to addressable decls.
20111 2013-04-09 Jeff Law <law@redhat.com>
20113 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20114 when the boolean was created by converting a wider object which
20115 had a boolean range.
20117 2013-04-09 Richard Biener <rguenther@suse.de>
20119 * tree-vectorizer.h (slp_void_p): Remove.
20120 (slp_tree): Typedef before _slp_tree declaration.
20121 (struct _slp_tree): Use a vector of slp_tree as children.
20122 (vect_get_place_in_interleaving_chain): Remove.
20123 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20125 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20127 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20128 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20129 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20130 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20131 Use slp_node instead of slp_void_p and adjust.
20133 2013-04-09 Richard Biener <rguenther@suse.de>
20135 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20136 work that is not necessary.
20138 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20140 PR tree-optimization/56854
20141 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20142 forward into clobber stmts if it would change MEM_REF lhs into
20145 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20147 * tree.c (type_hash_lookup, type_hash_add): Make static.
20148 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20150 2013-04-09 Richard Biener <rguenther@suse.de>
20152 * tree.h (unsave_expr_now): Remove.
20153 * tree-inline.c (mark_local_for_remap_r): Remove.
20154 (unsave_expr_1): Likewise.
20155 (unsave_r): Likewise.
20156 (unsave_expr_now): Likewise.
20157 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20158 (propagate_tree_value): Likewise.
20160 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20162 * doc/rtl.texi (sequence): Rewrite documentation to match the
20163 current use of SEQUENCE rtl objects.
20164 * rtl.def (SEQUENCE): Likewise.
20166 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20167 Update documentation.
20168 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20169 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20171 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20173 2013-04-08 Teresa Johnson <tejohnson@google.com>
20175 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20176 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20178 (estimate_edge_size_and_time): Add comment to suggest using rounding
20180 (estimate_node_size_and_time): Ditto.
20181 (remap_edge_change_prob): Use helper rounding divide methods.
20182 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20183 (gimple_mod_pow2_value_transform): Ditto.
20184 (gimple_mod_subtract_transform): Ditto.
20185 (gimple_ic_transform): Ditto.
20186 (gimple_stringops_transform): Ditto.
20187 * stmt.c (conditional_probability): Ditto.
20188 (emit_case_dispatch_table): Ditto.
20189 * lto-cgraph.c (merge_profile_summaries): Ditto.
20190 * tree-optimize.c (execute_fixup_cfg): Ditto.
20191 * cfgcleanup.c (try_forward_edges): Ditto.
20192 * cfgloopmanip.c (scale_loop_profile): Ditto.
20194 (duplicate_loop_to_header_edge): Ditto.
20195 (lv_adjust_loop_entry_edge): Ditto.
20196 * tree-vect-loop.c (vect_transform_loop): Ditto.
20197 * profile.c (compute_branch_probabilities): Ditto.
20198 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20199 * lto-streamer-in.c (input_cfg): Ditto.
20200 * gimple-streamer-in.c (input_bb): Ditto.
20201 * ipa-cp.c (update_profiling_info): Ditto.
20202 (update_specialized_profile): Ditto.
20203 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20204 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20206 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20207 (compute_trg_info): Ditto.
20208 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20209 (purge_dead_edges): Ditto.
20210 * loop-unswitch.c (unswitch_loop): Ditto.
20211 * cgraphclones.c (cgraph_clone_edge): Ditto.
20212 (cgraph_clone_node): Ditto.
20213 * tree-inline.c (copy_bb): Ditto.
20214 (copy_edges_for_bb): Ditto.
20215 (initialize_cfun): Ditto.
20216 (copy_cfg_body): Ditto.
20217 (expand_call_inline): Ditto.
20219 2013-04-08 Kai Tietz <ktietz@redhat.com>
20221 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20222 TARGET_CYGWIN64 by TARGET_64BIT.
20224 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20226 * config/epiphany/epiphany.md (GPR_1): New constant.
20227 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20228 * config/epiphany/epiphany.c (gen_compare_reg):
20229 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20230 is already in place.
20231 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20232 Don't require being called during rtl expansion; If y operlaps r0,
20234 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20235 (epiphany_expand_epilogue): Likewise.
20237 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20238 Don't use CC_FPmode for ORDERED / UNORDERED.
20239 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20241 * config/epiphany/constraints.md (CnL): New constraint.
20242 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20243 * config/epiphany/predicates.md (add_operand): Allow 1024.
20245 * config/epiphany/epiphany.md (logical_op): New code iterator.
20246 (op_mnc): New code attribute.
20247 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20248 (mov_f+1, mov_f+2): New peephole2 patterns.
20250 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20251 (cstoresi4): Also allow re-use of zero result when doing a NE
20252 comparison to a non-zero operand.
20253 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20255 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20256 Use gen_addsi3_i / gen_subsi3_i.
20258 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20262 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20263 contain anything but clobbers, at most one __builtin_stack_restore,
20264 optionally debug stmts and final resx, and if it has at least one
20265 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20266 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20267 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20268 which isn't defaut definition, remove them.
20269 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20270 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20271 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20272 with MEM_REF LHS with SSA_NAME address.
20274 2013-04-08 Jeff Law <law@redhat.com>
20276 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20278 2013-04-08 Richard Biener <rguenther@suse.de>
20280 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20282 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20283 determined vector type.
20284 (vect_analyze_data_refs): Likewise.
20285 (vect_get_new_vect_var): Adjust.
20286 (vect_create_destination_var): Preserve SSA name versions.
20287 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20288 not dump anything here.
20290 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20292 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20293 Add member lr_slot_known.
20294 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20296 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20297 Remove code that sets lr_slot_offset according to what a previous
20298 version of epiphany_emit_save_restore used to do.
20299 (epiphany_emit_save_restore): When doing an lr save or restore,
20300 set/verify lr_slot_known and lr_slot_offset.
20302 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20305 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20308 2013-04-08 Richard Biener <rguenther@suse.de>
20310 * alias.c (find_base_term): Fix thinko in previous change.
20312 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20314 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20315 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20316 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20317 if possible to compute val.
20318 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20319 For QImode integers don't require anything about precision. Use
20320 const_with_all_bytes_same to find out if the constant doesn't have
20321 repeated bytes in it.
20323 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20325 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20328 2013-04-08 Marek Polacek <polacek@redhat.com>
20330 PR rtl-optimization/48182
20331 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20334 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20337 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20338 nuses, make sure we have a label.
20340 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20343 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20344 (rs6000_emit_swdiv_low_precision): Remove.
20345 (rs6000_emit_swdiv): Rewrite to handle between one and four
20346 iterations of Newton-Raphson generally; modify required number of
20347 iterations for some cases.
20348 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20350 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20352 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20353 set-but-unused variable.
20355 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20356 basic blocks of released function bodies garbage-collectable.
20358 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20359 (struct rtl_opt_pass): Add TODO_df_finish.
20361 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20363 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20365 * config/arm/constraints.md (q): New constraint.
20366 * config/arm/ldrdstrd.md: New file.
20367 * config/arm/arm.md (ldrdstrd.md) New include.
20368 (arm_movdi): Use "q" instead of "r" constraint
20369 for double-word memory access.
20370 (movdf_soft_insn): Likewise.
20371 * config/arm/vfp.md (movdi_vfp): Likewise.
20372 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20373 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20374 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20375 (mem_ok_for_ldrd_strd): Likewise.
20376 (output_move_double): Update assertion.
20378 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20380 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20382 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20384 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20385 define_insn_and_split.
20386 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20388 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20390 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20391 define_insn_and_split.
20392 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20393 (shiftsi3_compare): New pattern.
20394 (rrx): New pattern.
20395 * config/arm/unspecs.md (UNSPEC_RRX): New.
20397 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20399 * config/arm/arm.md (negdi_extendsidi): New pattern.
20400 (negdi_zero_extendsidi): Likewise.
20402 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20404 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20405 define_insn_and_split.
20406 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20407 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20409 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20411 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20412 define_insn_and_split.
20413 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20414 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20416 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20418 * config/arm/arm.md (subsi3_carryin): New pattern.
20419 (subsi3_carryin_const): Likewise.
20420 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20421 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20423 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20425 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20427 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20429 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20430 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20432 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20434 * config/arm/arm.c (arm_expand_builtin): Change fcode
20435 type to unsigned int.
20437 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20439 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20441 2013-04-04 Ian Lance Taylor <iant@google.com>
20443 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20446 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20448 PR middle-end/56729
20449 * df-scan.c (df_insn_delete): Disable failing assert.
20451 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20453 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20454 New function prototype.
20455 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20456 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20457 (arm_builtin_vectorized_function): New function.
20459 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20461 * config/arm/arm_neon_builtins.def: New file.
20462 * config/arm/arm.c (neon_builtin_data): Move contents to
20463 arm_neon_builtins.def.
20464 (enum arm_builtins): Include neon builtin definitions.
20465 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20466 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20468 2013-04-04 Marek Polacek <polacek@redhat.com>
20470 PR tree-optimization/48186
20471 * predict.c (maybe_hot_frequency_p): Return false if
20472 HOT_BB_FREQUENCY_FRACTION is 0.
20473 (cgraph_maybe_hot_edge_p): Likewise.
20475 2013-04-04 Richard Biener <rguenther@suse.de>
20477 PR tree-optimization/56826
20478 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20481 2013-04-04 Richard Biener <rguenther@suse.de>
20483 PR tree-optimization/56213
20484 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20485 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20487 2013-04-04 Richard Biener <rguenther@suse.de>
20489 PR tree-optimization/56837
20490 * tree-loop-distribution.c (classify_partition): For non-zero
20491 values require that the value has the same precision as its
20492 mode to be useful as memset value.
20494 2013-04-03 Nick Clifton <nickc@redhat.com>
20496 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20497 (fmssf4): Use fmsf.s on E3V5 architectures.
20498 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20499 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20501 2013-04-03 Jeff Law <law@redhat.com>
20503 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20504 (lra-eliminations.o): Likewise.
20506 2013-04-03 Teresa Johnson <tejohnson@google.com>
20508 * gcov-io.c (compute_working_sets): Moved most of body of old
20509 compute_working_sets here from profile.c.
20510 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20511 (gcov_working_set_t): Moved typedef here from basic-block.h
20512 (compute_working_set): Declare.
20513 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20514 (get_working_sets): Renamed from compute_working_set,
20515 replace most of body with call to new compute_working_sets.
20516 (get_exec_counts): Replace call to compute_working_sets
20517 to get_working_sets.
20518 * profile.h (get_working_sets): Renamed from compute_working_set.
20519 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20520 to get_working_sets.
20521 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20522 * gcov-dump.c (dump_working_sets): New function.
20524 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20526 * hwint.c (sext_hwi, zext_hwi): New functions.
20527 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20528 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20529 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20530 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20531 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20532 (sext_hwi, zext_hwi): New functions.
20534 2013-04-03 Jeff Law <law@redhat.com>
20536 PR tree-optimization/56799
20537 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20538 back test for widening conversion erroneously dropped in prior change.
20540 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20543 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20544 instead of next_real_insn.
20546 2013-04-03 Marek Polacek <polacek@redhat.com>
20549 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20551 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20554 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20556 (thumb1_output_casesi): Likewise.
20557 (thumb2_output_casesi): Likewise.
20559 2013-04-03 Richard Biener <rguenther@suse.de>
20561 PR tree-optimization/56817
20562 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20564 (tree_unroll_loops_completely_1): ... new function to manually
20565 walk the loop tree, properly defering outer loops of unrolled
20566 loops to later iterations.
20568 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20570 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20571 (vectorizable_load): Likewise.
20572 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20573 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20575 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20577 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20580 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20582 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20584 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20586 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20588 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20590 PR tree-optimization/56790
20591 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20594 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20596 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20598 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20599 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20602 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20605 * tree.h (force_folding_builtin_constant_p): New decl.
20606 * builtins.c (force_folding_builtin_constant_p): New variable.
20607 (fold_builtin_constant_p): Fold immediately also if
20608 force_folding_builtin_constant_p.
20610 2013-04-03 Richard Biener <rguenther@suse.de>
20612 PR tree-optimization/56812
20613 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20614 DRs of the same interleaving chain are independent.
20616 2013-04-02 Jason Merrill <jason@redhat.com>
20618 * gdbinit.in (pbb): Use debug fn.
20620 2013-04-02 Lawrence Crowl <crowl@google.com>
20622 * sese.h (struct ivtype_map_elt_s): Remove unused.
20623 (extern debug_ivtype_map): Remove unused.
20624 (extern eq_ivtype_map_elts): Remove unused.
20625 * sese.c (debug_ivtype_map): Removed unused.
20626 (debug_ivtype_map_1): Removed unused.
20627 (debug_ivtype_elt): Remove unused.
20628 (eq_ivtype_map_elts): Remove unused.
20631 2013-04-02 Kai Tietz <ktietz@redhat.com>
20634 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20635 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20636 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20638 (legitimize_pe_coff_symbol): Likewise.
20639 (is_imported_p): New helper-function.
20640 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20641 for Windows x64 targets.
20642 (ix86_expand_prologue): Optimize for pe-coff targets.
20643 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20644 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20645 medium/large code-model.
20646 (legitimize_pic_address): Likewise.
20647 (legitimize_tls_address): Likewise.
20648 (ix86_expand_call): Likewise.
20649 (x86_output_mi_thunk): Likewise.
20650 (get_dllimport_decl): Add new beimport argument.
20651 (construct_plt_address): Don't assert for x64 pe-coff targets.
20652 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20654 (SYMBOL_FLAG_STUBVAR): New macro.
20655 (SYMBOL_REF_STUBVAR_P): Likewise.
20656 * config/i386/winnt.c (stub_list): New structure.
20657 (stub_head): New local variable.
20658 (i386_pe_record_stub): New function.
20659 (i386_pe_file_end): Emit refptr-stubs.
20661 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20663 PR rtl-optimization/56745
20664 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20665 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20668 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20669 and both of them are MEM_REFs, just compare first argument for
20670 equality and attempt to deal even with differing offsets.
20673 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20674 of gimple_clobber_p to be MEM_REF.
20675 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20676 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20677 after gimplification.
20678 * asan.c (get_mem_ref_of_assignment): Don't instrument
20679 gimple_clobber_p stmts.
20680 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20681 gimple_clobber_p stmt if they have MEM_REF lhs and
20682 are dead because of another gimple_clobber_p stmt.
20683 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20684 gimple_clobber_p stmts like debug stmts.
20685 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20686 that refer to unused VAR_DECLs or uninitialized values.
20687 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20688 gimple_clobber_p stmts if they refer to removed parameters.
20689 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20692 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20694 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20695 using SWI48 mode attribute.
20697 2013-04-02 Wei Mi <wmi@google.com>
20699 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20700 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20701 *<rotate_insn><mode>3_mask in i386.md.
20703 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20705 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20707 2013-04-02 Richard Biener <rguenther@suse.de>
20709 PR tree-optimization/56778
20710 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20711 Runtime alias tests are not supported for gather loads.
20712 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20713 stmts referenced from SSA operands before updating SSA form.
20715 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20716 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20718 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20719 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20720 * config/arm/cortex-a53.md: New file.
20721 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20722 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20723 * config/arm/arm.c (arm_issue_rate): Likewise.
20724 * config/arm/arm-tune.md: Regenerate
20725 * config/arm/arm-tables.opt: Regenerate.
20726 * config/arm/arm-cores.def: Add cortex-a53.
20728 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20730 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20733 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20735 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20736 scalar load/store operations using B/H registers.
20737 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20739 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20741 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20743 * config/aarch64/aarch64.c
20744 (aarch64_simd_scalar_immediate_valid_for_move): New.
20745 * config/aarch64/aarch64-protos.h
20746 (aarch64_simd_scalar_immediate_valid_for_move): New.
20747 * config/aarch64/constraints.md (Dh, Dq): New.
20748 * config/aarch64/iterators.md (hq): New.
20750 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20752 * reorg.c (get_branch_condition): Deal with conditional returns.
20753 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20755 2013-04-01 Wei Mi <wmi@google.com>
20757 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20758 Truncate operand 2 using %b asm operand modifier.
20759 (*<shift_insn><mode>3_mask): Ditto.
20760 (*<rotate_insn><mode>3_mask): Ditto.
20762 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20764 PR middle-end/56798
20765 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20767 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20769 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20771 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20773 2013-03-30 Lawrence Crowl <crowl@google.com>
20775 * dse.c (clear_alias_sets): Remove never set.
20776 (disqualified_clear_alias_sets): Remove never set.
20777 (clear_alias_mode_pool): Remove never set.
20778 (dse_step0): Remove condition that is never true.
20779 (canon_address): Remove condition that is never true.
20780 (dse_step7): Remove condition that is never true.
20781 (rest_of_handle_dse): Remove condition that is never true.
20782 (rest_of_handle_dse::did_global): Remove never read from above.
20783 (dse_step2_spill): Remove never called from above.
20784 (dse_step5_spill): Remove never called from above.
20786 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20788 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20789 JUMP_TABLE_DATA changes.
20790 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20791 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20792 (Insns) <jump_table_data>: New entry.
20793 * doc/tm.texi: Regenerate.
20795 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20797 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20798 for table jump at the end of a basic block using tablejump_p.
20799 * targhooks.c (default_invalid_within_doloop): Likewise.
20800 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20801 target hook implementation that is identical to the default hook.
20802 (rs6000_invalid_within_doloop): Remove.
20804 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20805 unused variable from tablejump_p call.
20807 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20808 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20809 (INSN_DELETED_P): Likewise.
20810 (emit_jump_table_data): New prototype.
20811 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20812 after 4th as unused.
20813 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20814 * sched-vis.c (print_insn): Likewise.
20815 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20816 insn for compatibility with back ends that use next_active_insn to
20817 identify jump table data.
20818 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20819 (remove_insn): Likewise.
20820 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20822 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20823 (emit_jump_table_data): New function.
20825 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20826 basic block, a JUMP_TABLE_DATA never is.
20827 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20828 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20829 off from code handling real insns.
20830 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20831 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20832 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20833 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20834 is not a NONDEBUG_INSN_P.
20835 * ira-costs.c (scan_one_insn): Likewise.
20836 * jump.c (mark_all_labels): Likewise.
20837 (mark_jump_label_1): Likewise.
20838 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20839 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20841 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20842 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20843 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20844 JUMP_TABLE_DATA_P insns.
20845 (calculate_elim_costs_all_insns): Likewise.
20846 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20847 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20848 (delete_output_reload): Code style fixups.
20849 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20850 insn flags on this non-insn.
20851 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20852 as scheduling barriers, for pre-change compatibility.
20853 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20854 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20856 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20857 redundant JUMP_TABLE_DATA_P test.
20858 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20859 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20860 (frv_for_each_packet): Likewise.
20861 * config/i386/i386.c (min_insn_size): Likewise.
20862 (ix86_avoid_jump_mispredicts): Likewise.
20863 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20864 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20865 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20866 (mips16_insn_length): Robustify.
20867 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20868 (mips16_split_long_branches): Likewise.
20869 * config/pa/pa.c (pa_combine_instructions): Likewise.
20870 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20871 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20872 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20873 as contributing to pool range lengths.
20874 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20875 Remove redundant JUMP_TABLE_DATA_P test.
20876 (sh_loop_align): Likewise.
20877 (split_branches): Likewise.
20878 (sh_insn_length_adjustment): Likewise.
20879 * config/spu/spu.c (get_branch_target): Likewise.
20881 2013-03-29 Jan Hubicka <jh@suse.cz>
20883 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20884 gcov streaming; stream hot bb threshold to ltrans.
20885 * predict.c (get_hot_bb_threshold): Break out from ....
20886 (maybe_hot_count_p): ... here.
20887 (set_hot_bb_threshold): New function.
20888 * lto-section-in.c (lto_section_name): Add profile.
20889 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20890 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20891 and data-streamer.h
20892 (histogram_entry): New structure.
20893 (histogram, histogram_pool): New global vars.
20894 (histogram_hash): New structure.
20895 (histogram_hash::hash): New method.
20896 (histogram_hash::equal): Likewise.
20897 (account_time_size): New function.
20898 (cmp_counts): New function.
20899 (dump_histogram): New function.
20900 (ipa_profile_generate_summary): New function.
20901 (ipa_profile_write_summary): New function.
20902 (ipa_profile_read_summary): New function.
20903 (ipa_profile): Decide on threshold.
20904 (pass_ipa_profile): Add ipa_profile_write_summary and
20905 ipa_profile_read_summary.
20906 * Makefile.in (ipa.o): Update dependencies.
20907 * lto-streamer.h (LTO_section_ipa_profile): New section.
20909 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20911 * tree.h (VAR_P): New.
20913 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20916 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20918 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20920 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20921 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20922 (control_flow_insn_p): Likewise.
20923 * cfgrtl.c (duplicate_insn_chain): Likewise.
20924 * final.c (get_attr_length_1): Likewise.
20925 (shorten_branches): Likewise.
20926 (final_scan_insn): Likewise.
20927 * function.c (instantiate_virtual_regs): Likewise.
20928 * gcse.c (insert_insn_end_basic_block): Likewise.
20929 * ira-costs.c (scan_one_insn): Likewise.
20930 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20931 * lra.c (check_rtl): Likewise.
20932 * reload1.c (elimination_costs_in_insn): Likewise.
20933 * reorg.c (follow_jumps): Likewise.
20935 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20936 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20937 (thumb_far_jump_used_p): Likewise.
20938 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20939 (workaround_speculation): Likewise.
20940 (add_sched_insns_for_speculation): Likewise.
20941 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20942 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20943 (frv_for_each_packet): Likewise.
20944 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20945 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20946 (final_emit_insn_group_barriers): Likewise.
20947 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20948 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20949 (mips16_insn_length): Likewise.
20950 * config/pa/pa.c (pa_reorg): Likewise.
20951 (pa_combine_instructions): Likewise.
20952 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20953 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20954 (sh_reorg): Likewise.
20955 (split_branches): Likewise.
20956 * config/spu/spu.c (get_branch_target): Likewise.
20958 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20961 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20963 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
20964 Fix declaration name.
20966 2013-03-28 Lawrence Crowl <crowl@google.com>
20968 * graphds.h (struct graph.indicies): Remove unused.
20969 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20970 (SCOP_ORIGINAL_PDDRS): Remove unused.
20971 * sese.h (extern insert_loop_close_phis): Removed unused.
20972 (extern insert_guard_phis): Removed unused.
20973 (extern ivtype_map_elt_info): Removed unused.
20974 (new_ivtype_map_elt): Removed unused.
20975 * sese.c (ivtype_map_elt_info): Removed unused.
20977 2013-03-28 Lawrence Crowl <crowl@google.com>
20979 * Makefile.in: Add several missing include dependences.
20981 (test-dump.o): New. This object is not added to any executable,
20982 but is present for ad-hoc testing.
20984 (debug (const bitmap_head_def &)): New.
20985 (debug (const bitmap_head_def *)): New.
20987 (extern debug (const bitmap_head_def &)): New.
20988 (extern debug (const bitmap_head_def *)): New.
20990 (debug (edge_def &)): New.
20991 (debug (edge_def *)): New.
20993 (debug (basic_block_def &)): New.
20994 (debug (basic_block_def *)): New.
20996 (dump_node (const_tree, int, FILE *)): Correct source file.
20998 (debug (die_struct &)): New.
20999 (debug (die_struct *)): New.
21001 (extern debug (die_struct &)): New.
21002 (extern debug (die_struct *)): New.
21003 * gimple-pretty-print.c
21004 (debug (gimple_statement_d &)): New.
21005 (debug (gimple_statement_d *)): New.
21006 * gimple-pretty-print.h
21007 (extern debug (gimple_statement_d &)): New.
21008 (extern debug (gimple_statement_d *)): New.
21010 (debug (ira_allocno_copy &)): New.
21011 (debug (ira_allocno_copy *)): New.
21012 (debug (ira_allocno &)): New.
21013 (debug (ira_allocno *)): New.
21015 (extern debug (ira_allocno_copy &)): New.
21016 (extern debug (ira_allocno_copy *)): New.
21017 (extern debug (ira_allocno &)): New.
21018 (extern debug (ira_allocno *)): New.
21020 (debug (live_range &)): New.
21021 (debug (live_range *)): New.
21023 (debug (lra_live_range &)): New.
21024 (debug (lra_live_range *)): New.
21026 (debug (lra_live_range &)): New.
21027 (debug (lra_live_range *)): New.
21029 (debug (omega_pb_d &)): New.
21030 (debug (omega_pb_d *)): New.
21032 (extern debug (omega_pb_d &)): New.
21033 (extern debug (omega_pb_d *)): New.
21035 (debug (const rtx_def &)): New.
21036 (debug (const rtx_def *)): New.
21038 (debug_tree (tree): Move within file.
21039 (debug_raw (const tree_node &)): New.
21040 (debug_raw (const tree_node *)): New.
21041 (dump_tree_via_hooks (const tree_node *, int)): New.
21042 (debug (const tree_node &)): New.
21043 (debug (const tree_node *)): New.
21044 (debug_verbose (const tree_node &)): New.
21045 (debug_verbose (const tree_node *)): New.
21046 (debug_head (const tree_node &)): New.
21047 (debug_head (const tree_node *)): New.
21048 (debug_body (const tree_node &)): New.
21049 (debug_body (const tree_node *)): New.
21050 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21051 (debug (vec<tree, va_gc> &)): New.
21052 (debug (vec<tree, va_gc> *)): New.
21054 (extern debug (const rtx_def &)): New.
21055 (extern debug (const rtx_def *)): New.
21057 (debug_raw (simple_bitmap_def &)): New.
21058 (debug_raw (simple_bitmap_def *)): New.
21059 (debug (simple_bitmap_def &)): New.
21060 (debug (simple_bitmap_def *)): New.
21062 (extern debug (simple_bitmap_def &)): New.
21063 (extern debug (simple_bitmap_def *)): New.
21064 (extern debug_raw (simple_bitmap_def &)): New.
21065 (extern debug_raw (simple_bitmap_def *)): New.
21067 (debug (vinsn_def &)): New.
21068 (debug (vinsn_def *)): New.
21069 (debug_verbose (vinsn_def &)): New.
21070 (debug_verbose (vinsn_def *)): New.
21071 (debug (expr_def &)): New.
21072 (debug (expr_def *)): New.
21073 (debug_verbose (expr_def &)): New.
21074 (debug_verbose (expr_def *)): New.
21075 (debug (vec<rtx> &)): New.
21076 (debug (vec<rtx> *)): New.
21078 (extern debug (vinsn_def &)): New.
21079 (extern debug (vinsn_def *)): New.
21080 (extern debug_verbose (vinsn_def &)): New.
21081 (extern debug_verbose (vinsn_def *)): New.
21082 (extern debug (expr_def &)): New.
21083 (extern debug (expr_def *)): New.
21084 (extern debug_verbose (expr_def &)): New.
21085 (extern debug_verbose (expr_def *)): New.
21086 (extern debug (vec<rtx> &)): New.
21087 (extern debug (vec<rtx> *)): New.
21089 (_list_iter_cond_expr): Make inline instead of static.
21091 (debug (sreal &)): New.
21092 (debug (sreal *)): New.
21094 (extern debug (sreal &)): New.
21095 (extern debug (sreal *)): New.
21097 (extern debug_raw (const tree_node &)): New.
21098 (extern debug_raw (const tree_node *)): New.
21099 (extern debug (const tree_node &)): New.
21100 (extern debug (const tree_node *)): New.
21101 (extern debug_verbose (const tree_node &)): New.
21102 (extern debug_verbose (const tree_node *)): New.
21103 (extern debug_head (const tree_node &)): New.
21104 (extern debug_head (const tree_node *)): New.
21105 (extern debug_body (const tree_node &)): New.
21106 (extern debug_body (const tree_node *)): New.
21107 (extern debug (vec<tree, va_gc> &)): New.
21108 (extern debug (vec<tree, va_gc> *)): New.
21110 (debug (struct loop &)): New.
21111 (debug (struct loop *)): New.
21112 (debug_verbose (struct loop &)): New.
21113 (debug_verbose (struct loop *)): New.
21114 * tree-dump.c: Add header dependence.
21116 (extern debug (struct loop &)): New.
21117 (extern debug (struct loop *)): New.
21118 (extern debug_verbose (struct loop &)): New.
21119 (extern debug_verbose (struct loop *)): New.
21121 (debug (data_reference &)): New.
21122 (debug (data_reference *)): New.
21123 (debug (vec<data_reference_p> &)): New.
21124 (debug (vec<data_reference_p> *)): New.
21125 (debug (vec<ddr_p> &)): New.
21126 (debug (vec<ddr_p> *)): New.
21128 (extern debug (data_reference &)): New.
21129 (extern debug (data_reference *)): New.
21130 (extern debug (vec<data_reference_p> &)): New.
21131 (extern debug (vec<data_reference_p> *)): New.
21132 (extern debug (vec<ddr_p> &)): New.
21133 (extern debug (vec<ddr_p> *)): New.
21135 (debug (pt_solution &)): New.
21136 (debug (pt_solution *)): New.
21138 (extern debug (pt_solution &)): New.
21139 (extern debug (pt_solution *)): New.
21141 (debug (_var_map &)): New.
21142 (debug (_var_map *)): New.
21143 (debug (tree_live_info_d &)): New.
21144 (debug (tree_live_info_d *)): New.
21146 (extern debug (_var_map &)): New.
21147 (extern debug (_var_map *)): New.
21148 (extern debug (tree_live_info_d &)): New.
21149 (extern debug (tree_live_info_d *)): New.
21151 2013-03-28 Jan Hubicka <jh@suse.cz>
21153 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21155 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21157 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21158 record only when desired or required.
21160 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21162 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21163 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21165 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21167 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21168 (*andsi3_compare0_uxtw): New pattern.
21169 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21170 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21172 2013-03-28 Jan Hubicka <jh@suse.cz>
21174 * data-streamer-in.c (streamer_read_gcov_count): New function.
21175 * gimple-streamer-out.c: Include value-prof.h.
21176 (output_gimple_stmt): Output histogram.
21177 (output_bb): Use streamer_write_gcov_count.
21178 * value-prof.c: Include data-streamer.h
21179 (dump_histogram_value): Add HIST_TYPE_MAX.
21180 (stream_out_histogram_value): New function.
21181 (stream_in_histogram_value): New function.
21182 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21183 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21184 * data-streamer-out.c (streamer_write_gcov_count): New function.
21185 (streamer_write_gcov_count_stream): New function.
21186 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21187 (lto_output_node): Likewise.
21188 (input_node, input_edge): Likewise.
21189 * lto-streamer-out.c (output_cfg): Update streaming.
21190 * lto-streamer-in.c (input_cfg): Likewise.
21191 * data-streamer.h (streamer_write_gcov_count,
21192 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21193 * gimple-streamer-in.c: Include value-prof.h
21194 (input_gimple_stmt): Input histograms.
21195 (input_bb): Update profile streaming.
21197 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21199 * genmodes.c (emit_max_int): New function.
21200 (emit_insn_modes_h): Added call to emit_max_function.
21201 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21203 * machmode.def: Fixed comment.
21205 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21207 * combine.c (try_combine): Removed useless assert.
21208 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21210 2013-03-28 Marek Polacek <polacek@redhat.com>
21211 Richard Biener <rguenther@suse.de>
21213 PR tree-optimization/56695
21214 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21215 build signed result of a vector comparison.
21216 * tree-cfg.c (verify_gimple_comparison): Check that a result
21217 of a vector comparison has signed type.
21219 2013-03-28 Richard Biener <rguenther@suse.de>
21221 PR tree-optimization/37021
21222 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21223 do not restrict gaps between groups.
21224 * tree-vect-stmts.c (vectorizable_load): Properly account for
21225 a gap between groups.
21227 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21229 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21230 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21233 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21235 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21236 * doc/extend.texi (Named Address Spaces): Ditto.
21237 (Variable Attributes): Ditto.
21239 2013-03-27 Kai Tietz <ktietz@redhat.com>
21241 * config.build: Add support for cygwin x64 target.
21242 * config.gcc: Likewise.
21243 * config.host: Likewise.
21244 * configure.ac: Likewise
21245 * configure: Regenerated.
21247 2013-03-27 Kai Tietz <ktietz@redhat.com>
21249 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21250 * config/i386/t-cygwin-w64: New file.
21251 * config/i386/cygwin-w64.h: New file.
21252 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21253 and add support for x64-cygwin target.
21254 (CPP_SPEC): Likewise.
21255 (CXX_WRAP_SPEC_LIST): Undefine before define.
21256 (LIBGCJ_SONAME): Use 15 as version.
21258 2013-03-27 Richard Biener <rguenther@suse.de>
21260 PR tree-optimization/56716
21261 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21262 dumping for ref nodes.
21264 2013-03-27 Martin Jambor <mjambor@suse.cz>
21266 PR tree-optimization/55334
21267 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21268 restricted pointers to arrays.
21270 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21272 * Makefile.in (.SUFFIXES): Add .cc.
21273 (.c.o): Apply same recipe for implicit rule .cc.o.
21275 2013-03-27 Richard Biener <rguenther@suse.de>
21277 PR tree-optimization/37021
21278 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21279 REALPART/IMAGPART_EXPRs around the supported refs.
21280 * tree-ssa-structalias.c (find_func_aliases): Assume that
21281 floating-point values are not used to transfer pointers.
21283 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21285 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21286 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21287 * doc/tm.texi: Regenerate.
21288 * targhooks.h (default_has_ifunc_p): New.
21289 * targhooks.c (default_has_ifunc_p): Ditto.
21290 * config/linux-protos.h: New file.
21291 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21292 hook for linux which disables support of indirect functions in android.
21293 * config/linux-android.c: New file.
21294 * config/t-linux-android.c: Ditto.
21295 * config.gcc: Added new object file linux-android.o.
21296 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21297 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21298 * varasm.c (do_assemble_alias): Likewise.
21299 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21300 doesn't support indirect functions.
21301 * configure: Regenerate.
21303 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21306 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21307 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21310 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21312 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21314 2013-03-27 Terry Guo <terry.guo@arm.com>
21316 * config/arm/arm-cores.def: Added core cortex-r7.
21317 * config/arm/arm-tune.md: Regenerated.
21318 * config/arm/arm-tables.opt: Regenerated.
21319 * doc/invoke.texi: Added entry for core cortex-r7.
21321 2013-03-27 Walter Lee <walt@tilera.com>
21323 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21324 double-decrement of next_scratch_regno.
21326 2013-03-27 Walter Lee <walt@tilera.com>
21328 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21330 (insn_v1mulus): Ditto.
21331 (insn_v2muls): Ditto.
21333 2013-03-27 Walter Lee <walt@tilera.com>
21335 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21336 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21338 2013-03-27 Walter Lee <walt@tilera.com>
21340 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21341 (*sibcall_value): Ditto.
21343 2013-03-27 Walter Lee <walt@tilera.com>
21345 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21346 (insn_mnz_v8qi): ... this ...
21347 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21349 (insn_v<n>mnz): Replaced by ...
21350 (insn_v1mnz): ... this ...
21351 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21353 (insn_mz_<mode>): Replaced by ...
21354 (insn_mz_v8qi): ... this ...
21355 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21357 (insn_v<n>mz): Replaced by ...
21358 (insn_v1mz): ... this ...
21359 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21362 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21364 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21366 2013-03-26 Roland McGrath <mcgrathr@google.com>
21368 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21369 than fprintf with a non-constant, non-format string.
21371 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21373 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21374 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21375 operand 0 predicate.
21376 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21377 attribute. Use general_x64nomem_operand as operand 1 predicate.
21378 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21379 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21380 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21381 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21382 (mov<mode>_insv_1): Remove expander. Merge insn with
21383 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21384 Use general_x64nomem_operand as operand 1 predicate.
21385 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21386 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21387 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21388 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21389 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21390 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21391 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21392 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21393 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21394 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21395 (general_x64nomem_operand): Ditto.
21397 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21399 * config/rtems.opt: Add -pthread option.
21401 2013-03-26 Richard Biener <rguenther@suse.de>
21403 * alias.c (find_base_term): Avoid redundant and not used recursion.
21404 (base_alias_check): Get the initial base term from the caller.
21405 (true_dependence_1): Compute and pass base terms to base_alias_check.
21406 (write_dependence_p): Likewise.
21407 (may_alias_p): Likewise.
21409 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21411 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21412 PC-relative load in SI modes and above only.
21414 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21416 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21417 * config/arm/iwmmxt.md (WCGR0): Update.
21418 (WCGR1, WCGR2, WCGR3): Likewise.
21420 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21422 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21423 Use x64 and nox64 isa attributes.
21425 2013-03-26 Richard Biener <rguenther@suse.de>
21427 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21428 alignment computations and rely on get_object_alignment_1
21429 for the !TYPE_P case.
21430 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21432 2013-03-26 Walter Lee <walt@tilera.com>
21434 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21435 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21437 2013-03-25 Jeff Law <law@redhat.com>
21439 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21440 check for INTEGRAL_TYPE_P that was missing due to checking in
21441 wrong version of prior patch.
21443 2013-03-25 Walter Lee <walt@tilera.com>
21445 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21446 TILEGX_INSN_SHUFFLEBYTES1.
21447 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21449 (tilegx_builtins): Ditto.
21450 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21452 2013-03-25 Walter Lee <walt@tilera.com>
21454 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21455 (floatunssisf2): New pattern.
21456 (floatsidf2): New pattern.
21457 (floatunssidf2): New pattern.
21459 2013-03-25 Walter Lee <walt@tilera.com>
21461 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21462 tests for constraint J, K, N, P.
21464 2013-03-25 Walter Lee <walt@tilera.com>
21466 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21467 Use indirect/pcrel encoding.
21468 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21471 2013-03-25 Steve Ellcey <sellcey@mips.com>
21473 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21474 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21475 * config/mips/mips.c (mips_option_override): Set IMADD default.
21476 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21477 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21478 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21479 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21480 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21482 2013-03-25 Jeff Law <law@redhat.com>
21484 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21485 slightly to avoid creating and folding useless trees. Simplify
21486 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21488 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21490 * config/i386/i386.md (*zero_extendsidi2): Merge with
21491 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21492 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21493 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21494 targets. Clarify comment.
21496 2013-03-25 Martin Jambor <mjambor@suse.cz>
21498 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21499 pass-through jump functions differently.
21500 (ipa_read_jump_function): Likewise. Also use setter functions to set
21503 2013-03-25 Martin Jambor <mjambor@suse.cz>
21505 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21506 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21508 (ipa_get_indirect_edge_target): New function.
21509 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21510 ipa_get_indirect_edge_target. Update all callers.
21511 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21512 ipa_get_indirect_edge_target_1 instead of calling
21513 ipa_get_indirect_edge_target.
21514 (create_specialized_node): Pass aggvlas to
21515 ipcp_discover_new_direct_edges.
21517 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21519 * config/arm/arm.md (f_sels, f_seld): New types.
21520 (*cmov<mode>): New pattern.
21521 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21524 2013-03-25 Kai Tietz <ktietz@redhat.com>
21526 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21527 POSIX-printf for mingw-hosted builds.
21529 2013-03-25 Richard Biener <rguenther@suse.de>
21531 PR middle-end/56694
21532 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21533 must-not-throw stmt location.
21535 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21537 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21538 Emit load-acquire versions when acq is true.
21539 (arm_emit_store_exclusive): Add rel parameter.
21540 Emit store-release versions when rel is true.
21541 (arm_split_compare_and_swap): Use acquire-release instructions
21543 of barriers when appropriate.
21544 (arm_split_atomic_op): Likewise.
21545 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21546 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21547 (VUNSPEC_SLX): Likewise.
21548 (VUNSPEC_LDA): Likewise.
21549 (VUNSPEC_STL): Likewise.
21550 * config/arm/sync.md (atomic_load<mode>): New pattern.
21551 (atomic_store<mode>): Likewise.
21552 (arm_load_acquire_exclusive<mode>): Likewise.
21553 (arm_load_acquire_exclusivesi): Likewise.
21554 (arm_load_acquire_exclusivedi): Likewise.
21555 (arm_store_release_exclusive<mode>): Likewise.
21557 2013-03-25 Catherine Moore <clm@codesourcery.com>
21559 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21560 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21561 * config/mip/predicates.md (lwsp_swsp_operand,
21562 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21563 sb16_operand, db4_operand, db7_operand, ib3_operand,
21564 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21565 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21566 andi16_operand): New predicates.
21567 * config/mips/mips.md (compression): New attribute.
21568 (enabled): New attribute.
21569 (length): Consider compression in computing length.
21570 (shift_compression): New code attribute.
21571 (*add<mode>3): New operands. Record compression.
21572 (sub<mode>3): Likewise.
21573 (one_cmpl<mode>2): Likewise.
21574 (*and<mode>3): Likewise.
21575 (*ior<mode>3): Likewise.
21576 (unnamed pattern for xor): Likewise.
21577 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21578 (*<optab><mode>3): Likewise.
21579 (*mov<mode>_internal: Likewise.
21580 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21581 (mips_unsigned_immediate_p): New.
21582 (umips_lwsp_swsp_address_p): New.
21583 (m16_based_address_p): New.
21584 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21585 (mips_unsigned_immediate_p): New prototype.
21586 (lwsp_swsp_address_p): New prototype.
21587 (m16_based_address_p): New prototype.
21588 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21589 (mips_signed_immediate_p): New function.
21590 (m16_based_address_p): New function.
21591 (lwsp_swsp_address_p): New function.
21592 (mips_print_operand_punctuation): Recognize short delay slot insns
21593 for microMIPS.add<mode>3"
21595 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21598 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21599 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21601 2013-03-25 Richard Biener <rguenther@suse.de>
21603 PR tree-optimization/56689
21604 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21607 2013-03-25 Richard Biener <rguenther@suse.de>
21609 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21611 (memory_references): Likewise.
21612 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21613 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21614 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21615 (gather_mem_refs_in_loops): Fold into ...
21616 (analyze_memory_references): ... this. Move initialization
21617 to tree_ssa_lim_initialize.
21618 (fill_always_executed_in): Rename to ...
21619 (fill_always_executed_in_1): ... this.
21620 (fill_always_executed_in): Move contains_call computation to
21621 this new function from ...
21622 (tree_ssa_lim_initialize): ... here.
21623 (tree_ssa_lim): Call fill_always_executed_in.
21625 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21627 * postreload.c (reload_combine): Fix code detecting returns.
21629 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21631 * function.c (emit_use_return_register_into_block): On cc0 targets,
21632 do not emit the sequence between cc0 setter and user.
21634 2013-03-25 Kai Tietz <ktietz@redhat.com>
21636 * config/i386/predicates.md (local_symbolic_operand): Interpret
21637 dll-imported symbols as none-local.
21639 2013-03-25 Richard Biener <rguenther@suse.de>
21641 * tree-ssa-loop-im.c (struct depend): Remove.
21642 (struct lim_aux_data): Make depends a vec of gimples.
21643 (free_lim_aux_data): Adjust.
21644 (add_dependency): Likewise.
21645 (set_level): Likewise.
21647 2013-03-25 Richard Biener <rguenther@suse.de>
21649 PR middle-end/56434
21650 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21651 the pointer returned by calls with ECF_MALLOC set.
21653 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21655 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21657 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21659 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21660 using MMXMODE mode iterator.
21661 (*move<mode>_internal): Merge with *movv2sf_internal and
21662 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21664 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21666 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21667 (record_last_mem_set_info): Likewise.
21669 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21670 of XNEWVEC followed by memset.
21671 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21673 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21675 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21676 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21677 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21678 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21679 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21680 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21681 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21682 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21683 BARRIER_P instead of GET_CODE.
21685 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21687 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21688 inaccuracy in the probing code.
21690 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21691 (ctrapdi4): Likewise.
21693 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21695 * calls.c (expand_call): Add missing guard to code handling return
21696 of non-BLKmode structures in MSB.
21697 * function.c (expand_function_end): Likewise.
21699 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21701 * combine.c (try_combine): Adjust comment. Do not add the set of
21702 insn #0 if the destination indirectly is set or dies in insn #2.
21703 Tidy up code to distribute a new note.
21705 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21707 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21708 also for alternatives 16 and 17.
21710 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21712 * config/i386/sse.md (*mov<mode>_internal): Merge with
21713 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21714 Emit insn template depending on type attribute. Use
21715 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21716 movd instead of movq mnemonic for interunit moves. Rewrite mode
21717 attribute calculation. Remove unit attribute calculation.
21718 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21719 Set prefix_data16 attribute for DImode ssemov types.
21720 Use Ym instead of y for SSE-MMX conversion alternatives.
21721 Reorder operand constraints.
21723 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21725 * df.h (df_insn_delete): Adjust prototype.
21726 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21727 and let it decide whether mark the basic block dirty.
21728 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21729 * df-scan.c (df_insn_info_delete): New helper function, split
21730 off from df_insn_delete.
21731 (df_scan_free_bb_info): Use it.
21732 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21734 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21735 that the insn is actually an insn and it has a non-NULL basic block.
21736 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21738 2013-03-22 Richard Biener <rguenther@suse.de>
21740 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21742 (mem_ref_alloc): Do not allocate them.
21743 (refs_independent_p): Do not query or maintain a cache.
21745 2013-03-22 Richard Biener <rguenther@suse.de>
21747 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21748 (gather_mem_refs_in_loops): Do not compute it.
21749 (analyze_memory_references): Do not allocate it.
21750 (tree_ssa_lim_finalize): Do not free it.
21751 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21753 2013-03-22 Richard Biener <rguenther@suse.de>
21755 * is-a.h (as_a): Use gcc_checking_assert.
21757 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21759 * config/aarch64/aarch64.c (aarch64_print_operand): New
21760 format specifier for printing a constant in hex.
21761 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21762 format specifier for printing second operand.
21764 2013-03-22 Richard Biener <rguenther@suse.de>
21766 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21768 (gather_mem_refs_in_loops): Perform store accumulation here.
21769 (create_vop_ref_mapping_loop): Remove.
21770 (create_vop_ref_mapping): Likewise.
21771 (analyze_memory_references): Initialize refs_stored_in_loop.
21772 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21773 (record_indep_loop): Remove.
21774 (record_dep_loop): New function.
21775 (ref_indep_loop_p_1): Adjust to only walk over references
21776 in the loop, not its subloops.
21777 (ref_indep_loop_p): Rename to ...
21778 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21779 maintaining a more fine-grained cache.
21780 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21781 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21783 2013-03-22 Richard Biener <rguenther@suse.de>
21785 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21786 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21787 aggregate mem_ref_loc.
21788 (free_mem_ref_locs): Inline into ...
21789 (memref_free): ... this and adjust.
21790 (mem_ref_alloc): Adjust.
21791 (mem_ref_locs_alloc): Remove.
21792 (record_mem_ref_loc): Adjust.
21793 (get_all_locs_in_loop): Rewrite into ...
21794 (for_all_locs_in_loop): ... this iterator.
21795 (rewrite_mem_ref_loc): New functor.
21796 (rewrite_mem_refs): Use for_all_locs_in_loop.
21797 (sm_set_flag_if_changed): New functor.
21798 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21799 (ref_always_accessed): New functor.
21800 (ref_always_accessed_p): Use for_all_locs_in_loop.
21802 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21804 * tree-pass.h (PROP_gimple_lvec): New.
21805 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21806 (init_optimization_passes): Move pass_lower_vector.
21807 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21809 (pass_lower_vector): Provide PROP_gimple_lvec.
21810 (pass_lower_vector_ssa): Likewise.
21811 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21813 2013-03-21 Mark Wielaard <mjw@redhat.com>
21815 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21817 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21819 * config/i386/i386.md (*movdi_internal): Disparage slightly
21820 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21821 conversion alternatives.
21823 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21825 PR middle-end/48087
21826 * diagnostic.def (DK_WERROR): New kind.
21827 * diagnostic.h (werrorcount): Define.
21828 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21829 promoted to DK_ERROR, increment DK_WERROR counter instead of
21831 * toplev.c (toplev_main): Call print_ignored_options even if
21832 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21833 even if just werrorcount is non-zero.
21836 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21838 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21839 (string_cst_pool_decl): New function.
21840 (optimize_one_addr_into_implicit_ptr): New function.
21841 (resolve_addr_in_expr): Optimize DWARF location expression
21842 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21843 which doesn't live in memory, but has DW_AT_location or
21844 DW_AT_const_value, or refers to a string literal, into
21845 DW_OP_GNU_implicit_pointer.
21846 (optimize_location_into_implicit_ptr): New function.
21847 (resolve_addr): If removing DW_AT_location of a variable because
21848 it was DW_OP_addr of address of the variable, but the variable doesn't
21849 live in memory, try to emit const value attribute for the initializer.
21851 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21853 * tree.h (VECTOR_TYPE_P): New macro.
21854 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21855 TYPE_MODE): Use it.
21856 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21857 VEC_COND_EXPR cannot be lvalues.
21858 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21860 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21862 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21863 Restrict the transformation to equal modes.
21865 2013-03-21 Richard Biener <rguenther@suse.de>
21867 PR tree-optimization/39326
21868 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21869 (MEM_ANALYZABLE): Adjust.
21870 (record_mem_ref_loc): Move bitmap ops ...
21871 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21872 unanalyzable refs, do not record locations for it.
21873 (analyze_memory_references): Allocate ref zero as shared
21875 (refs_independent_p): Do not test for unanalyzed mems here.
21876 (ref_indep_loop_p_1): Special-case disambiguation against
21877 the unanalyzed ref.
21878 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
21880 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21882 * config/arm/arm-protos.h (tune_params): Add
21883 prefer_neon_for_64bits field.
21884 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21885 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21886 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21887 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21888 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21889 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21890 (arm_option_override): Handle -mneon-for-64bits new option.
21891 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21892 (prefer_neon_for_64bits): Declare new variable.
21893 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
21894 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
21895 (arch_enabled): Handle new arch types. Remove support for onlya8
21897 (one_cmpldi2): Use new arch names.
21898 (zero_extend<mode>di2, extend<mode>di2): Ditto.
21899 * config/arm/arm.opt (mneon-for-64bits): Add option.
21900 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21901 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21902 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21904 * doc/invoke.texi (-mneon-for-64bits): Document.
21906 2013-03-21 Richard Biener <rguenther@suse.de>
21908 PR tree-optimization/39326
21909 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21910 (sort_bbs_in_loop_postorder_cmp): New function.
21911 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
21913 2013-03-21 Richard Biener <rguenther@suse.de>
21915 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21916 (vect_insert_into_interleaving_chain): Likewise.
21917 (vect_drs_dependent_in_basic_block): Inline ...
21918 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21920 (vect_analyze_data_ref_dependence): ... here. Simplify.
21921 (vect_check_interleaving): Simplify.
21922 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21923 (vect_slp_analyze_data_ref_dependences): ... this new function.
21924 (dr_group_sort_cmp): New function.
21925 (vect_analyze_data_ref_accesses): Compute data-reference groups
21926 here instead of in vect_analyze_data_ref_dependence. Use
21927 a more efficient algorithm.
21928 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21929 vect_slp_analyze_data_ref_dependences. Call
21930 vect_analyze_data_ref_accesses earlier.
21931 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21932 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21933 (vect_slp_analyze_data_ref_dependences): New prototype.
21935 2013-03-21 Richard Biener <rguenther@suse.de>
21937 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21938 ref is stored in the loop.
21939 (find_refs_for_sm): Walk only over all stores.
21940 (store_motion_loop): Allocate from lim_bitmap_obstack.
21941 (store_motion): Likewise.
21943 2013-03-21 Richard Biener <rguenther@suse.de>
21945 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21946 Update virtual SSA form.
21948 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21950 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21951 * configure: Regenerate.
21952 * config.in: Regenerate.
21953 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21954 if !HAVE_LD_EH_FRAME_CIEV3.
21956 2013-03-21 Richard Biener <rguenther@suse.de>
21958 * tree-cfg.c (verify_expr_no_block): New function.
21959 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21960 nor DECL_VALUE_EXPR have locations with associated blocks.
21961 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21962 (clear_unused_block_pointer): Remove code dealing with
21963 blocks in DECL_DEBUG_EXPR locations.
21965 2013-03-21 Richard Biener <rguenther@suse.de>
21967 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21968 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21969 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21970 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21971 instead of DECL_DEBUG_EXPR_IS_FROM.
21972 * gimplify.c (gimplify_modify_expr): Likewise.
21973 * tree-cfg.c (verify_expr_location_1): Likewise.
21974 * tree-complex.c (create_one_component_var): Likewise.
21975 * tree-sra.c (create_access_replacement): Likewise.
21976 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21977 (clear_unused_block_pointer): Likewise.
21978 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21979 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
21980 * var-tracking.c (var_debug_decl): Likewise.
21981 (track_expr_p): Likewise.
21982 * tree-inline.c (add_local_variables): Likewise. Set
21983 DECL_HAS_DEBUG_EXPR_P after copying it.
21984 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
21985 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
21987 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21990 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
21991 * configure: Regenerate.
21992 * config.in: Regenerate.
21993 * config/i386/i386.md (*movdf_internal): Use
21994 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21995 movd instead of movq mnemonic for interunit moves.
21996 (*movdi_internal): Ditto.
21998 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22000 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22001 (abd<mode>_3): New pattern.
22002 (aba<mode>_3): New pattern.
22003 (fabd<mode>_3): New pattern.
22005 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22007 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22008 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22009 occurrence of REGISTER_PREFIX as its empty string.
22011 2013-03-20 Jeff Law <law@redhat.com>
22013 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22014 addititional equivalences for equality comparisons between an SSA_NAME
22015 and a constant where the SSA_NAME was set from a widening conversion.
22017 2013-03-20 Walter Lee <walt@tilera.com>
22019 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22021 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22023 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22024 depending on type attribute.
22025 (*movti_internal): Ditto.
22026 (*movtf_internal): Ditto.
22027 (*movxf_internal): Ditto.
22028 (*movdf_internal): Ditto.
22029 (*movsf_internal): Ditto.
22031 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22033 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22034 maybe_vex for sselog1 and ssemov types.
22035 (*movdi_internal): Reorder operand constraints.
22036 (*movsi_internal): Ditto. Set prefix attribute to
22037 maybe_vex for sselog1 and ssemov types.
22038 (*movtf_internal): Set prefix attribute to maybe_vex
22039 for sselog1 and ssemov types.
22040 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22041 DImode ssemov types. Reorder operand constraints.
22042 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22043 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22044 attribute for SImode ssemov types. Reorder operand constraints.
22046 2013-03-20 Martin Jambor <mjambor@suse.cz>
22048 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22049 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22051 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22053 * config/rs6000/predicates.md (indexed_address, update_address_mem
22054 update_indexed_address_mem): New predicates.
22055 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22056 attribute for load/store instructions.
22057 * config/rs6000/dfp.md (movsd_store): Likewise.
22058 (movsd_load): Likewise.
22059 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22060 (unnamed HI->DI extend define_insn): Likewise.
22061 (unnamed SI->DI extend define_insn): Likewise.
22062 (unnamed QI->SI extend define_insn): Likewise.
22063 (unnamed QI->HI extend define_insn): Likewise.
22064 (unnamed HI->SI extend define_insn): Likewise.
22065 (unnamed HI->SI extend define_insn): Likewise.
22066 (extendsfdf2_fpr): Likewise.
22067 (movsi_internal1): Likewise.
22068 (movsi_internal1_single): Likewise.
22069 (movhi_internal): Likewise.
22070 (movqi_internal): Likewise.
22071 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22072 attribute for load/store instructions.
22073 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22075 (mov<mode>_softfloat): Likewise.
22076 (mov<mode>_hardfloat32): Likewise.
22077 (mov<mode>_hardfloat64): Likewise.
22078 (mov<mode>_softfloat64): Likewise.
22079 (movdi_internal32): Likewise.
22080 (movdi_internal64): Likewise.
22081 (probe_stack_<mode>): Likewise.
22083 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22085 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22086 floating point, and decimal floating point to reload iterator.
22088 * config/rs6000/constraints.md (wl constraint): New constraints to
22089 return FLOAT_REGS if certain options are used to reduce the number
22090 of separate patterns that exist in the file.
22091 (wx constraint): Likewise.
22092 (wz constraint): Likewise.
22094 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22095 -mdebug=reg, print wg, wl, wx, and wz constraints.
22096 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22097 Initialize the reload functions for 64-bit binary/decimal floating
22099 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22100 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22101 create the buffer on the stack to overcome not having a 32-bit
22103 (rs6000_emit_move): Likewise.
22104 (rs6000_secondary_memory_needed_rtx): Likewise.
22105 (rs6000_alloc_sdmode_stack_slot): Likewise.
22106 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22107 via xxlxor, just like DFmode 0.0.
22109 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22110 define as 1 if we are running on a power7 or newer.
22111 (enum r6000_reg_class_enum): Add new constraints.
22113 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22114 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22115 with other moves by using conditional constraits (wg). Use LFIWZX
22116 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22117 (movsd splitter): Likewise.
22118 (movsd_hardfloat): Likewise.
22119 (movsd_softfloat): Likewise.
22121 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22122 binary and decimal floating point moves.
22123 (fmove_ok): New attributes to combine binary and decimal floating
22124 point moves, and to combine power6x (mfpgpr) moves along normal
22126 (real_value_to_target): Likewise.
22127 (f32_lr): Likewise.
22128 (f32_lm): Likewise.
22129 (f32_li): Likewise.
22130 (f32_sr): Likewise.
22131 (f32_sm): Likewise.
22132 (f32_si): Likewise.
22133 (movsf): Combine binary and decimal floating point moves. Combine
22134 power6x (mfpgpr) moves with other moves by using conditional
22135 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22136 (mov<mode> for SFmode/SDmode); Likewise.
22137 (SFmode/SDmode splitters): Likewise.
22138 (movsf_hardfloat): Likewise.
22139 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22140 (movsf_softfloat): Likewise.
22141 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22143 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22144 wx and wz constraints.
22146 * config/rs6000/constraints.md (wg constraint): New constraint to
22147 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22149 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22152 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22153 -mdebug=reg, print wg, wl, wx, and wz constraints.
22154 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22155 Initialize the reload functions for 64-bit binary/decimal floating
22157 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22158 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22159 create the buffer on the stack to overcome not having a 32-bit
22161 (rs6000_emit_move): Likewise.
22162 (rs6000_secondary_memory_needed_rtx): Likewise.
22163 (rs6000_alloc_sdmode_stack_slot): Likewise.
22164 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22165 via xxlxor, just like DFmode 0.0.
22167 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22168 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22169 with other moves by using conditional constraits (wg). Use LFIWZX
22170 and STFIWX for loading SDmode on power7.
22171 (movdd splitters): Likewise.
22172 (movdd_hardfloat32): Likewise.
22173 (movdd_softfloat32): Likewise.
22174 (movdd_hardfloat64_mfpgpr): Likewise.
22175 (movdd_hardfloat64): Likewise.
22176 (movdd_softfloat64): Likewise.
22178 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22179 64-bit binary and decimal floating point moves.
22180 (FMOVE64X): Likewise.
22181 (movdf): Combine 64-bit binary and decimal floating point moves.
22182 Combine power6x (mfpgpr) moves with other moves by using
22183 conditional constraits (wg).
22184 (mov<mode> for DFmode/DDmode): Likewise.
22185 (DFmode/DDmode splitters): Likewise.
22186 (movdf_hardfloat32): Likewise.
22187 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22188 (movdf_softfloat32): Likewise.
22189 (movdf_hardfloat64_mfpgpr): Likewise.
22190 (movdf_hardfloat64): Likewise.
22191 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22192 (movdf_softfloat64): Likewise.
22193 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22194 (reload_<mode>_load): Move to later in the file so they aren't in
22195 the middle of the floating point move insns.
22196 (reload_<mode>_store): Likewise.
22198 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22201 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22202 constraint if -mdebug=reg.
22203 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22204 Enable using dd reload support if needed.
22206 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22207 binary and decimal floating point moves in rs6000.md.
22208 (movtd_internal): Likewise.
22210 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22211 decimal floating point moves.
22213 (movtf_internal): Likewise.
22214 (mov<mode>_internal, TDmode/TFmode): Likewise.
22215 (movtf_softfloat): Likewise.
22216 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22218 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22219 movdi_internal64, using wg constraint for move direct operations.
22220 (movdi_internal64): Likewise.
22222 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22223 MODES_TIEABLE_P for selected modes. Print the numerical value of
22224 the various virtual registers. Use GPR/FPR first/last values,
22225 instead of hard coding the register numbers. Print which modes
22226 have reload functions registered.
22227 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22228 settings before/after setting cpu, target and subtarget settings.
22229 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22230 and for secondary reload failures in rs6000_secondary_reload_inner.
22231 (rs6000_secondary_reload_fail): Likewise.
22232 (rs6000_secondary_reload_inner): Likewise.
22234 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22235 macros for first/last GPR and FPR registers.
22236 (LAST_GPR_REGNO): Likewise.
22237 (FIRST_FPR_REGNO): Likewise.
22238 (LAST_FPR_REGNO): Likewise.
22240 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22241 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22242 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22243 (vcond<mode><mode>): Likewise.
22244 (vcondu<mode><mode>): Likewise.
22245 (vector_gtu<mode>): Likewise.
22246 (vector_gte<mode>): Likewise.
22247 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22248 to prevent the compiler from converting DImode operations to TImode.
22249 (ior<mode>3): Likewise.
22250 (and<mode>3): Likewise.
22251 (one_cmpl<mode>2): Likewise.
22252 (nor<mode>3): Likewise.
22253 (andc<mode>3): Likewise.
22255 * config/rs6000/constraints.md (wt constraint): New constraint
22256 that returns VSX_REGS if TImode is allowed in VSX registers.
22258 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22259 constant under VSX.
22261 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22262 similar to TImode, but it is restricted to being in the GPRs.
22264 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22265 TImode to occupy a single VSX register.
22267 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22268 -mvsx-timode for power7/power8.
22269 (power7 cpu): Likewise.
22270 (power8 cpu): Likewise.
22272 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22273 sure that TFmode/TDmode take up two registers if they are ever
22274 allowed in the upper VSX registers.
22275 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22277 (rs6000_init_hard_regno_mode_ok): Likewise.
22278 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22279 constraint. Print if LRA is turned on.
22280 (rs6000_option_override_internal): Give an error if -mvsx-timode
22281 and VSX is not enabled.
22282 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22283 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22284 to reg+offset addressing. Use PTImode when checking offset
22285 addresses for validity.
22286 (reg_offset_addressing_ok_p): Likewise.
22287 (rs6000_legitimate_offset_address_p): Likewise.
22288 (rs6000_legitimize_address): Likewise.
22289 (rs6000_legitimize_reload_address): Likewise.
22290 (rs6000_legitimate_address_p): Likewise.
22291 (rs6000_eliminate_indexed_memrefs): Likewise.
22292 (rs6000_emit_move): Likewise.
22293 (rs6000_secondary_reload): Likewise.
22294 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22295 reloads to fpr registers to continue to use reg+offset addressing,
22296 but 64-bit reloads to altivec registers need reg+reg addressing.
22297 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22298 it. Treat LO_SUM like a PLUS operation.
22299 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22300 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22301 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22302 registers to share a register with a smaller sized type, since VSX
22303 puts scalars in the upper 64-bits.
22304 (print_operand): Add support for PTImode.
22305 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22306 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22307 registers, but don't have arithmetic support.
22308 (rs6000_memory_move_cost): Add test for VSX.
22309 (rs6000_opt_masks): Add -mvsx-timode.
22311 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22314 (VSr): Use wt constraint for TImode.
22315 (VSv): Drop TImode support.
22316 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22317 (vsx_movti_64bit): Likewise.
22318 (vsx_movti_32bit): Likewise.
22319 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22320 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22321 one '?' on the appropriate output constraint. Do not allow TImode
22322 logical operations on 32-bit systems.
22323 (vsx_ior<mode>3): Likewise.
22324 (vsx_xor<mode>3): Likewise.
22325 (vsx_one_cmpl<mode>2): Likewise.
22326 (vsx_nor<mode>3): Likewise.
22327 (vsx_andc<mode>3): Likewise.
22328 (vsx_concat_<mode>): Likewise.
22329 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22331 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22332 OPTION_MASK_VSX_TIMODE.
22333 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22334 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22336 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22337 (TI2 iterator): New iterator for TImode, PTImode.
22338 (wd mode attribute): Add values for vector types.
22339 (movti_string): Replace TI move operations with operations for TImode
22340 and PTImode. Add support for TImode being allowed in VSX registers.
22341 (mov<mode>_string, TImode/PTImode): Likewise.
22342 (movti_ppc64): Likewise.
22343 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22344 (TI mode splitters): Likewise.
22346 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22349 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22351 PR tree-optimization/56355
22352 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22353 Also handle integers with undefined overflow.
22355 2013-03-20 Catherine Moore <clm@codesourcery.com>
22356 Maciej W. Rozycki <macro@codesourcery.com>
22357 Tom de Vries <tom@codesourcery.com>
22358 Nathan Sidwell <nathan@codesourcery.com>
22359 Iain Sandoe <iain@codesourcery.com>
22360 Nathan Froyd <froydnj@codesourcery.com>
22361 Chao-ying Fu <fu@mips.com>
22363 * doc/extend.texi: (micromips, nomicromips, nocompression):
22364 Document new function attributes.
22365 * doc/invoke.texi (minterlink-compressed, mmicromips,
22366 m14k, m14ke, m14kec): Document new options.
22367 (minterlink-mips16): Update documentation.
22368 * doc/md.texi (ZC, ZD): Document new constraints.
22369 * configure.ac (gcc_cv_as_micromips): Check if linker
22370 supports the .set micromips directive.
22371 * configure: Regenerate.
22372 * config.in: Regenerate.
22373 * config/mips/mips-tables.opt: Regenerate.
22374 * config/mips/micromips.md: New file.
22375 * constraints.md (ZC, ZD): New constraints.
22376 * config/mips/predicates.md (movep_src_register): New predicate.
22377 (movep_src_operand): New predicate.
22378 (non_volatile_mem_operand): New predicate.
22379 * config/mips/mips.md (multimem): New type.
22380 (length): Differentiate between 17-bit and 18-bit branch offsets.
22381 (MOVEP1, MOVEP2): New mode iterator.
22382 (mov_<load>l): Use ZC constraint.
22383 (mov_<load>r): Likewise.
22384 (mov_<store>l): Likewise.
22385 (mov_<store>r): Likewise.
22386 (*branch_equality<mode>_inverted): Add microMIPS support.
22387 (*branch_equality<mode>): Likewise.
22388 (*jump_absolute): Likewise.
22389 (indirect_jump_<mode>): Likewise.
22390 (tablejump_<mode>): Likewise.
22391 (<optab>_internal): Likewise.
22392 (sibcall_internal): Likewise.
22393 (sibcall_value_internal): Likewise.
22394 (prefetch): Use constraint ZD.
22395 * config/mips/mips.opt (minterlink-compressed): New option.
22396 (minterlink-mips16): Now an alias for minterlink-compressed.
22397 (mmicromips): New option.
22398 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22399 (compare_and_swap_12): Likewise.
22400 (sync_add<mode>): Likewise.
22401 (sync_<optab>_12): Likewise.
22402 (sync_old_<optab>_12): Likewise.
22403 (sync_new_<optab>_12): Likewise.
22404 (sync_nand_12): Likewise.
22405 (sync_old_nand_12): Likewise.
22406 (sync_new_nand_12): Likewise.
22407 (sync_sub<mode>): Likewise.
22408 (sync_old_add<mode>): Likewise.
22409 (sync_old_sub<mode>): Likewise.
22410 (sync_new_add<mode>): Likewise.
22411 (sync_new_sub<mode>): Likewise.
22412 (sync_<optab><mode>): Likewise.
22413 (sync_old_<optab><mode>): Likewise.
22414 (sync_new_<optab><mode>): Likewise.
22415 (sync_nand<mode>): Likewise.
22416 (sync_old_nand<mode>): Likewise.
22417 (sync_new_nand<mode>): Likewise.
22418 (sync_lock_test_and_set<mode>): Likewise.
22419 (test_and_set_12): Likewise.
22420 (atomic_compare_and_swap<mode>): Likewise.
22421 (atomic_exchange<mode>_llsc): Likewise.
22422 (atomic_fetch_add<mode>_llsc): Likewise.
22423 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22424 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22425 (umips_save_restore_pattern_p): Likewise.
22426 (umips_load_store_pair_p): Likewise.
22427 (umips_output_load_store_pair): Likewise.
22428 (umips_movep_target_p): Likewise.
22429 (umips_12bit_offset_address_p): Likewise.
22430 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22431 (mips_base_mips16): Rename this...
22432 (mips_base_compression_flags): ...to this. Update all uses.
22433 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22434 (mips_mips16_decl_p): Delete.
22435 (mips_nomips16_decl_p): Delete.
22436 (mips_get_compress_on_flags): New function.
22437 (mips_get_compress_off_flags): New function.
22438 (mips_get_compress_mode): New function.
22439 (mips_get_compress_on_name): New function.
22440 (mips_get_compress_off_name): New function.
22441 (mips_insert_attributes): Support multiple compression types.
22442 (mips_merge_decl_attributes): Likewise.
22443 (umips_12bit_offset_address_p): New function.
22444 (mips_start_function_definition): Emit .set micromips directive.
22445 (mips_call_may_need_jalx_p): New function.
22446 (mips_function_ok_for_sibcall): Add microMIPS support.
22447 (mips_print_operand_punctuation): Support short delay slots and
22449 (umips_swm_mask, umips_swm_encoding): New.
22450 (umips_build_save_restore): New function.
22451 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22452 (was_mips16_p): Remove.
22453 (old_compression_mode): New.
22454 (mips_set_compression_mode): New function.
22455 (mips_set_current_function): Add microMIPS support.
22456 (mips_option_override): Likewise.
22457 (umips_save_restore_pattern_p): New function.
22458 (umips_output_save_restore): New function.
22459 (umips_load_store_pair_p_1): New function.
22460 (umips_load_store_pair_p): New function.
22461 (umips_output_load_store_pair_1): New function.
22462 (umips_output_load_store_pair): New function.
22463 (umips_movep_target_p) New function.
22464 (mips_prepare_pch_save): Add microMIPS support.
22465 * config/mips/mips.h (TARGET_COMPRESSION): New.
22466 (TARGET_CPU_CPP_BUILTINS): Update macro
22467 to use new compression flags and to support microMIPS.
22468 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22469 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22470 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22471 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22472 (ASM_SPEC): Support mmicromips and mno-micromips.
22473 (M16STORE_REG_P): New macro.
22474 (MIPS_CALL): Support TARGET_MICROMIPS.
22475 (MICROMIPS_J): New macro.
22476 (mips_base_mips16): Rename this...
22477 (mips_base_compression_flags): ...to this.
22478 (UMIPS_12BIT_OFFSET_P): New macro.
22479 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22480 (MULTILIB_DIRNAMES): Likewise.
22481 2013-03-20 Richard Biener <rguenther@suse.de>
22483 PR tree-optimization/56661
22484 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22485 the result does not have to be distinct.
22487 2013-03-20 Richard Biener <rguenther@suse.de>
22489 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22492 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22493 Steven Bosscher <steven@gcc.gnu.org>
22495 PR rtl-optimization/56605
22496 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22498 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22501 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22502 that require movd instead of movq.
22504 2013-03-20 Richard Biener <rguenther@suse.de>
22506 * tree-ssa-structalias.c (struct variable_info): Add pointer
22507 to the first field of an aggregate with sub-vars. Make
22508 this and the pointer to the next subfield its ID.
22509 (vi_next): New function.
22510 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22511 storedanything_id, integer_id): Increment by one.
22512 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22513 get_call_clobber_vi): Adjust.
22514 (solution_set_expand): Simplify and speedup.
22515 (solution_set_add): Inline into ...
22516 (set_union_with_increment): ... this. Adjust accordingly.
22517 (do_sd_constraint): Likewise.
22518 (do_ds_constraint): Likewise.
22519 (do_complex_constraint): Simplify.
22520 (build_pred_graph): Adjust.
22521 (solve_graph): Likewise. Simplify and speedup.
22522 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22523 get_constraint_for_component_ref, get_constraint_for_1,
22524 first_vi_for_offset, first_or_preceding_vi_for_offset,
22525 create_function_info_for, create_variable_info_for_1,
22526 create_variable_info_for, intra_create_variable_infos): Adjust.
22527 (init_base_vars): Push NULL for ID zero.
22528 (compute_points_to_sets): Adjust.
22530 2013-03-20 Richard Biener <rguenther@suse.de>
22532 * cfgloop.c (verify_loop_structure): Streamline and avoid
22533 ICEing on corrupt loop tree.
22534 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22537 2013-03-20 Richard Biener <rguenther@suse.de>
22539 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22540 check whether an SSA update is needed.
22542 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22544 * config/mips/constraints.md (T): Rename to...
22548 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22549 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22551 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22553 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22554 (*subsi3_carryin_uxtw): Likewise.
22556 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22558 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22559 (*rorsi3_insn_uxtw): Likewise.
22561 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22563 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22564 (*extrsi5_insn_uxtw): Likewise.
22566 2013-03-19 Richard Biener <rguenther@suse.de>
22568 PR tree-optimization/56273
22569 * passes.c (init_optimization_passes): Move second VRP after DOM.
22571 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22573 * config/i386/i386.md (*movti_internal): Merge from
22574 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22575 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22576 nox64 isa attributes.
22578 2013-03-18 Richard Biener <rguenther@suse.de>
22580 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22582 (merge_node_constraints): Likewise.
22583 (build_succ_graph): Likewise.
22584 (valid_graph_edge): Inline into single caller.
22585 (unify_nodes): Likewise. Use bitmap_set_bit return value
22587 (scc_visit): Fix formatting and variable use.
22588 (do_sd_constraint): Use gcc_checking_assert.
22589 (do_ds_constraint): Likewise.
22590 (do_complex_constraint): Likewise.
22591 (condense_visit): Likewise. Cleanup.
22592 (dump_pred_graph): New function.
22593 (perform_var_substitution): Dump the pred-graph before
22594 variable substitution.
22595 (find_equivalent_node): Use gcc_checking_assert.
22596 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22598 2013-03-18 Richard Biener <rguenther@suse.de>
22600 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22601 Remove cond_expr_stmt_list argument and do not gimplify the
22603 (vect_loop_versioning): Adjust.
22604 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22605 Cleanup to use less temporaries.
22606 (vect_create_data_ref_ptr): Cleanup.
22608 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22610 PR tree-optimization/56635
22611 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22612 require types_compatible_p types.
22614 2013-03-18 Nick Clifton <nickc@redhat.com>
22616 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22617 spurious backslash.
22619 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22620 Add missing line to comment describing function.
22622 2013-03-18 Richard Biener <rguenther@suse.de>
22624 PR tree-optimization/56210
22625 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22626 Handle string / character search functions.
22627 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22629 2013-03-18 Richard Biener <rguenther@suse.de>
22631 PR middle-end/56483
22632 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22633 and implement properly.
22634 * gimple.h (gimple_cond_single_var_p): Remove.
22636 2013-03-18 Richard Biener <rguenther@suse.de>
22638 * tree-data-ref.h (find_data_references_in_loop): Declare.
22639 * tree-data-ref.c (get_references_in_stmt): Use a stack
22640 vector pre-allocated in the callers.
22641 (find_data_references_in_stmt): Adjust.
22642 (graphite_find_data_references_in_stmt): Likewise.
22643 (create_rdg_vertices): Likewise.
22644 (find_data_references_in_loop): Export.
22645 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22646 Compute dependences here...
22647 (vect_analyze_data_refs): ...not here. When we encounter
22648 a non-vectorizable data reference in basic-block vectorization
22649 truncate the data reference vector. Do not bother to
22650 fixup data-dependence information for gather loads.
22651 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22652 of data references, as reported.
22654 2013-03-18 Richard Biener <rguenther@suse.de>
22656 PR tree-optimization/3713
22657 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22658 has_constants and expr.
22659 (stmt_has_constants): Properly valueize SSA names when deciding
22660 whether the stmt has constants.
22662 2013-03-18 Richard Biener <rguenther@suse.de>
22664 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22665 whole function when there is nothing to do.
22666 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22667 * tree-vectorizer.c (vectorize_loops): Update virtual and
22668 loop-closed SSA once.
22669 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22671 2013-03-18 Richard Biener <rguenther@suse.de>
22673 PR middle-end/56113
22674 * domwalk.c (bb_postorder): New global static.
22675 (cmp_bb_postorder): New function.
22676 (walk_dominator_tree): Replace scheme imposing an order for
22677 visiting dominator sons by one sorting them at the time they
22678 are pushed on the stack.
22680 2013-03-18 Richard Biener <rguenther@suse.de>
22682 PR tree-optimization/39326
22683 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22684 (struct mem_ref): Replace mem member with ao_ref typed member.
22685 (MEM_ANALYZABLE): Adjust.
22686 (memref_eq): Likewise.
22687 (mem_ref_alloc): Likewise.
22688 (gather_mem_refs_stmt): Likewise.
22689 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22690 (execute_sm_if_changed_flag_set): Adjust.
22691 (execute_sm): Likewise.
22692 (ref_always_accessed_p): Likewise.
22693 (refs_independent_p): Likewise.
22694 (can_sm_ref_p): Likewise.
22696 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22699 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22700 return 1 even for !unsignedp.
22702 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22704 * config/i386/i386.md (isa): Add x64 and nox64.
22705 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22706 (*pushtf): Enable *roF alternative for x64 isa only.
22707 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22708 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22709 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22710 (*movtf_internal): Merge from *movtf_internal_rex64 and
22711 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22712 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22713 nox64 isa attributes.
22714 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22715 nox64 isa attributes.
22716 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22718 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22720 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22722 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22725 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22726 class names. Remove trailing comma after "ALL_REGS".
22728 2013-03-16 Jan Hubicka <jh@suse.cz>
22730 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22731 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22732 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22733 of cgraph_get_create_node.
22734 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22736 2013-03-16 Jason Merrill <jason@redhat.com>
22739 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22740 with DW_AT_default_value.
22742 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22744 * BASE-VER: Set to 4.9.0.
22746 2013-03-14 Andi Kleen <ak@linux.intel.com>
22749 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22750 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22751 Document _x* TSX intrinsics.
22753 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22754 David Holsgrove <david.holsgrove@xilinx.com>
22756 * configure.ac: Add MicroBlaze TLS support detection.
22757 * configure: Regenerate.
22758 * config/microblaze/microblaze-protos.h
22759 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22760 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22761 * config/microblaze/microblaze.c (microblaze_address_type): Add
22762 ADDRESS_TLS and tls_reloc address types.
22763 (microblaze_address_info): Add tls_reloc.
22764 (TARGET_HAVE_TLS): Define.
22765 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22766 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22767 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22768 load_tls_operand, microblaze_call_tls_get_addr,
22769 microblaze_legitimize_tls_address): New functions.
22770 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22771 (get_base_reg): Use microblaze_tls_symbol_p.
22772 (microblaze_classify_address): Handle TLS.
22773 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22774 label_mentioned_p and microblaze_tls_referenced_p.
22775 (microblaze_legitimize_address): Handle TLS.
22776 (microblaze_address_insns): Handle ADDRESS_TLS.
22777 (pic_address_needs_scratch): Handle TLS.
22778 (print_operand_address): Handle TLS.
22779 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22780 (microblaze_expand_move): Handle TLS.
22781 (microblaze_legitimate_constant_p): Check
22782 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22783 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22784 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22785 (PIC_OFFSET_TABLE_REGNUM): Set.
22786 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22787 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22788 (addsi3, movsi_internal2, movdf_internal): Update constraints
22789 * config/microblaze/predicates.md (arith_plus_operand): Define
22790 (move_operand): Redefine as move_src_operand,
22791 check microblaze_tls_referenced_p.
22793 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22795 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22796 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22798 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22800 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22803 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22805 PR tree-optimization/53265
22806 * common.opt (Waggressive-loop-optimizations): New option.
22807 * tree-ssa-loop-niter.c: Include tree-pass.h.
22808 (do_warn_aggressive_loop_optimizations): New function.
22809 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22810 if number_of_latch_executions returned constant.
22811 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22812 early. If number_of_latch_executions returned constant, set
22813 nb_iterations_upper_bound back to it.
22814 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22816 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22817 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22819 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22820 (MULTILIB_OSDIRNAMES): Set.
22821 * genmultilib: If defaultosdirname doesn't start with :: , set
22822 defaultosdirname2 instead, clear it and emit two . multilib_raw
22823 entries instead of just one.
22825 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22827 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22828 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22829 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22830 (SUBTARGET_OVERRIDE_OPTIONS): New.
22832 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22835 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22836 (musermode): Convert to Var(TARGET_USERMODE).
22837 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22838 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22839 * config/sh/sh.c (sh_option_override): Use
22840 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22841 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22843 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22845 (udivsi3_i4_single, divsi3_i4_single): Use
22846 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22848 2013-03-13 Dave Korn <dave.korn.cygwin@....>
22850 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22853 2013-03-13 Richard Biener <rguenther@suse.de>
22855 PR tree-optimization/56608
22856 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22857 calls when vectorizing basic-blocks.
22859 2013-03-13 Jakub Jelinek <jakub@redhat.com>
22862 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22865 2013-03-12 Jakub Jelinek <jakub@redhat.com>
22867 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22869 2013-03-11 Jan Hubicka <jh@suse.cz>
22872 * lto-streamer-out.c (output_symbol_p): Skip references from
22873 constructors of external variables.
22875 2013-03-11 Jan Hubicka <jh@suse.cz>
22877 PR middle-end/56571
22878 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22880 * emit-rtl.c (verify_rtx_sharing): Likewise.
22881 (copy_insn_1): Likewise.
22882 * rtl.c (copy_rtx): Likewise.
22884 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
22887 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22888 output_operand_lossage message.
22890 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
22893 * arm.c (shift_op): Validate RTL pattern on the fly.
22894 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22897 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22900 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22901 new call patterns that clobber %fr12.
22902 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22903 split and postreload patterns.
22904 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22905 registers %fr12 and %fr12R as call used.
22907 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22909 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22910 (canon_address, record_store, replace_read, check_mem_read_rtx,
22911 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22912 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22913 rest_of_handle_dse): Likewise.
22915 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22917 PR middle-end/56524
22918 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22920 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22921 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22922 (save_optabs_if_changed): Replace with...
22923 (init_tree_optimization_optabs): ...this.
22924 * optabs.c (save_optabs_if_changed): Rename to...
22925 (init_tree_optimization_optabs): ...this. Take the optimization node
22926 as argument. Do nothing if the base optabs are already correct.
22927 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22928 to recompute optabs.
22929 * function.h (function): Remove optabs field.
22930 * function.c (invoke_set_current_function_hook): Call
22931 init_tree_optimization_optabs. Use the result to initialize
22934 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
22936 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22937 if GTMA_HAS_NO_INSTRUMENTATION.
22938 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22939 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22940 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22941 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22942 GTMA_HAS_NO_INSTRUMENTATION.
22944 2013-03-08 Jakub Jelinek <jakub@redhat.com>
22946 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22949 2013-03-08 Marek Polacek <polacek@redhat.com>
22950 Jakub Jelinek <jakub@redhat.com>
22952 PR tree-optimization/56478
22953 * predict.c (is_comparison_with_loop_invariant_p): Change the
22954 type of loop_step to tree.
22955 (predict_loops): Adjust.
22956 (predict_iv_comparison): Perform the computations on double_ints.
22958 2013-03-08 Richard Biener <rguenther@suse.de>
22960 PR tree-optimization/56570
22961 * tree-cfg.c (verify_expr_location_1): Verify locations for
22963 * tree-sra.c (create_access_replacement): Strip locations
22964 from DECL_DEBUG_EXPRs.
22966 2013-03-08 Richard Biener <rguenther@suse.de>
22968 * tree-inline.c (expand_call_inline): Do not associate
22969 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22970 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22972 2013-03-08 Richard Biener <rguenther@suse.de>
22974 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22975 or block changes with -Og. Fix for location / block encoding
22976 changes and PHI arguments with locations.
22978 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
22980 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
22982 (struct output_info): Likewise.
22983 (register_overhead): Remove bad gcc_assert.
22984 (bitmap_find_bit): If there is only a single bitmap element, do not
22985 count a miss as a search.
22986 (print_statistics): Update for counter type changes.
22987 (dump_bitmap_statistics): Likewise. Print headers such that they
22988 are properly lined up with the printed counters.
22990 2013-03-07 Jakub Jelinek <jakub@redhat.com>
22992 PR tree-optimization/56559
22993 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
22994 check that it has only a single use.
22996 2013-03-07 Richard Biener <rguenther@suse.de>
22998 * doc/invoke.texi (fwhole-program): Discourage use in combination
23001 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23003 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23005 PR tree-optimization/56539
23006 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23007 instead of GSI_CONTINUE_LINKING as last argument to
23008 force_gimple_operand_gsi. Adjust function comment.
23010 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23013 PR middle-end/56548
23014 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23015 promoted mode, convert the result back to the original mode.
23017 2013-03-06 Richard Biener <rguenther@suse.de>
23019 PR middle-end/56294
23020 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23021 (insert_updated_phi_nodes_compare_uids): New function.
23022 (update_ssa): Sort symbols_to_rename after UID before
23023 traversing it to insert PHI nodes.
23025 2013-03-06 Richard Biener <rguenther@suse.de>
23027 PR middle-end/50494
23028 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23029 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23032 2013-02-13 Richard Biener <rguenther@suse.de>
23035 * varasm.c (output_constant_def_1): Get the decl representing
23036 the constant as argument.
23037 (output_constant_def): Wrap output_constant_def_1.
23038 (make_decl_rtl): Use output_constant_def_1 with the decl
23039 representing the constant.
23040 (build_constant_desc): Optionally re-use a decl already
23041 representing the constant.
23042 (tree_output_constant_def): Adjust.
23044 2013-03-06 Joey Ye <joey.ye@arm.com>
23047 * gcc.c (convert_white_space): New function.
23048 (main): Handles white space in function name.
23050 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23053 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23054 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23055 to SH_DIV_CALL_TABLE for TARGET_SH2.
23056 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23058 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23059 call-table options.
23061 2013-03-05 Sterling Augustine <saugustine@google.com>
23062 Cary Coutant <ccoutant@google.com>
23065 * dwarf2out.c (resolve_addr): Don't call
23066 remove_loc_list_addr_table_entries a second time for the same
23069 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23072 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23073 (avoid_complex_debug_insns): New function.
23074 (expand_debug_locations): Call it.
23076 PR rtl-optimization/56484
23077 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23078 lifetimes of hard registers on small register class machines.
23080 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23082 * config/microblaze/microblaze-protos.h: Rename
23083 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23084 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23086 (microblaze_fast_interrupt_function_p): New function.
23087 (microblaze_is_interrupt_handler): Rename to
23088 microblaze_is_interrupt_variant and add fast_interrupt check.
23089 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23090 (save_restore_insns): Likewise.
23091 (compute_frame_size): Likewise.
23092 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23093 (microblaze_globalize_label): Likewise.
23094 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23095 * config/microblaze/microblaze.md: Use wrapper
23096 microblaze_is_interrupt_variant.
23098 2013-03-05 Kai Tietz <ktietz@redhat.com>
23100 * sdbout.c (sdbout_one_type): Switch to current function's section
23101 supporting cold/hot.
23103 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23105 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23108 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23110 PR middle-end/56461
23111 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23112 if VALGRIND_GET_VBITS is defined, temporarily make object
23113 memory all defined, and restore previous valgrind addressability
23114 and definability afterwards. Free this_object at the end.
23116 PR middle-end/56461
23117 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23118 right before calling lra_create_live_ranges, also call it
23119 when clearing live_p. Only call lra_clear_live_ranges
23120 at the end if live_p.
23122 PR middle-end/56461
23123 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23125 2013-03-05 Richard Biener <rguenther@suse.de>
23127 PR tree-optimization/56521
23128 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23131 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23134 * except.h (remove_unreachable_eh_regions): New prototype.
23135 * except.c (remove_eh_handler_splicer): New function, split out
23136 of remove_eh_handler.
23137 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23138 warning about running it on many EH regions one at a time.
23139 (remove_unreachable_eh_regions_worker): New function, walk the
23140 EH tree in depth-first order and remove non-marked regions.
23141 (remove_unreachable_eh_regions): New function.
23142 * tree-eh.c (mark_reachable_handlers): New function, split out
23143 from remove_unreachable_handlers.
23144 (remove_unreachable_handlers): Use mark_reachable_handlers and
23145 remove_unreachable_eh_regions.
23146 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23147 and remove_unreachable_eh_regions.
23149 2013-03-05 Richard Biener <rguenther@suse.de>
23151 PR middle-end/56525
23152 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23153 not freeing them until the end.
23155 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23157 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23159 2013-03-05 Richard Biener <rguenther@suse.de>
23161 PR tree-optimization/56270
23162 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23163 of loads after scheduling an SLP instance.
23165 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23167 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23169 (check_gcc_parallelize): Run guality.exp as a separate job from
23170 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23171 struct-layout-1.exp with stackalign.exp.
23173 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23175 PR middle-end/56461
23176 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23177 load_index sbitmap even if some bit in it isn't set.
23179 PR middle-end/56461
23180 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23181 (discover_iteration_bound_by_body_walk): Change queues to
23182 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23183 spelling in comment. Call safe_push on queues[bound_index] directly.
23184 Release queues[queue_index] in every iteration unconditionally.
23185 Release bounds vector.
23187 PR middle-end/56461
23188 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23189 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23190 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23193 2013-03-05 Richard Biener <rguenther@suse.de>
23196 * tree-inline.c (remap_blocks_to_null): New function.
23197 (expand_call_inline): When expanding a call stmt without
23198 an associated block inline remap all callee blocks to NULL.
23200 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23202 PR rtl-optimization/56494
23203 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23204 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23205 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23207 PR middle-end/56461
23208 * sel-sched-ir.c (free_sched_pools): Release
23209 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23210 if succs_info_pool.max_top isn't -1.
23213 * opts.c (opts_obstack, opts_concat): Moved to...
23214 * opts-common.c (opts_obstack, opts_concat): ... here.
23216 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23218 PR middle-end/56461
23219 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23221 2013-03-04 Martin Jambor <mjambor@suse.cz>
23223 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23224 all appropriate places.
23226 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23228 PR tree-optimization/56424
23229 * ipa-split.c (split_function): Do not set the RSO flag if result is
23230 not by reference and its type is a register type.
23232 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23234 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23235 (microblaze_legitimate_pic_operand): Likewise
23236 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23237 new function microblaze_legitimate_pic_operand
23238 * config/microblaze/microblaze-protos.h
23239 (microblaze_legitimate_pic_operand): Declare.
23241 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23243 * config/microblaze/predicates.md (call_insn_simple_operand):
23244 New predicate for supported rtx code types.
23245 * config/microblaze/microblaze.md (call_internal1): Use
23246 call_insn_simple_operand predicate.
23248 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23250 PR middle-end/56461
23251 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23252 partitions.ordered_remove.
23254 PR middle-end/56461
23255 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23256 vec_oprnds0.create (1) for modifier == NONE.
23258 PR middle-end/56461
23259 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23260 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23261 vec_oprnds1 right before pushing anything to it for
23264 PR middle-end/56461
23265 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23266 set nbbs to 0 instead of having separate code path.
23267 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23268 instead of false as last argument if returning NULL.
23270 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23272 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23273 the attribute is now called "target" instead of "option".
23274 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23275 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23276 attribute/pragma name for TARGET_OPTION_VALID_P and
23277 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23278 * doc/tm.texi: Regenerated.
23280 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23282 * config/microblaze/microblaze.c:
23283 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23284 * config/microblaze/microblaze.h: Add -mxl-reorder to
23286 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23287 instructions emitted if TARGET_REORDER.
23288 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23289 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23292 2013-03-01 Xinliang David Li <davidxl@google.com>
23294 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23297 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23299 PR middle-end/56461
23300 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23301 vector even when returning true. Fix up function comment formatting.
23303 PR middle-end/56461
23304 * ira-build.c (ira_loop_nodes_count): New variable.
23305 (create_loop_tree_nodes): Initialize it.
23306 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23308 PR middle-end/56461
23309 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23310 method on dr_chain and result_chain.
23311 * tree-vect-stmts.c (vectorizable_store): Only call
23312 result_chain.create if j == 0.
23314 PR middle-end/56461
23315 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23316 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23317 before overwriting it.
23319 2013-03-01 Tobias Burnus <burnus@net-b.de>
23321 * doc/extended.texi (C Extensions): Change order in @menu
23323 (Other MIPS Built-in Functions): Move last MIPS entry before
23324 "picoChip Built-in Functions".
23325 (SH Built-in Functions): Move after RX Built-in Functions.
23326 * doc/gcc.texi (Introduction): Change order in @menu
23328 * doc/md.texi (Constraints): Ditto.
23329 * gty.texi (Type Information): Ditto.
23330 (User-provided marking routines for template types): Make
23332 * doc/invoke.texi (AArch64 Options): Move before
23333 "Adapteva Epiphany Options".
23335 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23336 Jakub Jelinek <jakub@redhat.com>
23339 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23340 no_address_safety_analysis attribute.
23341 * doc/extend.texi (no_address_safety_attribute): Rename to
23342 no_sanitize_address attribute, mention no_address_safety_analysis
23343 attribute as deprecated alias.
23345 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23347 PR middle-end/56461
23348 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23349 type to vec<vec<tree> > *.
23350 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23351 to be vec<tree> instead of vec<tree> *, set vec_defs
23352 to vNULL and call vec_defs.create (number_of_vects), adjust other
23354 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23355 vectorizable_condition): Adjust vect_get_slp_defs callers.
23357 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23359 * config/aarch64/aarch64.c
23360 (aarch64_float_const_representable): Remove unused variable.
23362 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23364 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23366 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23368 * config/aarch64/aarch64-builtins.c
23369 (aarch64_init_simd_builtins): Make static.
23371 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23373 * config/aarch64/aarch64.c
23374 (aarch64_simd_make_constant): Make static.
23376 2013-02-28 Martin Jambor <mjambor@suse.cz>
23378 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23379 with no initialization to the RHS of debug statements.
23381 2013-02-28 Martin Jambor <mjambor@suse.cz>
23383 PR tree-optimization/56294
23384 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23386 (get_access_replacement): Do not call create_access_replacement.
23387 Assert a replacement exists.
23388 (get_repl_default_def_ssa_name): Create the replacement declaration
23391 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23393 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23394 final_end_function.
23396 2013-02-28 Marek Polacek <polacek@redhat.com>
23398 PR rtl-optimization/56466
23399 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23400 if we're changing a loop.
23401 (peel_loops_completely): Likewise.
23403 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23406 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23408 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23411 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23412 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23413 INTX_FTYPE_FX, FX_FTYPE_INTX.
23414 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23416 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23418 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23419 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23420 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23421 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23422 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23423 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23424 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23425 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23426 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23427 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23428 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23429 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23430 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23431 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23432 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23433 (avrxmega6): Increase max flash segments from 5 to 6.
23434 * config/avr/t-multilib: Regenerate.
23435 * config/avr/avr-tables.opt: Regenerate.
23436 * doc/avr-mmcu.texi: Regenerate.
23438 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23440 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23441 (avr_device_to_arch): Rename to avr_device_to_ld.
23442 (avr_device_to_as): New prototype.
23443 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23444 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23445 * config/avr/driver-avr.c (avr_device_to_as): New.
23446 (avr_device_to_arch): Rename to avr_device_to_ld.
23448 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23450 PR middle-end/56461
23451 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23452 method on dr_chain and result_chain.
23454 PR middle-end/56461
23455 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23456 pointer_set_destroy on not_executed_last_iteration.
23458 PR middle-end/56461
23459 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23461 PR middle-end/56461
23462 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23463 FOR_EACH_DEFINED_FUNCTION when freeing state.
23465 PR middle-end/56461
23466 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23468 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23471 PR middle-end/56461
23472 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23473 known_aggs[i].items and release known_aggs vector.
23475 PR middle-end/56461
23476 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23478 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23480 * config/microblaze/microblaze.c (microblaze_emit_compare):
23481 Use xor for EQ/NE comparisions.
23482 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23483 (cbranchsf4): Adjust operator to comparison_operator.
23485 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23487 PR middle-end/56461
23488 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23490 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23491 vec_safe_push, always update *slot.
23492 (redirect_edge_var_map_clear): Use vec_free.
23493 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23494 (free_var_map_entry): Use vec_free.
23495 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23496 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23498 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23500 PR middle-end/45472
23501 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23502 when the may_trap_p bit of the exprs being merged differs.
23503 Reorder tests for speculativeness in the logical and operator.
23505 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23507 * incpath.c (add_standard_paths): Use reconcat instead of concat
23508 where appropriate and avoid leaking memory.
23510 * opts.h: Include obstack.h.
23511 (opts_concat): New prototype.
23512 (opts_obstack): New declaration.
23513 * opts.c (opts_concat): New function.
23514 (opts_obstack): New variable.
23515 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23516 (finish_options): Use opts_concat instead of concat
23517 and XOBNEWVEC instead of XNEWVEC.
23518 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23519 generate_option): Likewise.
23520 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23521 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23524 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23525 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23527 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23529 PR middle-end/56461
23530 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23532 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23534 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23535 (arm_block_move_unaligned_straight): Likewise.
23536 (arm_adjust_block_mem): Likewise.
23538 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23541 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23542 temp, cond and label.
23543 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23546 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23547 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23550 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23551 of prologue/epilogue functions.
23554 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23555 Remove unused variable cfa_offset.
23556 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23559 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23560 type promotion to unsigned.
23563 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23564 for HOST_WIDE_INT of 32 bit / same size as int.
23565 (arm_block_move_unaligned_straight): Likewise.
23566 (arm_adjust_block_mem): Likewise.
23569 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23572 2013-02-26 Marek Polacek <polacek@redhat.com>
23574 PR tree-optimization/56426
23575 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23577 2013-02-26 Richard Biener <rguenther@suse.de>
23580 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23581 unused variable loops.
23583 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23585 PR tree-optimization/56448
23586 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23587 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23588 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23589 later operands of the references, or even first operand for
23590 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23592 PR tree-optimization/56443
23593 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23594 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23595 to type_for_mode langhook.
23597 2013-02-25 Matt Turner <mattst88@gmail.com>
23599 * doc/invoke.texi: Document r4700.
23601 2013-02-25 Richard Biener <rguenther@suse.de>
23603 PR tree-optimization/56175
23604 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23606 (simplify_bitwise_binary): ... here. Also guard the conversion
23607 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23609 2013-02-25 Catherine Moore <clm@codesourcery.com>
23612 2013-02-24 Catherine Moore <clm@codesourcery.com>
23613 Maciej W. Rozycki <macro@codesourcery.com>
23614 Tom de Vries <tom@codesourcery.com>
23615 Nathan Sidwell <nathan@codesourcery.com>
23616 Iain Sandoe <iain@codesourcery.com>
23617 Nathan Froyd <froydnj@codesourcery.com>
23618 Chao-ying Fu <fu@mips.com>
23620 * doc/extend.texi: (micromips, nomicromips, nocompression):
23621 Document new function attributes.
23622 * doc/invoke.texi (minterlink-compressed, mmicromips,
23623 m14k, m14ke, m14kec): Document new options.
23624 (minterlink-mips16): Update documentation.
23625 * doc/md.texi (ZC, ZD): Document new constraints.
23626 * configure.ac (gcc_cv_as_micromips): Check if linker
23627 supports the .set micromips directive.
23628 * configure: Regenerate.
23629 * config.in: Regenerate.
23630 * config/mips/mips-tables.opt: Regenerate.
23631 * config/mips/micromips.md: New file.
23632 * constraints.md (ZC, AD): New constraints.
23633 * config/mips/predicates.md (movep_src_register): New predicate.
23634 (movep_src_operand): New predicate.
23635 (non_volatile_mem_operand): New predicate.
23636 * config/mips/mips.md (multimem): New type.
23637 (length): Differentiate between 17-bit and 18-bit branch offsets.
23638 (MOVEP1, MOVEP2): New mode iterator.
23639 (mov_<load>l): Use ZC constraint.
23640 (mov_<load>r): Likewise.
23641 (mov_<store>l): Likewise.
23642 (mov_<store>r): Likewise.
23643 (*branch_equality<mode>_inverted): Add microMIPS support.
23644 (*branch_equality<mode>): Likewise.
23645 (*jump_absolute): Likewise.
23646 (indirect_jump_<mode>): Likewise.
23647 (tablejump_<mode>): Likewise.
23648 (<optab>_internal): Likewise.
23649 (sibcall_internal): Likewise.
23650 (sibcall_value_internal): Likewise.
23651 (prefetch): Use constraint ZD.
23652 * config/mips/mips.opt (minterlink-compressed): New option.
23653 (minterlink-mips16): Now an alias for minterlink-compressed.
23654 (mmicromips): New option.
23655 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23656 (compare_and_swap_12): Likewise.
23657 (sync_add<mode>): Likewise.
23658 (sync_<optab>_12): Likewise.
23659 (sync_old_<optab>_12): Likewise.
23660 (sync_new_<optab>_12): Likewise.
23661 (sync_nand_12): Likewise.
23662 (sync_old_nand_12): Likewise.
23663 (sync_new_nand_12): Likewise.
23664 (sync_sub<mode>): Likewise.
23665 (sync_old_add<mode>): Likewise.
23666 (sync_old_sub<mode>): Likewise.
23667 (sync_new_add<mode>): Likewise.
23668 (sync_new_sub<mode>): Likewise.
23669 (sync_<optab><mode>): Likewise.
23670 (sync_old_<optab><mode>): Likewise.
23671 (sync_new_<optab><mode>): Likewise.
23672 (sync_nand<mode>): Likewise.
23673 (sync_old_nand<mode>): Likewise.
23674 (sync_new_nand<mode>): Likewise.
23675 (sync_lock_test_and_set<mode>): Likewise.
23676 (test_and_set_12): Likewise.
23677 (atomic_compare_and_swap<mode>): Likewise.
23678 (atomic_exchange<mode>_llsc): Likewise.
23679 (atomic_fetch_add<mode>_llsc): Likewise.
23680 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23681 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23682 (umips_save_restore_pattern_p): Likewise.
23683 (umips_load_store_pair_p): Likewise.
23684 (umips_output_load_store_pair): Likewise.
23685 (umips_movep_target_p): Likewise.
23686 (umips_12bit_offset_address_p): Likewise.
23687 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23688 (mips_base_mips16): Rename this...
23689 (mips_base_compression_flags): ...to this. Update all uses.
23690 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23691 (mips_mips16_decl_p): Delete.
23692 (mips_nomips16_decl_p): Delete.
23693 (mips_get_compress_on_flags): New function.
23694 (mips_get_compress_off_flags): New function.
23695 (mips_get_compress_mode): New function.
23696 (mips_get_compress_on_name): New function.
23697 (mips_get_compress_off_name): New function.
23698 (mips_insert_attributes): Support multiple compression types.
23699 (mips_merge_decl_attributes): Likewise.
23700 (umips_12bit_offset_address_p): New function.
23701 (mips_start_function_definition): Emit .set micromips directive.
23702 (mips_call_may_need_jalx_p): New function.
23703 (mips_function_ok_for_sibcall): Add microMIPS support.
23704 (mips_print_operand_punctuation): Support short delay slots and
23706 (umips_swm_mask, umips_swm_encoding): New.
23707 (umips_build_save_restore): New function.
23708 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23709 (was_mips16_p): Remove.
23710 (old_compression_mode): New.
23711 (mips_set_compression_mode): New function.
23712 (mips_set_current_function): Add microMIPS support.
23713 (mips_option_override): Likewise.
23714 (umips_save_restore_pattern_p): New function.
23715 (umips_output_save_restore): New function.
23716 (umips_load_store_pair_p_1): New function.
23717 (umips_load_store_pair_p): New function.
23718 (umips_output_load_store_pair_1): New function.
23719 (umips_output_load_store_pair): New function.
23720 (umips_movep_target_p) New function.
23721 (mips_prepare_pch_save): Add microMIPS support.
23722 * config/mips/mips.h (TARGET_COMPRESSION): New.
23723 (TARGET_CPU_CPP_BUILTINS): Update macro
23724 to use new compression flags and to support microMIPS.
23725 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23726 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23727 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23728 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23729 (ASM_SPEC): Support mmicromips and mno-micromips.
23730 (M16STORE_REG_P): New macro.
23731 (MIPS_CALL): Support TARGET_MICROMIPS.
23732 (MICROMIPS_J): New macro.
23733 (mips_base_mips16): Rename this...
23734 (mips_base_compression_flags): ...to this.
23735 (UMIPS_12BIT_OFFSET_P): New macro.
23736 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23737 (MULTILIB_DIRNAMES): Likewise.
23739 2013-02-25 Tom de Vries <tom@codesourcery.com>
23741 PR rtl-optimization/56131
23742 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23743 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23744 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23746 2013-02-25 Tobias Burnus <burnus@net-b.de>
23748 * doc/invoke.texi (-fsanitize=): Move from optimization
23749 to debugging options.
23751 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23753 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23755 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23756 Alexander Monakov <amonakov@ispras.ru>
23758 PR middle-end/56077
23759 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23760 flush pending lists also on non-jumps. Adjust comment.
23762 2013-02-24 Catherine Moore <clm@codesourcery.com>
23763 Maciej W. Rozycki <macro@codesourcery.com>
23764 Tom de Vries <tom@codesourcery.com>
23765 Nathan Sidwell <nathan@codesourcery.com>
23766 Iain Sandoe <iain@codesourcery.com>
23767 Nathan Froyd <froydnj@codesourcery.com>
23768 Chao-ying Fu <fu@mips.com>
23770 * doc/extend.texi: (micromips, nomicromips, nocompression):
23771 Document new function attributes.
23772 * doc/invoke.texi (minterlink-compressed, mmicromips,
23773 m14k, m14ke, m14kec): Document new options.
23774 (minterlink-mips16): Update documentation.
23775 * doc/md.texi (ZC, ZD): Document new constraints.
23776 * configure.ac (gcc_cv_as_micromips): Check if linker
23777 supports the .set micromips directive.
23778 * configure: Regenerate.
23779 * config.in: Regenerate.
23780 * config/mips/mips-tables.opt: Regenerate.
23781 * config/mips/micromips.md: New file.
23782 * constraints.md (ZC, AD): New constraints.
23783 * config/mips/predicates.md (movep_src_register): New predicate.
23784 (movep_src_operand): New predicate.
23785 (non_volatile_mem_operand): New predicate.
23786 * config/mips/mips.md (multimem): New type.
23787 (length): Differentiate between 17-bit and 18-bit branch offsets.
23788 (MOVEP1, MOVEP2): New mode iterator.
23789 (mov_<load>l): Use ZC constraint.
23790 (mov_<load>r): Likewise.
23791 (mov_<store>l): Likewise.
23792 (mov_<store>r): Likewise.
23793 (*branch_equality<mode>_inverted): Add microMIPS support.
23794 (*branch_equality<mode>): Likewise.
23795 (*jump_absolute): Likewise.
23796 (indirect_jump_<mode>): Likewise.
23797 (tablejump_<mode>): Likewise.
23798 (<optab>_internal): Likewise.
23799 (sibcall_internal): Likewise.
23800 (sibcall_value_internal): Likewise.
23801 (prefetch): Use constraint ZD.
23802 * config/mips/mips.opt (minterlink-compressed): New option.
23803 (minterlink-mips16): Now an alias for minterlink-compressed.
23804 (mmicromips): New option.
23805 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23806 (compare_and_swap_12): Likewise.
23807 (sync_add<mode>): Likewise.
23808 (sync_<optab>_12): Likewise.
23809 (sync_old_<optab>_12): Likewise.
23810 (sync_new_<optab>_12): Likewise.
23811 (sync_nand_12): Likewise.
23812 (sync_old_nand_12): Likewise.
23813 (sync_new_nand_12): Likewise.
23814 (sync_sub<mode>): Likewise.
23815 (sync_old_add<mode>): Likewise.
23816 (sync_old_sub<mode>): Likewise.
23817 (sync_new_add<mode>): Likewise.
23818 (sync_new_sub<mode>): Likewise.
23819 (sync_<optab><mode>): Likewise.
23820 (sync_old_<optab><mode>): Likewise.
23821 (sync_new_<optab><mode>): Likewise.
23822 (sync_nand<mode>): Likewise.
23823 (sync_old_nand<mode>): Likewise.
23824 (sync_new_nand<mode>): Likewise.
23825 (sync_lock_test_and_set<mode>): Likewise.
23826 (test_and_set_12): Likewise.
23827 (atomic_compare_and_swap<mode>): Likewise.
23828 (atomic_exchange<mode>_llsc): Likewise.
23829 (atomic_fetch_add<mode>_llsc): Likewise.
23830 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23831 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23832 (umips_save_restore_pattern_p): Likewise.
23833 (umips_load_store_pair_p): Likewise.
23834 (umips_output_load_store_pair): Likewise.
23835 (umips_movep_target_p): Likewise.
23836 (umips_12bit_offset_address_p): Likewise.
23837 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23838 (mips_base_mips16): Rename this...
23839 (mips_base_compression_flags): ...to this. Update all uses.
23840 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23841 (mips_mips16_decl_p): Delete.
23842 (mips_nomips16_decl_p): Delete.
23843 (mips_get_compress_on_flags): New function.
23844 (mips_get_compress_off_flags): New function.
23845 (mips_get_compress_mode): New function.
23846 (mips_get_compress_on_name): New function.
23847 (mips_get_compress_off_name): New function.
23848 (mips_insert_attributes): Support multiple compression types.
23849 (mips_merge_decl_attributes): Likewise.
23850 (umips_12bit_offset_address_p): New function.
23851 (mips_start_function_definition): Emit .set micromips directive.
23852 (mips_call_may_need_jalx_p): New function.
23853 (mips_function_ok_for_sibcall): Add microMIPS support.
23854 (mips_print_operand_punctuation): Support short delay slots and
23856 (umips_swm_mask, umips_swm_encoding): New.
23857 (umips_build_save_restore): New function.
23858 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23859 (was_mips16_p): Remove.
23860 (old_compression_mode): New.
23861 (mips_set_compression_mode): New function.
23862 (mips_set_current_function): Add microMIPS support.
23863 (mips_option_override): Likewise.
23864 (umips_save_restore_pattern_p): New function.
23865 (umips_output_save_restore): New function.
23866 (umips_load_store_pair_p_1): New function.
23867 (umips_load_store_pair_p): New function.
23868 (umips_output_load_store_pair_1): New function.
23869 (umips_output_load_store_pair): New function.
23870 (umips_movep_target_p) New function.
23871 (mips_prepare_pch_save): Add microMIPS support.
23872 * config/mips/mips.h (TARGET_COMPRESSION): New.
23873 (TARGET_CPU_CPP_BUILTINS): Update macro
23874 to use new compression flags and to support microMIPS.
23875 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23876 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23877 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23878 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23879 (ASM_SPEC): Support mmicromips and mno-micromips.
23880 (M16STORE_REG_P): New macro.
23881 (MIPS_CALL): Support TARGET_MICROMIPS.
23882 (MICROMIPS_J): New macro.
23883 (mips_base_mips16): Rename this...
23884 (mips_base_compression_flags): ...to this.
23885 (UMIPS_12BIT_OFFSET_P): New macro.
23886 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23887 (MULTILIB_DIRNAMES): Likewise.
23889 2013-02-24 Jakub Jelinek <jakub@redhat.com>
23892 * target-globals.c (save_target_globals): For init_reg_sets and
23893 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23895 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23897 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23898 * config/aarch64/t-aarch64
23899 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23901 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23903 PR inline-asm/56148
23904 * lra-constraints.c (process_alt_operands): Reload operand
23905 conflicting with earlier clobber only if no more other conflicting
23908 2013-02-22 Jakub Jelinek <jakub@redhat.com>
23911 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23912 if not linking a shared library.
23914 2013-02-22 Seth LaForge <sethml@google.com>
23916 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23918 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23920 * config/arm/arm.md (split for extendsidi): Update condition.
23921 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23922 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23923 (qhs_zextenddi_cstr): Likewise.
23925 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23927 PR middle-end/56420
23928 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23929 avoid signed wrapping.
23930 (expand_mult): Handle properly multiplication by
23931 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23932 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23933 in the compiler if coeff is HOST_WIDE_INT_MIN.
23934 (expand_divmod): Don't make ext_op1 static, change it's type to
23935 uhwi. Avoid undefined behavior in -INTVAL (op1).
23937 PR rtl-optimization/50339
23938 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23940 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23941 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23942 into splitting_ashiftrt field.
23943 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23945 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23948 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
23950 PR middle-end/56108
23951 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23952 are sure to go irrevocable.
23954 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
23956 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23957 scalars are valid operands.
23959 2013-02-21 Martin Jambor <mjambor@suse.cz>
23961 PR tree-optimization/56310
23962 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23963 only matching indices and non-negative final offsets.
23964 (intersect_aggregates_with_edge): Pass src_idx to
23965 agg_replacements_to_vector. Pass src_idx insstead of index to
23966 intersect_with_agg_replacements.
23968 2013-02-21 Martin Jambor <mjambor@suse.cz>
23970 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23971 instead of hard-wired defaults.
23973 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23975 * doc/invoke.texi (MIPS Options): Update documentation of the
23976 floating-point multiply-accumulate instruction restrictions.
23978 2013-02-21 Kostya Serebryany <kcc@google.com>
23980 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
23981 asan_shadow_offset on x86_64 linux.
23983 2013-02-21 Richard Biener <rguenther@suse.de>
23985 PR tree-optimization/56415
23987 2013-02-11 Richard Biener <rguenther@suse.de>
23989 PR tree-optimization/56273
23990 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23993 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23996 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
23999 PR inline-asm/56405
24000 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24001 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24003 2013-02-20 Jan Hubicka <jh@suse.cz>
24005 PR tree-optimization/56265
24006 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24007 when target is referenced for first time.
24009 2013-02-20 Richard Biener <rguenther@suse.de>
24011 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24012 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24013 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24014 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24015 not return anything.
24016 (rename_ssa_copies): Do not remove unused locals.
24017 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24018 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24019 * passes.c (execute_function_todo): Do not schedule unused locals
24020 removal if cleanup_tree_cfg did something.
24021 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24022 about the number of removed locals.
24024 2013-02-20 Richard Biener <rguenther@suse.de>
24026 PR tree-optimization/56398
24027 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24029 2013-02-20 Martin Jambor <mjambor@suse.cz>
24031 PR tree-optimization/55334
24032 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24033 restricted pointers to arrays.
24035 2013-02-20 Richard Biener <rguenther@suse.de>
24036 Jakub Jelinek <jakub@redhat.com>
24038 PR tree-optimization/56396
24039 * tree-ssa-ccp.c (n_const_val): New static variable.
24040 (get_value): Return NULL for SSA names we don't have a lattice
24042 (ccp_initialize): Initialize n_const_val.
24043 * tree-ssa-copy.c (n_copy_of): New static variable.
24044 (init_copy_prop): Initialize n_copy_of.
24045 (get_value): Return NULL_TREE for SSA names we don't have a
24048 2013-02-20 Martin Jambor <mjambor@suse.cz>
24050 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24052 2013-02-20 Richard Biener <rguenther@suse.de>
24054 * genpreds.c (write_lookup_constraint): Do not compare first
24055 letter of the constraint again.
24057 2013-02-20 Richard Biener <rguenther@suse.de>
24059 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24061 (get_use_iv_cost): Terminate hashtable walk when coming across
24064 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24066 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24067 reassociation for avx2 targets.
24069 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24071 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24072 Add version check for v8.10.a to enable microblaze_has_clz
24073 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24074 version and TARGET_PATTERN_COMPARE check
24075 * config/microblaze/microblaze.md: New clzsi2 instruction
24077 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24079 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24080 function before branching.
24082 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24084 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24086 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24088 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24090 PR middle-end/55889
24091 * sel-sched.c: Include ira.h.
24092 (implicit_clobber_conflict_p): New function.
24093 (moveup_expr): Use it.
24094 * Makefile.in (sel-sched.o): Depend on ira.h.
24096 2013-02-19 Richard Biener <rguenther@suse.de>
24098 PR tree-optimization/56384
24099 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24100 (vn_hash_type): Split out from ...
24101 (vn_hash_constant_with_type): ... here.
24102 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24103 (vn_phi_eq): Compare types from vn_phi_s structure.
24104 (vn_phi_lookup): Populate vn_phi_s type.
24105 (vn_phi_insert): Likewise.
24107 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24109 PR tree-optimization/56350
24110 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24111 if haven't found reduction or nested cycle operand, rather than
24112 asserting we must find it.
24114 PR tree-optimization/56381
24115 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24118 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24119 Jakub Jelinek <jakub@redhat.com>
24122 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24123 (swap_optab_enable): Same.
24124 (init_all_optabs): Use argument instead of global.
24125 * tree.h (struct tree_optimization_option): New field target_optabs.
24126 * expr.h (init_all_optabs): Add argument to prototype.
24127 (TREE_OPTIMIZATION_OPTABS): New.
24128 (save_optabs_if_changed): Protoize.
24129 * optabs.h: Declare this_fn_optabs.
24130 * optabs.c (save_optabs_if_changed): New.
24131 Declare this_fn_optabs.
24132 (init_optabs): Add argument to init_all_optabs() call.
24133 * function.c (invoke_set_current_function_hook): Handle per
24135 * function.h (struct function): New field optabs.
24136 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24137 optimization_current_node has changed.
24138 * target-globals.h (save_target_globals_default_opts): Protoize.
24139 * target-globals.c (save_target_globals_default_opts): New.
24141 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24144 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24145 registers %fr12 and %fr12R as call used.
24148 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24149 and HImode, require all displacements to be an integer multiple of
24151 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24152 only allow QImode and HImode when reload is in progress and strict is
24153 true. Likewise for symbolic addresses. Use base14_operand to check
24154 displacements in REG+BASE addresses.
24156 2013-02-18 Richard Biener <rguenther@suse.de>
24158 PR tree-optimization/56366
24159 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24160 handle sign-conversion of outer-loop initial induction value.
24162 2013-02-18 Richard Biener <rguenther@suse.de>
24164 PR middle-end/56349
24165 * cfghooks.c (merge_blocks): If we merge a latch into another
24166 block adjust references to it.
24167 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24168 (verify_loop_structure): Verify that a recorded latch is in fact
24171 2013-02-18 Richard Biener <rguenther@suse.de>
24173 PR tree-optimization/56321
24174 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24175 order SSA name release and virtual operand unlinking.
24177 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24179 * config/microblaze/microblaze.md (save_stack_block): Define.
24180 (restore_stack_block): Likewise.
24182 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24184 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24185 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24186 * config/microblaze/microblaze.c (microblaze_option_override):
24187 Bail out early for PIC modes when target does not support PIC.
24189 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24191 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24192 Replace with a microblaze version.
24193 (microblaze_trampoline_init): Adapt for microblaze.
24194 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24197 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24198 Dodji Seketeli <dodji@redhat.com>
24201 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24202 (instrument_mem_region_access): Do not forget to always put
24203 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24204 0) statement, even for cases where either 'base' or 'base + len'
24205 are not instrumented -- because they have been previously
24206 instrumented. Simplify the logic by putting all the statements
24207 instrument 'base + len' inside a sequence, and then insert that
24208 sequence right before the current insertion point. Then, to
24209 instrument 'base + len', just get an iterator on that statement.
24210 And do not forget to update the pointer to iterator the function
24211 received as argument.
24213 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24215 PR rtl-optimization/56348
24216 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24218 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24220 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24221 (clean_graph_dump_file): Pass base to start_graph_dump.
24223 2013-02-14 Richard Henderson <rth@redhat.com>
24226 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24228 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24230 * collect2-aix.h: Define F_LOADONLY.
24232 2013-02-14 Richard Biener <rguenther@suse.de>
24235 * varasm.c (output_constant_def_1): Get the decl representing
24236 the constant as argument.
24237 (output_constant_def): Wrap output_constant_def_1.
24238 (make_decl_rtl): Use output_constant_def_1 with the decl
24239 representing the constant.
24240 (build_constant_desc): Optionally re-use a decl already
24241 representing the constant.
24242 (tree_output_constant_def): Adjust.
24244 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24247 * asan.c (instrument_builtin_call): Really put the length of the
24248 second source argument into src1_len.
24250 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24252 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24253 argument. If it is false, don't create edge from then_bb to
24255 (insert_if_then_before_iter): Pass true to it.
24256 (build_check_stmt): Pass false to it.
24257 (transform_statements): Flush hash table only on extended basic
24258 block boundaries, rather than at the beginning of every bb.
24259 Don't flush hash table on nonfreeing_call_p calls.
24260 * tree-flow.h (nonfreeing_call_p): New prototype.
24261 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24263 2013-02-13 David S. Miller <davem@davemloft.net>
24265 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24267 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24270 * ira.c (max_regno_before_ira): Move from ...
24272 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24275 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24277 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24279 2013-02-13 Richard Biener <rguenther@suse.de>
24282 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24283 globals in MEM_REFs.
24285 2013-02-13 Richard Biener <rguenther@suse.de>
24287 * loop-init.c (loop_optimizer_init): Clear loop state when
24288 re-initializing preserved loops.
24289 * loop-unswitch.c (unswitch_single_loop): Return whether
24290 we unswitched the loop. Do not verify loop state here.
24291 (unswitch_loops): When we unswitched a loop discover new loops.
24293 2013-02-13 Kostya Serebryany <kcc@google.com>
24295 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24297 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24299 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24301 Avoid instrumenting duplicated memory access in the same basic block
24302 * Makefile.in (asan.o): Add new dependency on hash-table.h
24303 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24304 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24305 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24306 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24307 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24308 (get_mem_ref_of_assignment): New functions.
24309 (get_mem_refs_of_builtin_call): Extract from
24310 instrument_builtin_call and tweak a little bit to make it fit with
24312 (instrument_builtin_call): Use the new
24313 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24314 of is_gimple_builtin_call.
24315 (instrument_derefs, instrument_mem_region_access): Insert the
24316 instrumented memory reference into the hash table.
24317 (maybe_instrument_assignment): Renamed instrument_assignment into
24318 this, and change it to advance the iterator when instrumentation
24319 actually happened and return true in that case. This makes it
24320 homogeneous with maybe_instrument_assignment, and thus give a
24321 chance to callers to be more 'regular'.
24322 (transform_statements): Clear the memory reference hash table
24323 whenever we enter a new BB, when we cross a function call, or when
24324 we are done transforming statements. Use
24325 maybe_instrument_assignment instead of instrumentation. No more
24326 need to special case maybe_instrument_assignment and advance the
24327 iterator after calling it; it's now handled just like
24328 maybe_instrument_call. Update comment.
24330 2013-02-13 Richard Biener <rguenther@suse.de>
24332 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24333 Fix loop discovery code.
24335 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24337 PR inline-asm/56148
24338 * lra-constraints.c (process_alt_operands): Match early clobber
24339 operand with itself. Check conflicts with earlyclobber only if
24340 the operand is not reloaded. Prefer to reload conflicting operand
24341 if earlyclobber and matching operands are the same.
24343 2013-02-12 Richard Biener <rguenther@suse.de>
24346 * lto-streamer-out.c (write_symbol): Do not output symbols
24347 for hard register variables.
24349 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24352 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24353 (umulsidi3_insn, mulsidi3_insn): New insns.
24355 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24357 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24358 (struct tune_params): Add vec_costs field.
24359 * config/arm/arm.c (arm_builtin_vectorization_cost)
24360 (arm_add_stmt_cost): New functions.
24361 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24362 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24363 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24364 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24365 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24366 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24367 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24369 2013-02-12 Richard Biener <rguenther@suse.de>
24372 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24373 decls again if possible.
24375 2013-02-12 Richard Biener <rguenther@suse.de>
24377 PR middle-end/56288
24378 * tree-ssa.c (verify_ssa_name): Fix check, move
24379 SSA_NAME_IN_FREE_LIST check up.
24381 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24382 Steven Bosscher <steven@gcc.gnu.org>
24384 PR rtl-optimization/56151
24385 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24386 equal to op0 or op1, and last_insn pattern is CODE operation
24387 with MEM dest and one of the operands matches that MEM.
24389 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24391 * doc/extend.texi: Document Function Multiversioning and "default"
24392 parameter string to target attribute.
24393 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24394 target attribute parameter is "default".
24395 (ix86_compare_version_priority): Remove checks for target attribute.
24396 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24397 Remove check for target attribute equal to NULL. Add assert.
24398 (ix86_generate_version_dispatcher_body): Change error to sorry.
24400 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24401 Jack Howarth <howarth@bromo.med.uc.edu>
24402 Patrick Marlier <patrick.marlier@gmail.com>
24405 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24406 define ENDFILE_SPEC as TM_DESTRUCTOR.
24407 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24409 2013-02-11 Alexander Potapenko <glider@google.com>
24410 Jack Howarth <howarth@bromo.med.uc.edu>
24411 Jakub Jelinek <jakub@redhat.com>
24414 * config/darwin.c (cdtor_record): Rename ctor_record.
24415 (sort_cdtor_records): Rename sort_ctor_records.
24416 (finalize_dtors): New routine to sort destructors by
24417 priority before use in assemble_integer.
24418 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24420 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24422 PR rtl-optimization/56275
24423 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24424 offset is non-negative and less than cmode size before
24425 calling simplify_subreg.
24427 2013-02-11 Richard Biener <rguenther@suse.de>
24429 PR tree-optimization/56264
24430 * cfgloop.h (fix_loop_structure): Adjust prototype.
24431 * loop-init.c (fix_loop_structure): Return the number of
24432 newly discovered loops.
24433 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24434 are discovered, do a full loop-closed SSA rewrite.
24436 2013-02-11 Richard Biener <rguenther@suse.de>
24438 PR tree-optimization/56273
24439 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24441 (check_array_ref): Fix missing newline in dumps.
24442 (search_for_addr_array): Likewise.
24444 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24446 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24448 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24451 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24453 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24455 PR rtl-optimization/56246
24456 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24458 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24459 constraints are satisfied.
24461 2013-02-08 Jeff Law <law@redhat.com>
24464 * emit-rtl.c (reg_is_parm_p): New function.
24465 * regs.h (reg_is_parm_p): New prototype.
24466 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24467 callee-clobbered registers.
24469 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24472 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24473 If there is no implicit builtin declaration, just return NULL.
24475 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24477 * config/i386/sse.md (FMAMODEM): New mode iterator.
24478 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24479 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24481 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24483 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24484 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24485 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24487 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24489 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24490 (microblaze*-*-elf): Likewise.
24491 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24493 * config/microblaze/microblaze-c.c: Add builtin defines for
24494 _LITTLE_ENDIAN and _BIG_ENDIAN.
24495 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24496 add to TARGET_DEFAULT flags.
24497 Expand ASM_SPEC and LINK_SPEC.
24498 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24499 * config/microblaze/microblaze.md: Update extendsidi2 and
24500 movdi_internal instructions to use low-order / high-order reg
24502 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24503 options and inversemask / mask of LITTLE_ENDIAN.
24504 * config/microblaze/t-microblaze: Expand multilib options to
24505 include mlittle-endian (le) and update exceptions patterns.
24507 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24509 PR rtl-optimization/56195
24510 * lra-constraints.c (get_reload_reg): Don't reuse regs
24511 if they have smaller mode than requested, if they have
24512 wider mode than requested, try to return a SUBREG.
24514 PR tree-optimization/56250
24515 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24516 if type is unsigned and code isn't MULT_EXPR.
24518 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24520 PR tree-optimization/56064
24521 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24522 bits according to mode.
24523 * fixed-value.h (fixed_from_double_int)
24524 (const_fixed_from_double_int): Adjust comments.
24526 2013-02-08 Richard Biener <rguenther@suse.de>
24529 * lto-streamer.h (struct data_in): Remove current_file, current_line
24530 and current_col members.
24531 * lto-streamer-out.c (lto_output_location): Stream changed bits
24532 en-block for efficiency.
24533 * lto-streamer-in.c (clear_line_info): Remove.
24534 (lto_input_location): Cache current file, line and column
24535 globally via local statics. Read changed bits en-block.
24536 (input_function): Do not call clear_line_info.
24537 (lto_read_body): Likewise.
24538 (lto_input_toplevel_asms): Likewise.
24540 2013-02-08 Michael Matz <matz@suse.de>
24542 PR tree-optimization/52448
24543 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24544 (nt_call_phase): New static.
24545 (add_or_mark_expr): Only mark accesses with newer phase than any
24547 (nonfreeing_call_p): New.
24548 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24549 (nt_fini_block): Keep blocks marked as visited.
24550 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24552 2013-02-08 Richard Biener <rguenther@suse.de>
24554 * ira.c (ira): Free broken dominator information.
24556 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24558 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24560 2013-02-08 Marek Polacek <polacek@redhat.com>
24562 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24564 2013-02-08 Richard Biener <rguenther@suse.de>
24566 PR middle-end/56181
24567 * cfgloop.h (flow_loops_find): Adjust.
24568 (bb_loop_header_p): Declare.
24569 * cfgloop.c (bb_loop_header_p): New function split out from ...
24570 (flow_loops_find): ... here. Adjust function signature,
24571 support incremental loop structure update.
24572 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24573 * cfgloopmanip.c (fix_loop_structure): Move ...
24574 * loop-init.c (fix_loop_structure): ... here.
24575 (apply_loop_flags): Split out from ...
24576 (loop_optimizer_init): ... here.
24577 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24578 in incremental mode, only remove dead loops here.
24580 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24583 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24584 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24585 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24586 (*round<mode>3.libgcc): New insns for fixed-modes.
24587 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24588 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24589 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24590 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24591 implementations. Define to __builtin_avr_absFX,
24592 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24593 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24594 __builtin_avr_countlsFX, respectively.
24595 * config/avr/avr-c.c (target.h): Include it.
24596 (enum avr_builtin_id): New enum.
24597 (avr_resolve_overloaded_builtin): New static function.
24598 (avr_register_target_pragmas): Use it to set
24599 targetm.resolve_overloaded_builtin.
24600 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24601 tree nodes used by DEF_BUILTIN.
24602 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24603 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24604 <AVR_BUILTIN_xxBITS>: Same.
24606 2013-02-08 Richard Biener <rguenther@suse.de>
24608 * cfgloop.c (verify_loop_structure): Properly handle
24609 a loop exiting to another loop header.
24610 * ira-int.h (ira_loops): Remove.
24611 * ira.c (ira_loops): Remove.
24612 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24613 (do_reload): Use loop_optimizer_finalize.
24614 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24615 number_of_loops to access the loop tree.
24616 (more_one_region_p): Likewise.
24617 (finish_loop_tree_nodes): Likewise.
24618 (rebuild_regno_allocno_maps): Likewise.
24619 (mark_loops_for_removal): Likewise.
24620 (mark_all_loops_for_removal): Likewise.
24621 (remove_unnecessary_regions): Likewise.
24622 (ira_build): Likewise.
24623 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24625 2013-02-08 Richard Biener <rguenther@suse.de>
24627 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24628 * ipa-pure-const.c (analyze_function): Avoid calling
24629 mark_irreducible_loops twice.
24630 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24632 2013-02-07 David S. Miller <davem@davemloft.net>
24634 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24636 * var-tracking.c (vt_add_function_parameter): Test the presence of
24637 HAVE_window_save properly and do not remap argument registers when
24638 we have a leaf function.
24640 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24643 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24645 * config/i386/i386.c (ix86_print_operand): Ditto.
24647 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24649 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24651 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24653 PR rtl-optimization/56225
24654 * lra-constraints.c (process_alt_operands): Check that reload hard
24655 reg can hold value for strict_low_part.
24657 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24660 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24661 dwarf2out_end_function.
24662 (in_first_function_p, maybe_at_text_label_p,
24663 first_loclabel_num_not_at_text_label): New variables.
24664 (dwarf2out_var_location): In the first function find out
24665 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24666 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24669 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24671 PR rtl-optimization/56178
24672 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24673 SUBREG of a register. Tidy up related block of code.
24674 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24675 note if the source is a register or a SUBREG of a register.
24677 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24680 * config/rs6000/rs6000.md (ptrm): New mode attr.
24681 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24682 call_value_indirect_aix<pttrsize>,
24683 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24686 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24688 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24689 if -bnortl. Convert to strcmp and strncmp.
24691 2013-02-07 Alan Modra <amodra@gmail.com>
24694 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24695 addresses won't wrap when offsetting.
24696 (rs6000_secondary_reload): Provide secondary reloads needed for
24697 wrapping LO_SUM addresses.
24699 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24701 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24702 MACH, just __MACH__.
24704 2013-02-06 Richard Biener <rguenther@suse.de>
24706 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24707 instead of calling fix_loop_structure.
24709 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24711 PR middle-end/56217
24712 * omp-low.c (use_pointer_for_field): Return false if
24713 lower_send_shared_vars doesn't generate any copy-out code.
24715 2013-02-06 Tom de Vries <tom@codesourcery.com>
24717 PR rtl-optimization/56131
24718 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24719 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24720 of the label is NULL. Add comment.
24722 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24724 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24727 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24728 (STATIC_LIBTSAN_LIBS): Likewise.
24729 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24730 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24731 is defined, don't add anything else beyond that.
24732 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24733 (LINK_COMMAND_SPEC): Use them.
24735 PR tree-optimization/56205
24736 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24737 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24738 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24740 2013-02-05 Richard Biener <rguenther@suse.de>
24742 PR tree-optimization/53342
24743 PR tree-optimization/53185
24744 * tree-vectorizer.h (vect_check_strided_load): Remove.
24745 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24746 not disallow peeling for vectorized strided loads.
24747 (vect_check_strided_load): Make static and simplify.
24748 (vect_analyze_data_refs): Adjust.
24749 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24750 correctly when vectorizing strided loads.
24752 2013-02-05 Richard Biener <rguenther@suse.de>
24754 * doc/install.texi: Refer to ISL, not PPL.
24756 2013-02-05 Jan Hubicka <jh@suse.cz>
24758 PR tree-optimization/55789
24759 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24761 2013-02-05 Jan Hubicka <jh@suse.cz>
24763 PR tree-optimization/55789
24764 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24765 the dead call anyway.
24767 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24770 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24772 2013-02-04 Alexander Potapenko <glider@google.com>
24773 Jack Howarth <howarth@bromo.med.uc.edu>
24774 Jakub Jelinek <jakub@redhat.com>
24777 * config/darwin.c (sort_ctor_records): Stabilized qsort
24778 on constructor priority by using original position.
24779 (finalize_ctors): New routine to sort constructors by
24780 priority before use in assemble_integer.
24781 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24783 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24786 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24787 about visibility on artificial decls.
24788 * config/sol2.c (solaris_assemble_visibility): Likewise.
24790 2013-02-04 Kai Tietz <ktietz@redhat.com>
24793 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24794 argument and improve checking of return-argument types for 16-byte
24796 (ix86_function_value_1): Add additional valtype argument on call
24797 of function_value_64.
24798 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24799 handling infunction_value_64 function.
24801 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24803 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24805 2013-02-04 Richard Biener <rguenther@suse.de>
24807 PR tree-optimization/56188
24808 * tree-ssa-structalias.c (label_visit): Consider case with
24809 initially non-empty points-to set.
24810 (perform_var_substitution): Dump node mapping and clean up.
24812 2013-02-04 Richard Guenther <rguenther@suse.de>
24815 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24816 node prevail as last resort.
24817 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24818 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24820 2013-02-04 Richard Biener <rguenther@suse.de>
24822 PR tree-optimization/56113
24823 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24825 (equiv_class_lookup_or_add): ... this.
24826 (label_visit): Adjust and fix error in previous patch.
24827 (perform_var_substitution): Adjust.
24829 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24831 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24832 * config/sh/sh4-300.md: Likewise.
24833 * config/sh/sh4a.md: Likewise.
24834 * config/sh/constraints.md: Likewise.
24835 * config/sh/sh.md: Likewise.
24836 * config/sh/netbsd-elf.h: Likewise.
24837 * config/sh/predicates.md: Likewise.
24838 * config/sh/sh-protos.h: Likewise.
24839 * config/sh/ushmedia.h: Likewise.
24840 * config/sh/linux.h: Likewise.
24841 * config/sh/sh.c: Likewise.
24842 * config/sh/superh.h: Likewise.
24843 * config/sh/elf.h: Likewise.
24844 * config/sh/sh4.md: Likewise.
24845 * config/sh/sh.h: Likewise.
24847 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24849 * config/pa/constraints.md: Adjust unused letters. Change "T"
24850 constraint to match_test floating_point_store_memory_operand().
24851 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24852 (base14_operand): New.
24853 (floating_point_store_memory_operand): New.
24854 (integer_store_memory_operand): Revise to use base14_operand and
24855 reg_plus_base_memory_operand.
24856 (move_dest_operand): Allow symbolic_memory_operands.
24857 (symbolic_memory_operand): Check for LO_SOM.
24858 (symbolic_operand): Change default case to break.
24859 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24860 CONST_DOUBLE values to be reloaded by putting them into memory when
24861 the destination is a floating point register.
24862 (movdf): Remove code to handle CONST_DOUBLE.
24864 (reload_indf_r1): New.
24865 (reload_insf_r1): New.
24866 Consistently use "Q" and "T" constraints with integer and floating
24867 point move instructions, respectively.
24868 (movdi): Remove FAIL.
24869 Change predicate for source operand unamed DImode move from
24870 general_operand to move_src_operand.
24871 (umulsidi3): Change predicate for destination operand to
24873 Likewise for similar unamed patterns.
24874 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24875 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24876 (hppa_legitimize_address): Simplify mask calculation.
24877 (pa_emit_move_sequence): Revised handling of secondary reloads from
24878 REG+D addresses for floating point loads and stores. Directly handle
24879 loading CONST0_RTX (mode) to a floating point register.
24880 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24881 to floating point registers. Don't restrict secondary reloads to
24882 floating point registers to integer modes. Revise some comments and
24884 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24885 (pa_legitimate_address_p): New.
24886 (pa_legitimize_reload_address): New.
24887 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24888 (STRICT_REG_OK_FOR_BASE_P): New.
24889 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24890 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24892 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
24893 Andrew Dixie <andrewd@gentrack.com>
24895 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24898 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24900 * expmed.c (extract_bit_field_1): Pass the full width of the
24901 structure to get_best_reg_extraction_insn.
24903 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24906 * configure.ac (use_cxa_atexit): Add AIX.
24907 * configure: Regenerate.
24909 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24911 2013-02-01 Jakub Jelinek <jakub@redhat.com>
24914 * final.c (need_profile_function): New variable.
24915 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24916 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24917 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24918 notes, targetm.asm_out.function_prologue doesn't emit anything,
24919 HAVE_prologue and profiler should be emitted before prologue,
24920 set need_profile_function instead of emitting it.
24921 (final_scan_insn): If need_profile_function, emit
24922 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24923 NOTE_INSN_FUNCTION_BEG note.
24925 2013-02-01 Richard Henderson <rth@redhat.com>
24927 * config/rs6000/rs6000.md (smulditi3): New.
24930 * config/alpha/alpha.md (umulditi3): New.
24932 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24934 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24935 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24937 2013-02-01 Richard Biener <rguenther@suse.de>
24939 PR tree-optimization/56113
24940 * tree-ssa-structalias.c (label_visit): Reduce work for
24941 single-predecessor nodes.
24943 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24945 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24946 range isn't testing for zero.
24948 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24950 PR middle-end/56113
24951 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24953 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24954 Nick Clifton <nickc@redhat.com>
24956 * config/v850/constraints.md (Q): Define as a memory constraint.
24957 * config/v850/predicates.md (label_ref_operand): New predicate.
24958 (e3v5_shift_operand): New predicate.
24959 (ior_operator): New predicate.
24960 * config/v850/t-v850: Add e3v5 multilib.
24961 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24962 (v850_gen_movdi): Prototype.
24963 * config/v850/v850.c: Add support for e3v5 architecture.
24964 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24966 (construct_save_jarl): Add e3v5 long JARL support.
24967 (v850_adjust_insn_length): New function. Adjust length of call
24968 insns when using e3v5 instructions.
24969 (v850_gen_movdi): New function: Generate instructions to move a
24971 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24972 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24973 (TARGET_USE_FPU): Enable for e3v5.
24974 (CONST_OK_FOR_W): New macro.
24975 (ADJUST_INSN_LENGTH): Define.
24976 * config/v850/v850.md (UNSPEC_LOOP): Define.
24977 (attr cpu): Add v850e3v5.
24978 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24979 (movdi): New pattern.
24980 (movdi_internal): New pattern.
24981 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
24982 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
24983 (cstoresf4): Likewise.
24984 (cstoredf4): Likewise.
24985 (insv): New pattern.
24986 (rotlso3_a): New pattern.
24987 (rotlsi3_b): New pattern
24988 (rotlsi3_v850e3v5): New pattern.
24989 (doloop_begin): New pattern.
24990 (fix_loop_counter): New pattern.
24991 (doloop_end): New pattern.
24992 (branch_normal): Add e3v5 long branch support.
24993 (branch_invert): Likewise.
24994 (branch_z_normal): Likewise.
24995 (branch_z_invert): Likewise.
24996 (branch_nz_normal): Likewise.
24997 (branch_nz_invert): Likewise.
24998 (call_internal_short): Add e3v5 register-indirect JARL support.
24999 (call_internal_long): Likewise.
25000 (call_value_internal_short): Likewise.
25001 (call_value_internal_long): Likewise.
25002 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25003 (mloop): New option.
25004 * config.gcc: Add support for configuring v840e3v5 target.
25005 * doc/invoke.texi: Document new v850 specific command line options.
25007 2013-01-31 Paul Koning <ni1d@arrl.net>
25011 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25012 children if parent is a class.
25013 (prune_unused_types_prune): Don't add DW_AT_declaration.
25015 2013-01-31 Richard Biener <rguenther@suse.de>
25017 PR tree-optimization/56157
25018 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25019 match up operand with SLP child.
25021 2013-01-31 Jason Merrill <jason@redhat.com>
25024 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25025 parameters the first time.
25026 (gen_scheduled_generic_parms_dies): Check completeness here.
25028 2013-01-31 Richard Biener <rguenther@suse.de>
25030 PR middle-end/53073
25031 * common.opt (faggressive-loop-optimizations): New flag,
25032 enabled by default.
25033 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25034 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25035 infer_loop_bounds_from_undefined by it.
25037 2013-01-31 Richard Biener <rguenther@suse.de>
25039 PR tree-optimization/56150
25040 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25041 visit virtual operands.
25042 (find_uses_to_rename_bb): Likewise.
25044 2013-01-31 Richard Biener <rguenther@suse.de>
25046 PR tree-optimization/56150
25047 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25048 mixed store non-store stmts.
25050 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25053 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25054 LIBASAN_EARLY_SPEC is defined.
25055 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25056 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25058 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25061 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25062 invalid args instead of ICEing on it.
25063 (ix86_valid_target_attribute_tree): Return error_mark_node if
25064 ix86_valid_target_attribute_inner_p failed.
25065 (ix86_valid_target_attribute_p): Return false only if
25066 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25067 target("default") attribute.
25068 (sorted_attr_string): Change argument from const char * to tree,
25069 merge in all target attribute arguments rather than just one.
25070 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25071 instead of free. Avoid using strcat.
25072 (ix86_mangle_function_version_assembler_name): Mangle
25073 target("default") as if no target attribute is present. Adjust
25074 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25075 instead of xmalloc and XDELETEVEC instead of free.
25076 (ix86_function_versions): Don't return true if one of the decls
25077 doesn't have target attribute. If they don't and one of the decls
25078 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25079 sorted_attr_string caller. Use XDELETEVEC instead of free.
25080 (ix86_supports_function_versions): Remove.
25081 (make_name): Fix up formatting.
25082 (make_dispatcher_decl): Remove resolver_name and its initialization.
25083 Avoid leaking memory.
25084 (is_function_default_version): Return true if there is
25085 target("default") attribute rather than no target attribute at all.
25086 (make_resolver_func): Avoid leaking memory.
25087 (ix86_generate_version_dispatcher_body): Likewise.
25088 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25089 * target.def (supports_function_versions): Remove.
25090 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25091 * doc/tm.texi: Regenerated.
25093 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25095 PR rtl-optimization/56144
25096 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25097 for values with side effects.
25099 2013-01-30 Richard Biener <rguenther@suse.de>
25101 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25102 (sparseset_pop): Likewise.
25103 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25104 to be able to use quick_push in the worker loop.
25106 2013-01-30 Marek Polacek <polacek@redhat.com>
25108 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25110 2013-01-30 Richard Biener <rguenther@suse.de>
25113 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25115 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25117 PR tree-optimization/56064
25118 * fixed-value.c (fixed_from_double_int): New function.
25119 * fixed-value.h (fixed_from_double_int): New prototype.
25120 (const_fixed_from_double_int): New static inline function.
25121 * fold-const.c (native_interpret_fixed): New static function.
25122 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25123 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25124 (native_encode_fixed): New static function.
25125 (native_encode_expr) <FIXED_CST>: Use it.
25126 (native_interpret_int): Move double_int worker code to...
25127 * double-int.c (double_int::from_buffer): ...this new static method.
25128 * double-int.h (double_int::from_buffer): Prototype it.
25130 2013-01-30 Richard Biener <rguenther@suse.de>
25132 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25133 New pointer-map and obstack.
25134 (init_alias_vars): Allocate pointer-map and obstack.
25135 (delete_points_to_sets): Free them.
25136 (find_what_var_points_to): Cache result.
25137 (find_what_p_points_to): Adjust for changed interface of
25138 find_what_var_points_to.
25139 (compute_points_to_sets): Likewise.
25140 (ipa_pta_execute): Likewise.
25142 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25144 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25145 * configure: Regenerate.
25146 * config.in: Regenerate.
25147 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25148 #nobits/#progbits if supported.
25150 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25153 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25154 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25155 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25157 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25159 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25160 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25162 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25164 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25165 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25167 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25169 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25171 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25172 * config/arm/cortex-a7.md: New bypasses using
25173 arm_mac_accumulator_is_result.
25175 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25177 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25178 (cortex_a7_neon_mla): Likewise.
25179 (cortex_a7_fpfmad): New reservation.
25180 (cortex_a7_fpmacs): Use ffmas and update required units.
25181 (cortex_a7_fpmuld): Update required units and latency.
25182 (cortex_a7_fpmacd): Likewise.
25183 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25184 (cortex_a7_neon). Likewise.
25185 (bypass) Update participating units.
25187 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25189 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25190 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25192 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25193 (vfp_fmul): Use ffmad.
25194 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25195 (cortex_r4_fmacd): Use ffmad.
25196 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25197 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25198 (cortex_a9_fmacd): Use ffmad.
25199 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25200 (cortex_a8_vfp_macd): Use ffmad.
25201 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25202 (cortex_a5_fpmacd): Use ffmad.
25203 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25204 (cortex_a15_vfp_macd): Use ffmad.
25205 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25207 2013-01-29 Jason Merrill <jason@redhat.com>
25210 * varasm.c (default_assemble_visibility): Don't warn about
25211 visibility on artificial decls.
25213 2013-01-29 Richard Biener <rguenther@suse.de>
25215 PR tree-optimization/56113
25216 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25218 (label_visit): Free duplicate bitmaps and record the leader instead.
25219 (perform_var_substitution): Adjust.
25221 2013-01-29 Richard Biener <rguenther@suse.de>
25223 PR tree-optimization/55270
25224 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25225 the CFG, schedule loops for fixup.
25227 2013-01-29 Nick Clifton <nickc@redhat.com>
25229 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25232 2013-01-28 Leif Ekblad <leif@rdos.net>
25234 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25235 * config/i386/i386.h (TARGET_RDOS): New macro.
25236 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25237 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25238 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25239 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25240 DEFAULT_LARGE_SECTION_THRESHOLD.
25241 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25242 * config/i386/rdos.h: New file.
25243 * config/i386/rdos64.h: New file.
25245 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25248 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25251 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25253 PR rtl-optimization/56117
25254 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25255 call cselib_lookup_from_insn on the MEM before calling
25256 add_insn_mem_dependence.
25258 2013-01-28 Richard Biener <rguenther@suse.de>
25260 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25261 to a stmt that didn't have one.
25262 (copy_phis_for_bb): Likewise for PHI arguments.
25263 (copy_debug_stmt): Likewise for debug stmts.
25265 2013-01-28 Richard Biener <rguenther@suse.de>
25267 PR tree-optimization/56034
25268 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25269 (partition_builtin_p): Adjust.
25270 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25271 it is the last partition.
25272 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25273 up the vertex for the definition.
25274 (classify_partition): Classify whether a partition is a
25275 PKIND_REDUCTION, thus has uses outside of the loop.
25276 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25277 Merge all PKIND_REDUCTION partitions into the last partition.
25278 (tree_loop_distribution): Seed partitions from reductions as well.
25280 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25282 PR tree-optimization/56125
25283 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25284 pow(x,c) into sqrt(x) * powi(x, n/2) or
25285 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25286 optimizing for size.
25287 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25288 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25291 PR tree-optimization/56094
25292 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25293 to UNKNOWN_LOCATION while gimplifying expr.
25295 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25298 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25299 operand 0 in movabs insn template for -masm=intel asm alternative.
25300 (*movabs<mode>_2): Ditto for operand 1.
25302 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25305 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25308 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25310 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25313 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25315 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25316 Undef to avoid warning.
25318 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25320 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25321 * configure: Regenerate.
25323 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25325 PR tree-optimization/56098
25326 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25327 for stmts with volatile ops.
25328 (cond_store_replacement): Don't optimize if assign has volatile ops.
25329 (cond_if_else_store_replacement_1): Don't optimize if either
25330 then_assign or else_assign have volatile ops.
25331 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25334 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25336 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25338 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25340 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25341 missing ':' in asm example.
25343 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25345 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25346 entries into lane and laneq entries.
25347 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25348 Remove AdvSIMD scalar modes.
25349 (aarch64_sq<r>dmulh_laneq<mode>): New.
25350 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25352 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25353 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25354 * config/aarch64/iterators.md (VCOND): New.
25357 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25360 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25361 Add NULL LIBNAME argument to existing definitions.
25362 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25363 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25364 * config/avr/avr.c (DEF_BUILTIN): Same.
25365 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25366 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25367 implementation is available (DECL_ASSEMBLER_NAME is set).
25368 (avr_fold_absfx): New static function.
25369 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25370 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25371 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25372 AVR_BUILTIN_ABSLLK.
25373 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25374 (abshk, absk, abslk, absllk): Provide as static inline functions.
25376 2013-01-25 Marek Polacek <polacek@redhat.com>
25378 PR tree-optimization/56035
25379 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25381 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25383 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25384 (*movtf_internal_rex64): Add (!o,C) alternative
25385 (*movxf_internal_rex64): Ditto.
25386 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25388 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25390 * doc/invoke.texi: fix typo.
25391 * doc/objc.texi: fix typo.
25393 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25395 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25396 for the first two alternatives.
25398 2013-01-24 Diego Novillo <dnovillo@google.com>
25400 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25401 (ggc-zone.o): Remove.
25402 * configure.ac: Remove option --with-gc.
25403 * configure: Re-generate.
25404 * doc/install.texi: Remove documentation for --with-gc.
25405 * gengtype.c (write_enum_defn): Remove. Update all users.
25406 (write_Types_process_field): Remove generation of gt_e_* argument.
25407 (output_type_enum): Remove. Update all users.
25408 (write_enum_defn): Remove. Update all users.
25409 (enum alloc_zone): Remove. Update all users.
25410 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25411 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25412 Update all callers.
25413 (struct ptr_data): Remove field TYPE. Update all users.
25414 (gt_pch_note_object): Remove argument TYPE. Update all users.
25415 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25417 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25418 (struct alloc_zone): Remove.
25419 (ggc_internal_alloc_zone_stat): Remove.
25420 (ggc_internal_cleared_alloc_zone_stat): Remove.
25421 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25422 (ggc_pch_count_object): Remove last argument. Update all users.
25423 (ggc_pch_alloc_object): Remove last argument. Update all users.
25424 (struct alloc_zone): Remove.
25425 * ggc-zone.c: Remove.
25426 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25427 (struct alloc_zone): Remove.
25428 (ggc_alloc_typed_stat): Remove.
25429 (ggc_alloc_typed): Remove.
25430 (ggc_splay_alloc): Remove first argument.
25431 (rtl_zone): Remove. Update all users.
25432 (tree_zone): Remove. Update all users.
25433 (tree_id_zone): Remove. Update all users.
25434 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25435 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25436 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25437 * tree-ssanames.c: Remove references to zone allocator in comments.
25439 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25441 * config/avr/avr.c (avr_out_fract): Make register numbers that
25442 might be outside of source operand signed.
25444 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25446 * config/i386/constraints.md (Yf): New constraint.
25447 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25448 of f constraint to conditionaly disable x87 register preferences.
25449 (*movdf_internal): Ditto.
25450 (*movsf_internal): Ditto.
25452 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25454 PR inline-asm/55934
25455 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25456 that have operands with impossible constraints.
25457 Add a FIXME for a speed-up opportunity.
25458 * lra-constraints.c (process_alt_operands): Verify that a class
25459 selected from constraints on asms is valid for the operand mode.
25460 (curr_insn_transform): Remove incorrect comment.
25462 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25464 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25465 TOC operand is a valid symbol ref in the constant pool.
25467 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25469 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25471 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25474 * config/avr/stdfix.h: New file.
25475 * t-avr (stdfix-gcc.h): New rule to build it.
25476 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25478 2013-01-23 Kostya Serebryany <kcc@google.com>
25480 * config/darwin.h: remove dependency on
25481 CoreFoundation (asan on Mac OS).
25483 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25486 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25487 instead of cmpdi_operand for first comparison operand.
25488 Don't assert that comparison operands aren't both constants.
25490 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25492 * doc/install.texi (Downloading the Source): Update references to
25493 downloading separate components.
25495 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25497 * doc/extend.texi (__int128): Improve grammar.
25499 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25502 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25503 alternative to (o,r).
25504 (*movdi_internal_rex64): Remove (!o,n) alternative.
25505 (DImode immediate->memory splitter): Remove.
25506 (DImode immediate->memory peephole2): Remove.
25507 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25508 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25509 alternative to (!o,*r).
25510 (*movtf_internal_sse): New pattern.
25511 (*movxf_internal_rex64): New pattern.
25512 (*movxf_internal): Disable for TARGET_64BIT.
25513 (*movdf_internal_rex64): Remove (!o,F) alternative.
25515 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25517 PR middle-end/56074
25518 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25519 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25520 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25521 stmt locations where LOCATION_LOCUS of the stmt location is
25522 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25525 * config/i386/i386.md (UNSPEC_STOS): New.
25526 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25527 *strsetqi_1): Add UNSPEC_STOS.
25529 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25532 * doc/invoke.texi: Remove left over -Wsynth example.
25534 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25536 PR tree-optimization/56051
25537 * fold-const.c (fold_binary_loc): Don't fold
25538 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25539 a narrowing conversion, or widening conversion from signed
25542 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25544 PR rtl-optimization/56023
25545 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25546 dependent on debug instruction.
25548 2013-01-21 Martin Jambor <mjambor@suse.cz>
25550 PR middle-end/56022
25551 * function.c (allocate_struct_function): Call
25552 invoke_set_current_function_hook earlier.
25554 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25556 * reload1.c (init_reload): Only initialize reload_obstack
25557 during the first call.
25559 2013-01-21 Marek Polacek <polacek@redhat.com>
25561 * cfgloop.c (verify_loop_structure): Fix up grammar.
25563 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25565 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25566 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25568 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25571 * config/arm/marvell-pj4.md: Update copyright year.
25572 Fix up use of alu to alu_reg and simple_alu_imm.
25574 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25576 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25578 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25581 * lra-constraints.c (curr_insn_transform): Don't reuse original
25582 insn for secondary memory move when memory mode should be different.
25584 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25586 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25587 atomic_storedi_1): New patterns.
25589 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25591 btver2 pipeline descriptions.
25592 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25594 * config/i386/i386.md (btver2_decode): New type attributes.
25595 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25597 * config/i386/btver2.md: New file describing btver2 pipelines.
25599 2013-01-19 Andrew Pinski <apinski@cavium.com>
25601 PR tree-optimization/52631
25602 * tree-ssa-sccvn (visit_use): Before looking up the original
25603 statement, try looking up the simplified expression.
25605 2013-01-19 Anthony Green <green@moxielogic.com>
25607 * config/moxie/moxie.c (moxie_expand_prologue): Set
25608 current_function_static_stack_size.
25610 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25612 PR tree-optimization/56029
25613 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25614 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25616 2013-01-18 Sharad Singhai <singhai@google.com>
25618 PR tree-optimization/55995
25619 * dumpfile.c (dump_loc): Print location only if available.
25620 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25622 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25625 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25626 secondary memory move.
25627 (inherit_reload_reg): Use rclass instead of cl for
25628 check_secondary_memory_needed_p.
25630 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25632 PR middle-end/56015
25633 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25634 the case where writing real complex part of target modifies op1.
25636 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25638 * config/aarch64/aarch64-simd.md
25639 (aarch64_vcond_internal<mode>): Handle unordered cases.
25640 * config/aarch64/iterators.md (v_cmp_result): New.
25642 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25643 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25645 * config/arm/marvell-pj4.md: New file.
25646 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25647 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25648 (generic_vfp): Likewise.
25649 * config/arm/arm-cores.def: Add marvell-pj4.
25650 * config/arm/arm-tune.md: Regenerate.
25651 * config/arm/arm-tables.opt: Regenerate.
25652 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25653 * doc/invoke.texi: Document marvell-pj4.
25655 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25657 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25659 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25664 * var-tracking.c (negative_power_of_two_p): New.
25665 (global_get_addr_cache, local_get_addr_cache): New.
25666 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25667 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25669 (vt_stack_offset_p): Remove.
25670 (vt_canon_true_dep): Always canonicalize loc's address.
25671 (clobber_overlapping_mems): Make sure we have a MEM.
25672 (local_get_addr_clear_given_value): New.
25673 (val_reset): Clear local cached entries.
25674 (compute_bb_dataflow): Create and release the local cache.
25675 Disable duplicate MEMs clobbering.
25676 (emit_notes_in_bb): Clobber MEMs likewise.
25677 (vt_emit_notes): Create and release the local cache.
25678 (vt_initialize, vt_finalize): Create and release the global
25679 cache, respectively.
25680 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25682 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25684 PR libmudflap/53359
25685 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25686 not found in the symtab.
25688 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25691 PR rtl-optimization/55547
25692 PR rtl-optimization/53827
25695 * alias.c (offset_overlap_p): New, factored out of...
25696 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25697 the conservative special case for symbolic constants. Don't
25698 adjust zero sizes on alignment.
25700 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25702 PR rtl-optimization/52573
25703 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25704 REG_UNUSED for the same register.
25706 2013-01-17 Richard Biener <rguenther@suse.de>
25707 Marek Polacek <polacek@redhat.com>
25709 PR rtl-optimization/55833
25710 * loop-unswitch.c (unswitch_loops): Move loop verification...
25711 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25712 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25713 Set it to true when we're removing a loop from hierarchy tree in
25714 an irreducible region.
25715 (fix_bb_placements): Adjust caller.
25716 (fix_loop_placements): Likewise.
25718 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25720 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25721 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25722 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25723 Remove ID. Adjust comments.
25724 * config/avr/avr-c.c (avr_builtin_name): Remove.
25725 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25726 * config/avr/avr.c (avr_tolower): New static function.
25727 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25728 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25729 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25732 2013-01-17 Jan Hubicka <jh@suse.cz>
25734 PR tree-optimization/55273
25735 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25737 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25740 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25741 store through atomic_store<mode>_1.
25742 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25744 2013-01-17 Martin Jambor <mjambor@suse.cz>
25746 PR tree-optimizations/55264
25747 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25748 for virtual methods.
25749 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25750 virtual methods before inlining is over.
25751 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25753 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25756 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25758 PR rtl-optimization/56005
25759 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25760 pending reads for prefetch.
25762 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25764 * config/aarch64/aarch64.md
25765 (*cstoresi_neg_uxtw): New pattern.
25766 (*cmovsi_insn_uxtw): New pattern.
25767 (*<optab>si3_uxtw): New pattern.
25768 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25769 (*<optab>si3_insn_uxtw): New pattern.
25770 (*bswapsi2_uxtw): New pattern.
25772 2013-01-16 Richard Biener <rguenther@suse.de>
25774 * tree-inline.c (tree_function_versioning): Remove set but
25775 never used variable.
25777 2013-01-16 Richard Biener <rguenther@suse.de>
25779 PR tree-optimization/55964
25780 * tree-flow.h (rename_variables_in_loop): Remove.
25781 (rename_variables_in_bb): Likewise.
25782 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25783 (copy_loop_before): Adjust and delete update-ssa status.
25784 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25785 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25786 (rename_variables_in_loop): Remove.
25787 (slpeel_update_phis_for_duplicate_loop): Likewise.
25788 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25789 use available cfg machinery instead of duplicating it.
25790 Update PHI nodes and perform poor-mans SSA update here.
25791 (slpeel_tree_peel_loop_to_edge): Adjust.
25793 2013-01-16 Richard Biener <rguenther@suse.de>
25795 PR tree-optimization/54767
25796 PR tree-optimization/53465
25797 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25798 (vrp_visit_phi_node): For PHI arguments coming via backedges
25799 drop all symbolical range information.
25800 (execute_vrp): Compute backedges.
25802 2013-01-16 Richard Biener <rguenther@suse.de>
25804 * doc/install.texi: Update CLooG and ISL requirements to
25807 2013-01-16 Christian Bruel <christian.bruel@st.com>
25810 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25811 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25812 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25814 2013-01-16 DJ Delorie <dj@redhat.com>
25816 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25817 (UNSPECV_SP_SWITCH_E): New.
25818 (sp_switch_1): Change to an unspec.
25819 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25822 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25824 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25825 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25826 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25827 (expand_mem_thread_fence): Ditto.
25828 (expand_mem_signal_fence): Ditto.
25829 (expand_atomic_load): Ditto.
25830 (expand_atomic_store): Ditto.
25832 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25834 PR rtl-optimization/55547
25835 PR rtl-optimization/53827
25838 * alias.c (memrefs_conflict_p): Set sizes to negative after
25841 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25844 * function.c (thread_prologue_and_epilogue_insns): Always
25845 add crtl->drap_reg to set_up_by_prologue.set, even if
25846 stack_realign_drap is false.
25848 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25850 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25851 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25852 *call): Fix indention.
25854 2013-01-15 Tom de Vries <tom@codesourcery.com>
25857 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25860 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25862 PR rtl-optimization/55153
25863 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25865 2013-01-15 Martin Jambor <mjambor@suse.cz>
25867 PR tree-optimization/55920
25868 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25869 accesses as grp_to_be_debug_replaced.
25871 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25873 PR tree-optimization/55920
25874 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25875 there is non-useless type conversion needed from debug rhs to lhs,
25876 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25878 2013-01-15 Joseph Myers <joseph@codesourcery.com>
25879 Mikael Pettersson <mikpe@it.uu.se>
25882 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25884 (ASM_OUTPUT_CASE_LABEL): Remove.
25885 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25886 * final.c (shorten_branches): Update alignment of labels before
25887 jump tables if CASE_VECTOR_SHORTEN_MODE.
25889 2013-01-15 Richard Biener <rguenther@suse.de>
25892 * system.h: Do not include gmp.h for building host tools.
25894 2013-01-15 Richard Biener <rguenther@suse.de>
25896 PR middle-end/55882
25897 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25898 account for bitpos when computing alignment.
25900 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25902 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25903 (ix86_target_macros_internal): Likewise.
25905 * config/i386/i386.c (m_CORE2I7): Removed.
25906 (m_CORE_HASWELL): New macro.
25907 (m_CORE_ALL): Likewise.
25908 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25909 (initial_ix86_arch_features): Likewise.
25910 (processor_target_table): Initializations for Core avx2.
25911 (cpu_names): New names "core-avx2".
25912 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25913 PROCESSOR_CORE_HASWELL.
25914 (ix86_issue_rate): New case.
25915 (ia32_multipass_dfa_lookahead): Likewise.
25916 (ix86_sched_init_global): Likewise.
25918 * config/i386/i386.h (TARGET_HASWELL): New macro.
25919 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25920 (processor_type): New PROCESSOR_HASWELL.
25922 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25924 PR tree-optimization/55955
25925 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25926 *SHIFT_EXPR and *ROTATE_EXPR codes.
25928 PR tree-optimization/48766
25929 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25930 -ftrapv disable -fwrapv.
25932 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25935 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25936 etc. to 1 and not to __flash.
25937 Use LL suffix for __INT24_MAX__ with -mint8.
25938 Use ULL suffix for __UINT24_MAX__ with -mint8.
25940 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25942 * config/avr/avr-arch.h
25943 (struct base_arch_s): Use typedef avr_arch_t instead.
25944 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25945 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25946 * config/avr/avr.c: Same.
25947 * config/avr/avr-devices.c: Same.
25948 * config/avr/driver-avr.c: Same.
25949 * config/avr/gen-avr-mmcu-texi.c: Same.
25950 * config/avr/avr-mcus.def: Adjust comment.
25952 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25954 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25955 * config/aarch64/iterators.md (VALLDI): New.
25957 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
25958 Andi Kleen <ak@linux.intel.com>
25961 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25962 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25965 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25967 * config/avr/avr-stdint.h: Remove trailing blanks.
25968 * config/avr/avr-log.h: Same.
25969 * config/avr/avr-arch.h: Same.
25970 * config/avr/avr-devices.c: Same.
25971 * config/avr/avr-dimode.md: Same.
25972 * config/avr/predicates.md: Same.
25973 * config/avr/avr-c.c: Same. And fix typo.
25975 * config/avr/avr-protos.h: Same. And:
25976 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25977 (init_cumulative_args): Rename to avr_init_cumulative_args.
25978 (expand_prologue): Rename to avr_expand_prologue.
25979 (expand_epilogue): Rename to avr_expand_epilogue.
25980 (adjust_insn_length): Rename to avr_adjust_insn_length.
25981 (notice_update_cc): Rename to avr_notice_update_cc.
25982 (final_prescan_insn): Rename to avr_final_prescan_insn.
25983 * config/avr/avr.c: Same.
25984 * config/avr/avr.h: Same.
25985 * config/avr/avr.md: Remove trailing blanks.
25986 (prologue): Use avr_expand_prologue.
25987 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
25989 2013-01-14 Richard Biener <rguenther@suse.de>
25991 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
25992 verify_location, collect_subblocks): New functions.
25993 (verify_gimple_in_cfg): Verify that locations only reference
25994 BLOCKs in the functions BLOCK tree.
25996 2013-01-14 Richard Biener <rguenther@suse.de>
25998 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26000 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26002 (insert_out_of_ssa_copy_on_edge): Likewise.
26003 (rewrite_close_phi_out_of_ssa): Likewise.
26004 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26006 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26007 propagated constants.
26008 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26011 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26013 * config/avr/avr-modes.def: Add GPL copyright notice.
26015 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26017 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26018 MEMMODEL_MASK to determine memory model.
26019 (atomic_store<mode>): Ditto from operands[2].
26020 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26022 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26025 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26026 (fold_gimple_assign): Don't call unshare_expr here.
26027 (fold_ctor_reference): Call unshare_expr.
26029 2013-01-13 Terry Guo <terry.guo@arm.com>
26031 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26032 * doc/fragments.texi: Document MULTILIB_REUSE.
26033 * gcc.c (multilib_reuse): New internal spec.
26034 (set_multilib_dir): Also search multilib from multilib_reuse.
26035 * genmultilib (tmpmultilib3): Refactor code.
26036 (tmpmultilib4): Ditto.
26037 (multilib_reuse): New multilib argument.
26039 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26041 * Makefile.in: Update copyright.
26043 2013-01-12 Tom de Vries <tom@codesourcery.com>
26045 PR middle-end/55890
26046 * calls.c (expand_call): Check if arg_nr is valid.
26048 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26050 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26051 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26052 documentation. Add missing '__' in front of
26053 __builtin_ia32_packssdw256.
26055 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26058 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26059 NO_REGS for larl operands.
26060 (s390_reload_larl_operand): Use s390_load_address instead of
26063 2013-01-11 Richard Biener <rguenther@suse.de>
26065 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26066 (verify_node_sharing): ... here.
26067 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26069 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26071 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26072 Substitute TREECHECKING.
26073 * configure: Regenerate.
26074 * Makefile.in (TREECHECKING): New.
26076 2013-01-11 Richard Guenther <rguenther@suse.de>
26078 PR tree-optimization/44061
26079 * tree-vrp.c (extract_range_basic): Compute zero as
26080 value-range for __builtin_constant_p of function parameters.
26082 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26084 Update copyright years.
26086 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26088 PR rtl-optimization/55672
26089 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26090 const to be eliminable.
26092 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26094 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26095 * configure: Regenerate.
26097 2013-01-10 Richard Biener <rguenther@suse.de>
26099 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26101 2013-01-10 Richard Biener <rguenther@suse.de>
26104 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26105 locations for virtual PHI arguments.
26106 (rewrite_update_phi_arguments): Likewise.
26108 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26110 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26113 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26115 PR tree-optimization/55921
26116 * tree-complex.c (expand_complex_asm): New function.
26117 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26119 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26122 * config/s390/s390.c (s390_symref_operand_p)
26123 (s390_loadrelative_operand_p): Merge the two functions.
26124 (s390_check_qrst_address, print_operand_address): Add parameters
26125 to s390_loadrelative_operand_p invokation.
26126 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26127 (s390_reload_larl_operand, s390_secondary_reload): Use
26128 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26129 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26131 2013-01-09 Mike Stump <mikestump@comcast.net>
26133 * dse.c (record_store): Remove unnecessary assert.
26135 2013-01-09 Jan Hubicka <jh@suse.cz>
26137 PR tree-optimization/55569
26138 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26139 * cfgloop.h (scale_loop_profile): Likewise.
26141 2013-01-09 Jan Hubicka <jh@suse.cz>
26144 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26146 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26148 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26150 PR middle-end/55114
26151 * expr.h (maybe_emit_group_store): Declare.
26152 * expr.c (maybe_emit_group_store): New function.
26153 * builtins.c (expand_builtin_int_roundingfn): Call it.
26154 (expand_builtin_int_roundingfn_2): Likewise.
26156 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26158 PR rtl-optimization/55829
26159 * lra-constraints.c (match_reload): Add code for absent output.
26160 (curr_insn_transform): Add code for reloads of matched inputs
26163 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26165 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26166 attribute of movddup insn to DF.
26167 (*vec_interleave_lowv2df): Ditto.
26168 (vec_dupv2df): Ditto.
26170 2013-01-09 Jan Hubicka <jh@suse.cz>
26172 PR tree-optimiation/55875
26173 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26174 EVERY_ITERATION parameter.
26175 (number_of_iterations_exit): Check if exit is executed every iteration.
26176 (idx_infer_loop_bounds): Similarly here.
26177 (n_of_executions_at_most): Simplify
26178 to only test for cases where statement is dominated by the
26179 particular bound; handle correctly the "postdominance" test.
26180 (scev_probably_wraps_p): Use max loop iterations info
26181 as a global bound first.
26183 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26184 Nick Clifton <nickc@redhat.com>
26186 * config/v850/v850.md (cbranchsf4): New pattern.
26187 (cstoresf4): New pattern.
26188 (cbranchdf4): New pattern.
26189 (cstoredf4): New pattern.
26190 (movsicc): Disallow floating point comparisons.
26191 (cmpsf_le_insn): Fix order of operators.
26192 (cmpsf_lt_insn): Likewise.
26193 (cmpsf_eq_insn): Likewise.
26194 (cmpdf_le_insn): Likewise.
26195 (cmpdf_lt_insn): Likewise.
26196 (cmpdf_eq_insn): Likewise.
26197 (cmpsf_ge_insn): Use LE comparison.
26198 (cmpdf_ge_insn): Likewise.
26199 (cmpsf_gt_insn): Use LT comparison.
26200 (cmpdf_gt_insn): Likewise.
26201 (cmpsf_ne_insn): Delete pattern.
26202 (cmpdf_ne_insn): Delete pattern.
26203 * config/v850/v850.c (v850_gen_float_compare): Use
26204 gen_cmpdf_eq_insn for NE comparison.
26205 (v850_float_z_comparison_operator)
26206 (v850_float_nz_comparison_operator): Move from here ...
26207 * config/v850/predicates.md: ... to here. Move GT and GE
26208 comparisons into v850_float_z_comparison_operator.
26209 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26211 (v850_float_nz_comparison_operator): Likewise.
26213 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26215 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26216 with calls to gen_insvsi/gen_insvdi.
26218 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26220 * config/i386/i386.c (initial_ix86_tune_features): Set up
26221 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26223 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26224 Jakub Jelinek <jakub@redhat.com>
26226 PR tree-optimization/48189
26227 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26228 If nitercst is 0, don't predict the exit edge.
26230 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26232 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26233 in asm_fprintf with reg_names.
26234 (aarch64_print_operand_address): Likewise.
26235 (aarch64_return_addr): Likewise.
26236 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26238 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26240 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26241 (INT_U6_BITS): Likewise.
26242 * config/pa/predicates.md (uint6_operand): New predicate.
26243 (shift5_operand, shift6_operand): Likewise.
26244 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26246 (lshrdi3): Use shift6_operand.
26247 (shrpsi4, shrpdi4): New insn patterns.
26248 (extzv): Delete expander.
26249 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26250 predicates in unamed zero extract patterns. Tighten common constraint.
26251 (extv): Delete expander.
26252 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26253 predicates in unamed sign extract patterns. Tighten common constraint.
26254 (insv): Delete expander.
26255 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26256 predicates in unamed insert patterns. Tighten common constraint.
26257 Change uint32_operand predicate to uint6_operand predicate in unamed
26258 DImode pattern to insert constant values of type 1...1xxxx.
26260 2013-01-04 Jan Hubicka <jh@suse.cz>
26262 PR tree-optimization/55823
26263 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26266 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26267 Uros Bizjak <ubizjak@gmail.com>
26269 PR rtl-optimization/55845
26270 * df-problems.c (can_move_insns_across): Stop scanning at
26271 volatile_insn_p source instruction or give up if
26272 across_from .. across_to range contains any volatile_insn_p
26275 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26277 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26278 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26280 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26281 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26283 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26286 * asan.c (asan_clear_shadow): New function.
26287 (asan_emit_stack_protection): Use it.
26289 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26291 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26292 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26293 with tab instead of space.
26295 2013-01-08 Nick Clifton <nickc@redhat.com>
26297 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26298 register bank 0 at the start of an interrupt handler.
26299 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26302 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26304 * config/aarch64/aarch64-simd.md
26305 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26306 (aarch64_simd_bsl): Likewise.
26307 (aarch64_vcond_internal<mode>): Likewise.
26308 (vcond<mode><mode>): Likewise.
26309 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26310 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26312 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26314 * config/aarch64/aarch64-builtins.c
26315 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26317 2013-01-08 Martin Jambor <mjambor@suse.cz>
26320 * tree-sra.c (analyze_access_subtree): Return true also after
26321 potentially creating a debug-only replacement.
26323 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26325 PR middle-end/55890
26326 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26328 PR tree-optimization/54120
26329 * tree-vrp.c (range_fits_type_p): Don't allow
26330 src_precision < precision from signed vr to unsigned_p
26331 if vr->min or vr->max is negative.
26332 (simplify_float_conversion_using_ranges): Test can_float_p
26333 against CODE_FOR_nothing.
26335 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26336 Richard Biener <rguenther@suse.de>
26338 PR middle-end/55851
26339 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26340 types instead of just INTEGER_TYPE types.
26342 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26344 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26347 2013-01-07 Steve Ellcey <sellcey@mips.com>
26350 * config/mips/mips.opt: Change mad to mmad to match documentation.
26352 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26355 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26356 .progmemx.data now.
26358 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26361 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26362 (avr_addrspace_t): Add .section_name field.
26363 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26365 (avr_addrspace): Same. Initialize .section_name. Remove last
26366 NULL entry. Put __memx into .progmemx.data.
26367 (progmem_section_prefix): Remove.
26368 (avr_asm_init_sections): No need to initialize progmem_section.
26369 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26370 section name prefix.
26371 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26372 retrieve the progmem section.
26373 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26374 boundary to run over avr_addrspace[].
26375 (avr_register_target_pragmas): Ditto.
26377 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26379 * varasm.c (output_constant_def_contents): For asan_protect_global
26380 protected strings, adjust DECL_ALIGN if needed, before testing for
26382 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26383 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26385 (output_object_block): For asan protected decls, emit asan padding
26386 after their contents.
26387 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26388 (asan_finish_file): Test it here instead.
26390 2013-01-07 Nick Clifton <nickc@redhat.com>
26391 Matthias Klose <doko@debian.org>
26392 Doug Kwan <dougkwan@google.com>
26393 H.J. Lu <hongjiu.lu@intel.com>
26396 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26398 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26400 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26402 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26404 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26406 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26409 * doc/install.texi (Cross-Compiler-Specific Options): Document
26412 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26414 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26415 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26416 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26417 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26418 vqmovun_high_s64): Fix source operand number and update copyright.
26420 2013-01-07 Richard Biener <rguenther@suse.de>
26422 PR middle-end/55890
26423 * gimple.h (gimple_call_builtin_p): New overload.
26424 * gimple.c (validate_call): New function.
26425 (gimple_call_builtin_p): Likewise.
26426 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26427 Use gimple_call_builtin_p.
26428 (find_func_clobbers): Likewise.
26429 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26430 (strlen_optimize_stmt): Likewise.
26432 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26434 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26435 (vld1q_dup_*): Likewise.
26436 (vld1_*): Likewise.
26437 (vld1q_*): Likewise.
26438 (vld1_lane_*): Likewise.
26439 (vld1q_lane_*): Likewise.
26441 2013-01-07 Richard Biener <rguenther@suse.de>
26443 * lto-streamer.h (LTO_minor_version): Bump to 2.
26445 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26447 * config/aarch64/aarch64-protos.h
26448 (aarch64_const_double_zero_rtx_p): Rename to...
26449 (aarch64_float_const_zero_rtx_p): ...this.
26450 (aarch64_float_const_representable_p): New.
26451 (aarch64_output_simd_mov_immediate): Likewise.
26452 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26453 move immediate case.
26454 * config/aarch64/aarch64.c
26455 (aarch64_const_double_zero_rtx_p): Rename to...
26456 (aarch64_float_const_zero_rtx_p): ...this.
26457 (aarch64_print_operand): Allow printing of new constants.
26458 (aarch64_valid_floating_const): New.
26459 (aarch64_legitimate_constant_p): Check for valid floating-point
26461 (aarch64_simd_valid_immediate): Likewise.
26462 (aarch64_vect_float_const_representable_p): New.
26463 (aarch64_float_const_representable_p): Likewise.
26464 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26465 (aarch64_output_simd_mov_immediate): New.
26466 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26467 (*movdf_aarch64): Likewise.
26468 * config/aarch64/constraints.md (Ufc): New.
26469 (Y): call aarch64_float_const_zero_rtx.
26470 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26472 2013-01-07 Richard Biener <rguenther@suse.de>
26474 PR tree-optimization/55888
26475 PR tree-optimization/55862
26476 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26477 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26478 not if it is contained therein.
26480 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26482 * config/avr/t-avr: Typo.
26484 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26487 * config/avr/t-avr: Don't automatically rebuild
26488 $(srcdir)/config/avr/t-multilib
26489 $(srcdir)/config/avr/avr-tables.opt
26490 $(srcdir)/doc/avr-mmcu.texi
26491 (avr-mcus): New phony target to build them on request.
26492 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26493 * avr/avr-mcus.def: Adjust comments.
26495 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26497 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26499 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26501 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26503 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26505 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26507 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26509 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26510 to generate profiling.
26511 * config/rs6000/aix64.h (LIB_SPEC): Same.
26513 2013-01-04 Andrew Pinski <apinski@cavium.com>
26515 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26517 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26519 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26521 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26523 (ix86_expand_move): Ditto.
26524 (ix86_zero_extend_to_Pmode): Ditto.
26525 (ix86_expand_call): Ditto.
26526 (ix86_expand_special_args_builtin): Ditto.
26527 (ix86_expand_builtin): Ditto.
26529 2013-01-04 Richard Biener <rguenther@suse.de>
26531 PR tree-optimization/55862
26532 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26533 translating them through PHI nodes.
26535 2013-01-04 Martin Jambor <mjambor@suse.cz>
26537 PR tree-optimization/55755
26538 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26539 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26541 2013-01-04 Marek Polacek <polacek@redhat.com>
26543 PR middle-end/55859
26544 * opts.c (default_options_optimization): Clarify error message.
26546 2013-01-04 Richard Biener <rguenther@suse.de>
26548 PR middle-end/55863
26549 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26552 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26555 * config/pa/pa.md (movsi): Revert previous change.
26556 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26559 2013-01-03 Richard Henderson <rth@redhat.com>
26561 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26562 after eliminating TLS symbols.
26564 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26567 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26568 * graphite-poly.c (debug_gmp_value): Likewise.
26570 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26573 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26574 selected code model, define __code_mode_small__, __code_model_medium__,
26575 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26576 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26577 xchg temporary register with %k. Declare temporary register as
26579 [__x86_64__]: For medium and large code models, preserve %rbx register.
26581 2013-01-03 Richard Biener <rguenther@suse.de>
26583 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26584 (dump_subscript): Adjust.
26585 (finalize_ddr_dependent): Do not dump redundant info.
26586 (analyze_siv_subscript): Adjust.
26587 (subscript_dependence_tester): Likewise.
26588 (compute_affine_dependence): Likewise.
26590 2013-01-03 Richard Biener <rguenther@suse.de>
26593 2013-01-03 Richard Biener <rguenther@suse.de>
26595 PR tree-optimization/55857
26596 * tree-vect-stmts.c (vectorizable_load): Do not setup
26597 re-alignment for invariant loads.
26599 2013-01-02 Richard Biener <rguenther@suse.de>
26601 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26602 invariant load do not generate a vector load from the scalar location.
26604 2013-01-03 Richard Biener <rguenther@suse.de>
26606 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26607 for not vectorizing.
26608 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26609 not build INDIRECT_REFs, call get_name once only.
26610 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26611 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
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 2013-01-03 Richard Biener <rguenther@suse.de>
26622 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26623 prefer a built-in decl.
26625 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26627 * gcc.c (process_command): Update copyright notice dates.
26628 * gcov.c (print_version): Likewise.
26629 * gcov-dump.c (print_version): Likewise.
26631 PR rtl-optimization/55838
26632 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26633 iv0.step, iv1.step and step.
26635 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26636 Marc Glisse <marc.glisse@inria.fr>
26638 PR tree-optimization/55832
26639 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26640 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26641 integer_{one,zero}_node.
26643 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26646 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26647 * var-tracking.c (reverse_op): Don't add reverse ops to
26648 VALUEs that have already
26649 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26651 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26653 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26655 2013-01-02 Teresa Johnson <tejohnson@google.com>
26657 * dumpfile.c (dump_loc): Print filename with location.
26658 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26659 new location_t parameter to emit complete unroll message with
26660 new dump framework.
26661 (canonicalize_loop_induction_variables): Compute loops location
26662 and pass to try_unroll_loop_completely.
26663 * loop-unroll.c (report_unroll_peel): New function.
26664 (peel_loops_completely): Use new dump format with location
26665 for main dumpfile message, and invoke report_unroll_peel on success.
26666 (decide_unrolling_and_peeling): Ditto.
26667 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26668 by report_unroll_peel.
26669 (decide_peel_completely): Ditto.
26670 (decide_unroll_constant_iterations): Ditto.
26671 (decide_unroll_runtime_iterations): Ditto.
26672 (decide_peel_simple): Ditto.
26673 (decide_unroll_stupid): Ditto.
26674 * cfgloop.c (get_loop_location): New function.
26675 * cfgloop.h (get_loop_location): Declare.
26677 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26679 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26682 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26684 PR middle-end/55198
26685 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26686 BLKmode objects when EXPAND_MEMORY is specified.
26688 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26690 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26692 (fold_builtin_cpu): Do not share cpu model decls across statements.
26694 2013-01-02 Jason Merrill <jason@redhat.com>
26697 * tree.c (build_array_type_1): Revert earlier change.
26699 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26701 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26703 * config/aarch64/aarch64-tune.md: Re-generate.
26705 2013-01-02 Richard Biener <rguenther@suse.de>
26707 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26708 invariant load do not generate a vector load from the scalar location.
26710 2013-01-02 Richard Biener <rguenther@suse.de>
26713 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26714 * configure: Regenerate.
26716 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26718 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26719 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26720 (expand_builtin_int_roundingfn_2): Keep the original target around
26721 for the fallback case.
26723 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26725 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26726 to be clear for sign changes.
26728 2013-01-01 Jan Hubicka <jh@suse.cz>
26730 * ipa-inline-analysis.c: Fix formatting.
26732 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26734 PR tree-optimization/55831
26735 * tree-vect-loop.c (get_initial_def_for_induction): Use
26736 gsi_after_labels instead of gsi_start_bb.
26738 Copyright (C) 2013 Free Software Foundation, Inc.
26740 Copying and distribution of this file, with or without modification,
26741 are permitted in any medium without royalty provided the copyright
26742 notice and this notice are preserved.