1 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
6 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
8 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
9 (MULTILIB_MATCHES): Add multilib for -march=v8.
11 2013-10-15 Sriraman Tallam <tmsriram@google.com>
14 * optc-save-gen.awk: Add extra parameter to the save and restore
16 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
17 * tree.c (build_optimization_node): New parameter. Add extra parameter
18 to call to cl_optimization_save.
19 (build_target_option_node): New parameter. Add extra parameter
20 to call to cl_target_option_save.
21 * tree.h (build_optimization_node): New parameter.
22 (build_target_option_node): New parameter.
23 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
24 build_optimization_node and build_target_option_node.
25 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
26 (handle_pragma_push_options): Ditto.
27 * toplev.c (process_options): Ditto.
28 * opts.c (init_options_struct): Check for opts_set non-null.
29 * target.def (target_option.save): New parameter.
30 (target_option.restore): New parameter.
32 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
33 (ix86_pragma_target_parse): Ditto.
34 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
36 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
37 to build_optimization_node and build_target_option_node.
38 (rs6000_valid_attribute_p): Ditto.
39 (rs6000_pragma_target_parse): Ditto.
40 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
43 TARGET_64BIT_P: New Macro
44 TARGET_MMX_P: New Macro.
45 TARGET_3DNOW_P: New Macro.
46 TARGET_3DNOW_A_P: New Macro.
47 TARGET_SSE_P: New Macro.
48 TARGET_SSE2_P: New Macro.
49 TARGET_SSE3_P: New Macro.
50 TARGET_SSSE3_P: New Macro.
51 TARGET_SSE4_1_P: New Macro.
52 TARGET_SSE4_2_P: New Macro.
53 TARGET_AVX_P: New Macro.
54 TARGET_AVX2_P: New Macro.
55 TARGET_AVX512F_P: New Macro.
56 TARGET_AVX512PF_P: New Macro.
57 TARGET_AVX512ER_P: New Macro.
58 TARGET_AVX512CD_P: New Macro.
59 TARGET_FMA_P: New Macro.
60 TARGET_SSE4A_P: New Macro.
61 TARGET_FMA4_P: New Macro.
62 TARGET_XOP_P: New Macro.
63 TARGET_LWP_P: New Macro.
64 TARGET_ABM_P: New Macro.
65 TARGET_BMI_P: New Macro.
66 TARGET_BMI2_P: New Macro.
67 TARGET_LZCNT_P: New Macro.
68 TARGET_TBM_P: New Macro.
69 TARGET_POPCNT_P: New Macro.
70 TARGET_SAHF_P: New Macro.
71 TARGET_MOVBE_P: New Macro.
72 TARGET_CRC32_P: New Macro.
73 TARGET_AES_P: New Macro.
74 TARGET_PCLMUL_P: New Macro.
75 TARGET_CMPXCHG16B_P: New Macro.
76 TARGET_FSGSBASE_P: New Macro.
77 TARGET_RDRND_P: New Macro.
78 TARGET_F16C_P: New Macro.
79 TARGET_RTM_P: New Macro.
80 TARGET_HLE_P: New Macro.
81 TARGET_RDSEED_P: New Macro.
82 TARGET_PRFCHW_P: New Macro.
83 TARGET_ADX_P: New Macro.
84 TARGET_FXSR_P: New Macro.
85 TARGET_XSAVE_P: New Macro.
86 TARGET_XSAVEOPT_P: New Macro.
87 TARGET_LP64_P: New Macro.
88 TARGET_X32_P: New Macro.
89 TARGET_FPMATH_DEFAULT_P: New Macro.
90 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
91 * config/i386/i386.c (ix86_option_override_internal): New parameters.
93 Change ix86_tune_string to access opts->x_ix86_tune_string.
94 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
95 Change ix86_arch_string to access opts->x_ix86_arch_string.
96 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
97 Change ix86_pmode to access opts->x_ix86_pmode.
98 Change ix86_abi to access opts->x_ix86_abi.
99 Change ix86_cmodel to access opts->x_ix86_cmodel.
100 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
101 Change ix86_isa_flags_explicit to access
102 opts->x_ix86_isa_flags_explicit.
103 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
104 Change ix86_regparm to access opts->x_ix86_regparm.
105 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
106 Change ix86_preferred_stack_boundary_arg to access
107 opts->x_ix86_preferred_stack_boundary_arg.
108 Change ix86_force_align_arg_pointer to access
109 opts->x_ix86_force_align_arg_pointer.
110 Change ix86_incoming_stack_boundar_arg to access
111 opts->x_ix86_incoming_stack_boundar_arg.
112 Change ix86_fpmath to access opts->x_ix86_fpmath.
113 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
114 Change ix86_recip_name to access opts->x_ix86_recip_name.
115 Change ix86_stack_protector_guard to access
116 opts->x_ix86_stack_protector_guard.
117 Change ix86_tune_memcpy_strategy to access
118 opts->x_ix86_tune_memcpy_strategy.
119 Change ix86_tune_memset_strategy to access
120 opts->x_ix86_tune_memset_strategy.
121 Change global_options to access opts.
122 Change global_options_set to access opts_set.
123 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
124 Change TARGET_MMX to TARGET_MMX_P (opts->...)
125 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...)
126 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...)
127 Change TARGET_SSE to TARGET_SSE_P (opts->...)
128 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
129 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...)
130 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...)
131 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...)
132 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...)
133 Change TARGET_AVX to TARGET_AVX_P (opts->...)
134 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...)
135 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...)
136 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...)
137 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...)
138 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...)
139 Change TARGET_FMA to TARGET_FMA_P (opts->...)
140 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...)
141 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...)
142 Change TARGET_XOP to TARGET_XOP_P (opts->...)
143 Change TARGET_LWP to TARGET_LWP_P (opts->...)
144 Change TARGET_ABM to TARGET_ABM_P (opts->...)
145 Change TARGET_BMI to TARGET_BMI_P (opts->...)
146 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...)
147 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...)
148 Change TARGET_TBM to TARGET_TBM_P (opts->...)
149 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...)
150 Change TARGET_SAHF to TARGET_SAHF_P (opts->...)
151 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...)
152 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...)
153 Change TARGET_AES to TARGET_AES_P (opts->...)
154 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...)
155 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...)
156 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...)
157 Change TARGET_RDRND to TARGET_RDRND_P (opts->...)
158 Change TARGET_F16C to TARGET_F16C_P (opts->...)
159 Change TARGET_RTM to TARGET_RTM_P (opts->...)
160 Change TARGET_HLE to TARGET_HLE_P (opts->...)
161 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...)
162 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...)
163 Change TARGET_ADX to TARGET_ADX_P (opts->...)
164 Change TARGET_FXSR to TARGET_FXSR_P (opts->...)
165 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...)
166 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...)
167 Change TARGET_LP64 to TARGET_LP64_P (opts->...)
168 Change TARGET_X32 to TARGET_X32_P (opts->...)
169 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...)
170 Change TARGET_FLOAT_RETURNS_IN_80387 to
171 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...)
172 (ix86_function_specific_save): New parameter. Use opts-> fields
173 to replace global fields.
174 (ix86_function_specific_restore): Ditto.
175 (ix86_valid_target_attribute_inner_p): New parameters.
177 Fix call to ix86_handle_option and set_option.
178 (ix86_valid_target_attribute_tree): New parameters.
179 Change global_options to access opts.
180 Change global_options_set to access opts_set.
181 Fix call to ix86_valid_target_attribute_inner_p.
182 Change ix86_tune_string to access opts->x_ix86_tune_string.
183 Change ix86_arch_string to access opts->x_ix86_arch_string.
184 Change ix86_fpmath to access opts->x_ix86_fpmath
185 Fix call to ix86_option_override_internal.
186 Fix call to ix86_add_new_builtins.
187 Fix calls to build_optimization_node and build_target_option_node.
188 (ix86_valid_target_attribute_p): Remove access to global_options.
189 Use new gcc_options structure func_options.
190 Fix call to ix86_valid_target_attribute_tree.
191 Fix call to build_optimization_node.
192 (get_builtin_code_for_version): Fix call to
193 ix86_valid_target_attribute_tree.
195 2013-10-15 David Malcolm <dmalcolm@redhat.com>
197 * Makefile.in (PICFLAG): New.
198 (enable_host_shared): New.
199 (INTERNAL_CFLAGS): Use PICFLAG.
200 (LIBIBERTY): Use pic build of libiberty.a if configured with
201 --enable-host-shared.
202 * configure.ac: Add --enable-host-shared, setting up new
204 * configure: Regenerate.
205 * doc/install.texi (--enable-shared): Add note contrasting it
207 (--enable-host-shared): New option.
209 2013-10-15 Richard Biener <rguenther@suse.de>
211 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
212 for built-in functions.
214 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
216 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
217 (optimize_range_tests_1): New function,
218 extracted from optimize_range_tests.
219 (optimize_range_tests_xor): Similarly.
220 (optimize_range_tests_diff): New function.
221 (optimize_range_tests): Use optimize_range_tests_1.
223 2013-10-15 Cong Hou <congh@google.com>
225 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
226 requirement of the reduction pattern so that one operand of the
227 reduction operation can come from outside of the loop.
229 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
231 * config/arm/neon-schedgen.ml: Remove.
232 * config/arm/cortex-a9-neon.md: Remove comment regarding
235 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
237 * config/arm/types: Remove old neon types.
239 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
241 * config/arm/cortex-a7.md
242 (cortex_a7_neon_type): New.
243 (cortex_a7_neon_mul): Update for new types.
244 (cortex_a7_neon_mla): Likewise.
245 (cortex_a7_neon): Likewise.
247 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
249 * config/arm/cortex-a15-neon.md
250 (cortex_a15_neon_type): New,
252 (cortex_a15_neon_int_1): Remove.
253 (cortex_a15_neon_int_2): Likewise.
254 (cortex_a15_neon_int_3): Likewise.
255 (cortex_a15_neon_int_4): Likewise.
256 (cortex_a15_neon_int_5): Likewise.
257 (cortex_a15_neon_vqneg_vqabs): Likewise.
258 (cortex_a15_neon_vmov): Likewise.
259 (cortex_a15_neon_vaba): Likewise.
260 (cortex_a15_neon_vaba_qqq): Likewise.
261 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
262 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
263 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
265 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
266 (cortex_a15_neon_mla_qqq_8_16): Likewise.
267 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
268 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
269 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
270 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
271 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
272 (cortex_a15_neon_shift_1): Likewise.
273 (cortex_a15_neon_shift_2): Likewise.
274 (cortex_a15_neon_shift_3): Likewise.
275 (cortex_a15_neon_vshl_ddd): Likewise.
276 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
277 (cortex_a15_neon_vsra_vrsra): Likewise.
278 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
279 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
280 (cortex_a15_neon_bp_3cycle): Likewise.
281 (cortex_a15_neon_ldm_2): Likewise.
282 (cortex_a15_neon_stm_2): Likewise.
283 (cortex_a15_neon_mcr): Likewise.
284 (cortex_a15_neon_mrc): Likewise.
285 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
286 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
287 (cortex_a15_neon_fp_vmul_ddd): Likewise.
288 (cortex_a15_neon_fp_vmul_qqd): Likewise.
289 (cortex_a15_neon_fp_vmla_ddd): Likewise.
290 (cortex_a15_neon_fp_vmla_qqq): Likewise.
291 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
292 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
293 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
294 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
295 (cortex_a15_neon_bp_simple): Likewise.
296 (cortex_a15_neon_bp_2cycle): Likewise.
297 (cortex_a15_neon_bp_3cycle): Likewise.
298 (cortex_a15_neon_vld1_1_2_regs): Likewise.
299 (cortex_a15_neon_vld1_3_4_regs): Likewise.
300 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
301 (cortex_a15_neon_vld2_4_regs): Likewise.
302 (cortex_a15_neon_vld3_vld4): Likewise.
303 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
304 (cortex_a15_neon_vst1_3_4_regs): Likewise.
305 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
306 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
307 (cortex_a15_neon_vst3_vst4): Rename to...
308 (cortex_a15_neon_vst4): This, update for new attributes.
309 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
310 (cortex_a15_neon_vld3_vld4_lane): Likewise.
311 (cortex_a15_neon_vst1_vst2_lane): Likewise.
312 (cortex_a15_neon_vst3_vst4_lane): Likewise.
313 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
314 (cortex_a15_neon_ldm_2): Likewise.
315 (cortex_a15_neon_stm_2): Likewise.
316 (cortex_a15_neon_mcr): Likewise.
317 (cortex_a15_neon_mcr_2_mcrr): Likewise.
318 (cortex_a15_neon_mrc): Likewise.
319 (cortex_a15_neon_mrrc): Likewise.
321 (cortex_a15_neon_abd): New.
322 (cortex_a15_neon_abd_q): Likewise.
323 (cortex_a15_neon_aba): Likewise.
324 (cortex_a15_neon_aba_q): Likewise.
325 (cortex_a15_neon_acc): Likewise.
326 (cortex_a15_neon_acc_q): Likewise.
327 (cortex_a15_neon_arith_basic): Likewise.
328 (cortex_a15_neon_arith_complex): Likewise.
329 (cortex_a15_neon_multiply): Likewise.
330 (cortex_a15_neon_multiply_q): Likewise.
331 (cortex_a15_neon_mla): Likewise.
332 (cortex_a15_neon_mla_q): Likewise.
333 (cortex_a15_neon_sat_mla_long): Likewise.
334 (cortex_a15_neon_shift_acc): Likewise.
335 (cortex_a15_neon_shift_imm_basic): Likewise.
336 (cortex_a15_neon_shift_imm_complex): Likewise.
337 (cortex_a15_neon_shift_reg_basic): Likewise.
338 (cortex_a15_neon_shift_reg_basic_q): Likewise.
339 (cortex_a15_neon_shift_reg_complex): Likewise.
340 (cortex_a15_neon_shift_reg_complex_q): Likewise.
341 (cortex_a15_neon_fp_negabs): Likewise
342 (cortex_a15_neon_fp_arith): Likewise
343 (cortex_a15_neon_fp_arith_q): Likewise
344 (cortex_a15_neon_fp_cvt_int): Likewise
345 (cortex_a15_neon_fp_cvt_int_q): Likewise
346 (cortex_a15_neon_fp_cvt_16): Likewise
347 (cortex_a15_neon_fp_mul): Likewise
348 (cortex_a15_neon_fp_mul_q): Likewise
349 (cortex_a15_neon_fp_mla): Likewise
350 (cortex_a15_neon_fp_mla_q): Likewise
351 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
352 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
353 (cortex_a15_neon_bitops): Likewise.
354 (cortex_a15_neon_bitops_q): Likewise.
355 (cortex_a15_neon_from_gp): Likewise.
356 (cortex_a15_neon_from_gp_q): Likewise.
357 (cortex_a15_neon_tbl3_tbl4): Likewise.
358 (cortex_a15_neon_zip_q): Likewise.
359 (cortex_a15_neon_to_gp): Likewise.
360 (cortex_a15_neon_load_a): Likewise.
361 (cortex_a15_neon_load_b): Likewise.
362 (cortex_a15_neon_load_c): Likewise.
363 (cortex_a15_neon_load_d): Likewise.
364 (cortex_a15_neon_load_e): Likewise.
365 (cortex_a15_neon_load_f): Likewise.
366 (cortex_a15_neon_store_a): Likewise.
367 (cortex_a15_neon_store_b): Likewise.
368 (cortex_a15_neon_store_c): Likewise.
369 (cortex_a15_neon_store_d): Likewise.
370 (cortex_a15_neon_store_e): Likewise.
371 (cortex_a15_neon_store_f): Likewise.
372 (cortex_a15_neon_store_g): Likewise.
373 (cortex_a15_neon_store_h): Likewise.
374 (cortex_a15_vfp_to_from_gp): Likewise.
376 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
378 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
380 (cortex_a9_neon_vshl_ddd): Remove.
381 (cortex_a9_neon_vst3_vst4): Likewise.
382 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
384 (cortex_a9_neon_bit_ops_q): New.
386 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
387 (cortex_a9_neon_int_2): Likewise.
388 (cortex_a9_neon_int_3): Likewise.
389 (cortex_a9_neon_int_4): Likewise.
390 (cortex_a9_neon_int_5): Likewise.
391 (cortex_a9_neon_vqneg_vqabs): Likewise.
392 (cortex_a9_neon_vmov): Likewise.
393 (cortex_a9_neon_vaba): Likewise.
394 (cortex_a9_neon_vaba_qqq): Likewise.
395 (cortex_a9_neon_shift_1): Likewise.
396 (cortex_a9_neon_shift_2): Likewise.
397 (cortex_a9_neon_shift_3): Likewise.
398 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
399 (cortex_a9_neon_vsra_vrsra): Likewise.
400 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
401 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
402 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
404 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
405 (cortex_a9_neon_mla_qqq_8_16): Likewise.
406 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
408 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
409 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
410 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
411 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
412 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
413 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
414 (cortex_a9_neon_fp_vsum): Likewise.
415 (cortex_a9_neon_fp_vmul_ddd): Likewise.
416 (cortex_a9_neon_fp_vmul_qqd): Likewise.
417 (cortex_a9_neon_fp_vmla_ddd): Likewise.
418 (cortex_a9_neon_fp_vmla_qqq): Likewise.
419 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
420 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
421 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
422 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
423 (cortex_a9_neon_bp_simple): Likewise.
424 (cortex_a9_neon_bp_2cycle): Likewise.
425 (cortex_a9_neon_bp_3cycle): Likewise.
426 (cortex_a9_neon_ldr): Likewise.
427 (cortex_a9_neon_str): Likewise.
428 (cortex_a9_neon_vld1_1_2_regs): Likewise.
429 (cortex_a9_neon_vld1_3_4_regs): Likewise.
430 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
431 (cortex_a9_neon_vld2_4_regs): Likewise.
432 (cortex_a9_neon_vld3_vld4): Likewise.
433 (cortex_a9_neon_vld1_vld2_lane): Likewise.
434 (cortex_a9_neon_vld3_vld4_lane): Likewise.
435 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
436 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
437 (cortex_a9_neon_vst1_3_4_regs): Likewise.
438 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
439 (cortex_a9_neon_vst1_vst2_lane): Likewise.
440 (cortex_a9_neon_vst3_vst4_lane): Likewise.
441 (cortex_a9_neon_mcr): Likewise.
442 (cortex_a9_neon_mcr_2_mcrr): Likewise.
443 (cortex_a9_neon_mrc): Likewise.
444 (cortex_a9_neon_mrrc): Likewise.
446 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
448 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
450 (cortex_a8_neon_vshl_ddd): Remove.
451 (cortex_a8_neon_vst3_vst4): Likewise.
452 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
454 (cortex_a8_neon_bit_ops_q): New.
456 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
457 (cortex_a8_neon_int_2): Likewise..
458 (cortex_a8_neon_int_3): Likewise.
459 (cortex_a8_neon_int_5): Likewise.
460 (cortex_a8_neon_vqneg_vqabs): Likewise.
461 (cortex_a8_neon_int_4): Likewise.
462 (cortex_a8_neon_vaba): Likewise.
463 (cortex_a8_neon_vaba_qqq): Likewise.
464 (cortex_a8_neon_shift_1): Likewise.
465 (cortex_a8_neon_shift_2): Likewise.
466 (cortex_a8_neon_shift_3): Likewise.
467 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
468 (cortex_a8_neon_vsra_vrsra): Likewise.
469 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
470 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
471 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
473 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
474 (cortex_a8_neon_mla_qqq_8_16): Likewise.
475 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
477 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
478 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
479 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
480 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
481 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
482 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
483 (cortex_a8_neon_fp_vsum): Likewise.
484 (cortex_a8_neon_fp_vmul_ddd): Likewise.
485 (cortex_a8_neon_fp_vmul_qqd): Likewise.
486 (cortex_a8_neon_fp_vmla_ddd): Likewise.
487 (cortex_a8_neon_fp_vmla_qqq): Likewise.
488 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
489 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
490 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
491 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
492 (cortex_a8_neon_bp_simple): Likewise.
493 (cortex_a8_neon_bp_2cycle): Likewise.
494 (cortex_a8_neon_bp_3cycle): Likewise.
495 (cortex_a8_neon_ldr): Likewise.
496 (cortex_a8_neon_str): Likewise.
497 (cortex_a8_neon_vld1_1_2_regs): Likewise.
498 (cortex_a8_neon_vld1_3_4_regs): Likewise.
499 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
500 (cortex_a8_neon_vld2_4_regs): Likewise.
501 (cortex_a8_neon_vld3_vld4): Likewise.
502 (cortex_a8_neon_vld1_vld2_lane): Likewise.
503 (cortex_a8_neon_vld3_vld4_lane): Likewise.
504 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
505 (cortex_a8_neon_vst1_3_4_regs): Likewise.
506 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
507 (cortex_a8_neon_vst1_vst2_lane): Likewise.
508 (cortex_a8_neon_vst3_vst4_lane): Likewise.
509 (cortex_a8_neon_mcr): Likewise.
510 (cortex_a8_neon_mcr_2_mcrr): Likewise.
511 (cortex_a8_neon_mrc): Likewise.
512 (cortex_a8_neon_mrrc): Likewise.
514 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
516 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
518 * config/aarch64/aarch64-simd.md (neon_type): Remove.
519 (aarch64_simd_dup<mode>): Add "type" attribute.
520 (aarch64_dup_lane<mode>): Likewise.
521 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
522 (*aarch64_simd_mov<mode>): Likewise.
523 (aarch64_simd_mov_from_<mode>low): Likewise.
524 (aarch64_simd_mov_from_<mode>high): Likewise.
525 (orn<mode>3): Likewise.
526 (bic<mode>3): Likewise.
527 (add<mode>3): Likewise.
528 (sub<mode>3): Likewise.
529 (mul<mode>3): Likewise.
530 (*aarch64_mul3_elt<mode>): Likewise.
531 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
532 (*aarch64_mul3_elt_to_128df): Likewise.
533 (*aarch64_mul3_elt_to_64v2df): Likewise.
534 (neg<mode>2): Likewise.
535 (abs<mode>2): Likewise.
536 (abd<mode>_3): Likewise.
537 (aba<mode>_3): Likewise.
538 (fabd<mode>_3): Likewise.
539 (*fabd_scalar<mode>3): Likewise.
540 (and<mode>3): Likewise.
541 (ior<mode>3): Likewise.
542 (xor<mode>3): Likewise.
543 (one_cmpl<mode>2): Likewise.
544 (aarch64_simd_vec_set<mode>): Likewise.
545 (aarch64_simd_lshr<mode>): Likewise.
546 (aarch64_simd_ashr<mode>): Likewise.
547 (aarch64_simd_imm_shl<mode>): Likewise.
548 (aarch64_simd_reg_sshl<mode): Likewise.
549 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
550 (aarch64_simd_reg_shl<mode>_signed): Likewise.
551 (aarch64_simd_vec_setv2di): Likewise.
552 (aarch64_simd_vec_set<mode>): Likewise.
553 (aarch64_mla<mode>): Likewise.
554 (*aarch64_mla_elt<mode>): Likewise.
555 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
556 (aarch64_mls<mode>): Likewise.
557 (*aarch64_mls_elt<mode>): Likewise.
558 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
559 (<su><maxmin><mode>3): Likewise.
560 (move_lo_quad_<mode>): Likewise.
561 (aarch64_simd_move_hi_quad_<mode>): Likewise.
562 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
563 (vec_pack_trunc_<mode>): Likewise.
564 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
565 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
566 (*aarch64_<su>mlal_lo<mode>): Likewise.
567 (*aarch64_<su>mlal_hi<mode>): Likewise.
568 (*aarch64_<su>mlsl_lo<mode>): Likewise.
569 (*aarch64_<su>mlsl_hi<mode>): Likewise.
570 (*aarch64_<su>mlal<mode>): Likewise.
571 (*aarch64_<su>mlsl<mode>): Likewise.
572 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
573 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
574 (add<mode>3): Likewise.
575 (sub<mode>3): Likewise.
576 (mul<mode>3): Likewise.
577 (div<mode>3): Likewise.
578 (neg<mode>2): Likewise.
579 (abs<mode>2): Likewise.
580 (fma<mode>4): Likewise.
581 (*aarch64_fma4_elt<mode>): Likewise.
582 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
583 (*aarch64_fma4_elt_to_128df): Likewise.
584 (*aarch64_fma4_elt_to_64v2df): Likewise.
585 (fnma<mode>4): Likewise.
586 (*aarch64_fnma4_elt<mode>): Likewise.
587 (*aarch64_fnma4_elt_<vswap_width_name><mode>
588 (*aarch64_fnma4_elt_to_128df): Likewise.
589 (*aarch64_fnma4_elt_to_64v2df): Likewise.
590 (<frint_pattern><mode>2): Likewise.
591 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
592 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
593 (vec_unpacks_lo_v4sf): Likewise.
594 (aarch64_float_extend_lo_v2df): Likewise.
595 (vec_unpacks_hi_v4sf): Likewise.
596 (aarch64_float_truncate_lo_v2sf): Likewise.
597 (aarch64_float_truncate_hi_v4sf): Likewise.
598 (aarch64_vmls<mode>): Likewise.
599 (<su><maxmin><mode>3): Likewise.
600 (<maxmin_uns><mode>3): Likewise.
601 (reduc_<sur>plus_<mode>): Likewise.
602 (reduc_<sur>plus_v2di): Likewise.
603 (reduc_<sur>plus_v2si): Likewise.
604 (reduc_<sur>plus_<mode>): Likewise.
605 (aarch64_addpv4sf): Likewise.
606 (clz<mode>2): Likewise.
607 (reduc_<maxmin_uns>_<mode>): Likewise.
608 (reduc_<maxmin_uns>_v2di): Likewise.
609 (reduc_<maxmin_uns>_v2si): Likewise.
610 (reduc_<maxmin_uns>_<mode>): Likewise.
611 (reduc_<maxmin_uns>_v4sf): Likewise.
612 (aarch64_simd_bsl<mode>_internal): Likewise.
613 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
614 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
615 (aarch64_get_lane<mode>): Likewise.
616 (*aarch64_combinez<mode>): Likewise.
617 (aarch64_combine<mode>): Likewise.
618 (aarch64_simd_combine<mode>): Likewise.
619 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
620 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
621 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
622 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
623 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
624 (aarch64_<sur>h<addsub><mode>): Likewise.
625 (aarch64_<sur><addsub>hn<mode>): Likewise.
626 (aarch64_<sur><addsub>hn2<mode>): Likewise.
627 (aarch64_pmul<mode>): Likewise.
628 (aarch64_<su_optab><optab><mode>): Likewise.
629 (aarch64_<sur>qadd<mode>): Likewise.
630 (aarch64_sqmovun<mode>): Likewise.
631 (aarch64_<sur>qmovn<mode>): Likewise.
632 (aarch64_s<optab><mode>): Likewise.
633 (aarch64_sq<r>dmulh<mode>): Likewise.
634 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
635 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
636 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
637 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
638 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
639 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
640 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
641 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
642 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
643 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
644 (aarch64_sqdmull<mode>): Likewise.
645 (aarch64_sqdmull_lane<mode>_internal): Likewise.
646 (aarch64_sqdmull_n<mode>): Likewise.
647 (aarch64_sqdmull2<mode>_internal): Likewise.
648 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
649 (aarch64_sqdmull2_n<mode>_internal): Likewise.
650 (aarch64_<sur>shl<mode>): Likewise.
651 (aarch64_<sur>q<r>shl<mode>
652 (aarch64_<sur>shll_n<mode>): Likewise.
653 (aarch64_<sur>shll2_n<mode>): Likewise.
654 (aarch64_<sur>shr_n<mode>): Likewise.
655 (aarch64_<sur>sra_n<mode>): Likewise.
656 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
657 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
658 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
659 (aarch64_cm<optab><mode>): Likewise.
660 (aarch64_cm<optab>di): Likewise.
661 (aarch64_cm<optab><mode>): Likewise.
662 (aarch64_cm<optab>di): Likewise.
663 (aarch64_cmtst<mode>): Likewise.
664 (aarch64_cmtstdi): Likewise.
665 (aarch64_cm<optab><mode>): Likewise.
666 (*aarch64_fac<optab><mode>): Likewise.
667 (aarch64_addp<mode>): Likewise.
668 (aarch64_addpdi): Likewise.
669 (sqrt<mode>2): Likewise.
670 (vec_load_lanesoi<mode>): Likewise.
671 (vec_store_lanesoi<mode>): Likewise.
672 (vec_load_lanesci<mode>): Likewise.
673 (vec_store_lanesci<mode>): Likewise.
674 (vec_load_lanesxi<mode>): Likewise.
675 (vec_store_lanesxi<mode>): Likewise.
676 (*aarch64_mov<mode>): Likewise.
677 (aarch64_ld2<mode>_dreg): Likewise.
678 (aarch64_ld2<mode>_dreg): Likewise.
679 (aarch64_ld3<mode>_dreg): Likewise.
680 (aarch64_ld3<mode>_dreg): Likewise.
681 (aarch64_ld4<mode>_dreg): Likewise.
682 (aarch64_ld4<mode>_dreg): Likewise.
683 (aarch64_tbl1<mode>): Likewise.
684 (aarch64_tbl2v16qi): Likewise.
685 (aarch64_combinev16qi): Likewise.
686 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
687 (aarch64_st2<mode>_dreg): Likewise.
688 (aarch64_st2<mode>_dreg): Likewise.
689 (aarch64_st3<mode>_dreg): Likewise.
690 (aarch64_st3<mode>_dreg): Likewise.
691 (aarch64_st4<mode>_dreg): Likewise.
692 (aarch64_st4<mode>_dreg): Likewise.
693 (*aarch64_simd_ld1r<mode>): Likewise.
694 (aarch64_frecpe<mode>): Likewise.
695 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
696 (aarch64_frecps<mode>): Likewise.
698 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
700 * config/arm/iterators.md (V_elem_ch): New.
702 (VQH_type): Likewise.
703 * config/arm/arm.md (is_neon_type): New.
704 (conds): Use is_neon_type.
705 (anddi3_insn): Update type attribute.
706 (xordi3_insn): Likewise.
707 (one_cmpldi2): Likewise.
708 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
709 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
710 (*movmisalign<mode>_neon_store): Likewise.
711 (*movmisalign<mode>_neon_load): Likewise.
712 (vec_set<mode>_internal): Likewise.
713 (vec_set<mode>_internal): Likewise.
714 (vec_setv2di_internal): Likewise.
715 (vec_extract<mode>): Likewise.
716 (vec_extract<mode>): Likewise.
717 (vec_extractv2di): Likewise.
718 (*add<mode>3_neon): Likewise.
719 (adddi3_neon): Likewise.
720 (*sub<mode>3_neon): Likewise.
721 (subdi3_neon): Likewise.
722 (fma<VCVTF:mode>4): Likewise.
723 (fma<VCVTF:mode>4_intrinsic): Likewise.
724 (*fmsub<VCVTF:mode>4): Likewise.
725 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
726 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
727 (ior<mode>3): Likewise.
728 (and<mode>3): Likewise.
729 (orn<mode>3_neon): Likewise.
730 (orndi3_neon): Likewise.
731 (bic<mode>3_neon): Likewise.
732 (bicdi3_neon): Likewise.
733 (xor<mode>3): Likewise.
734 (one_cmpl<mode>2): Likewise.
735 (abs<mode>2): Likewise.
736 (neg<mode>2): Likewise.
737 (negdi2_neon): Likewise.
738 (*umin<mode>3_neon): Likewise.
739 (*umax<mode>3_neon): Likewise.
740 (*smin<mode>3_neon): Likewise.
741 (*smax<mode>3_neon): Likewise.
742 (vashl<mode>3): Likewise.
743 (vashr<mode>3_imm): Likewise.
744 (vlshr<mode>3_imm): Likewise.
745 (ashl<mode>3_signed): Likewise.
746 (ashl<mode>3_unsigned): Likewise.
747 (neon_load_count): Likewise.
748 (ashldi3_neon_noclobber): Likewise.
749 (ashldi3_neon): Likewise.
750 (signed_shift_di3_neon): Likewise.
751 (unsigned_shift_di3_neon): Likewise.
752 (ashrdi3_neon_imm_noclobber): Likewise.
753 (lshrdi3_neon_imm_noclobber): Likewise.
754 (<shift>di3_neon): Likewise.
755 (widen_ssum<mode>3): Likewise.
756 (widen_usum<mode>3): Likewise.
757 (quad_halves_<code>v4si): Likewise.
758 (quad_halves_<code>v4sf): Likewise.
759 (quad_halves_<code>v8hi): Likewise.
760 (quad_halves_<code>v16qi): Likewise.
761 (reduc_splus_v2di): Likewise.
762 (neon_vpadd_internal<mode>): Likewise.
763 (neon_vpsmin<mode>): Likewise.
764 (neon_vpsmax<mode>): Likewise.
765 (neon_vpumin<mode>): Likewise.
766 (neon_vpumax<mode>): Likewise.
767 (*ss_add<mode>_neon): Likewise.
768 (*us_add<mode>_neon): Likewise.
769 (*ss_sub<mode>_neon): Likewise.
770 (*us_sub<mode>_neon): Likewise.
771 (neon_vadd<mode>_unspec): Likewise.
772 (neon_vaddl<mode>): Likewise.
773 (neon_vaddw<mode>): Likewise.
774 (neon_vhadd<mode>): Likewise.
775 (neon_vqadd<mode>): Likewise.
776 (neon_vaddhn<mode>): Likewise.
777 (neon_vmul<mode>): Likewise.
778 (neon_vfms<VCVTF:mode>): Likewise.
779 (neon_vmlal<mode>): Likewise.
780 (neon_vmls<mode>): Likewise.
781 (neon_vmlsl<mode>): Likewise.
782 (neon_vqdmulh<mode>): Likewise.
783 (neon_vqdmlal<mode>): Likewise.
784 (neon_vqdmlsl<mode>): Likewise.
785 (neon_vmull<mode>): Likewise.
786 (neon_vqdmull<mode>): Likewise.
787 (neon_vsub<mode>_unspec): Likewise.
788 (neon_vsubl<mode>): Likewise.
789 (neon_vsubw<mode>): Likewise.
790 (neon_vqsub<mode>): Likewise.
791 (neon_vhsub<mode>): Likewise.
792 (neon_vsubhn<mode>): Likewise.
793 (neon_vceq<mode>): Likewise.
794 (neon_vcge<mode>): Likewise.
795 (neon_vcgeu<mode>): Likewise.
796 (neon_vcgt<mode>): Likewise.
797 (neon_vcgtu<mode>): Likewise.
798 (neon_vcle<mode>): Likewise.
799 (neon_vclt<mode>): Likewise.
800 (neon_vcage<mode>): Likewise.
801 (neon_vcagt<mode>): Likewise.
802 (neon_vtst<mode>): Likewise.
803 (neon_vabd<mode>): Likewise.
804 (neon_vabdl<mode>): Likewise.
805 (neon_vaba<mode>): Likewise.
806 (neon_vabal<mode>): Likewise.
807 (neon_vmax<mode>): Likewise.
808 (neon_vmin<mode>): Likewise.
809 (neon_vpaddl<mode>): Likewise.
810 (neon_vpadal<mode>): Likewise.
811 (neon_vpmax<mode>): Likewise.
812 (neon_vpmin<mode>): Likewise.
813 (neon_vrecps<mode>): Likewise.
814 (neon_vrsqrts<mode>): Likewise.
815 (neon_vqabs<mode>): Likewise.
816 (neon_vqneg<mode>): Likewise.
817 (neon_vcls<mode>): Likewise.
818 (clz<mode>2): Likewise.
819 (popcount<mode>2): Likewise.
820 (neon_vrecpe<mode>): Likewise.
821 (neon_vrsqrte<mode>): Likewise.
822 (neon_vget_lane<mode>_sext_internal): Likewise.
823 (neon_vget_lane<mode>_zext_internal): Likewise.
824 (neon_vdup_n<mode>): Likewise.
825 (neon_vdup_n<mode>): Likewise.
826 (neon_vdup_nv2di): Likewise.
827 (neon_vdup_lane<mode>_interal): Likewise.
828 (*neon_vswp<mode>): Likewise.
829 (neon_vcombine<mode>): Likewise.
830 (float<mode><V_cvtto>2): Likewise.
831 (floatuns<mode><V_cvtto>2): Likewise.
832 (fix_trunc<mode><V_cvtto>2): Likewise.
833 (fixuns_trunc<mode><V_cvtto>2
834 (neon_vcvt<mode>): Likewise.
835 (neon_vcvt<mode>): Likewise.
836 (neon_vcvtv4sfv4hf): Likewise.
837 (neon_vcvtv4hfv4sf): Likewise.
838 (neon_vcvt_n<mode>): Likewise.
839 (neon_vcvt_n<mode>): Likewise.
840 (neon_vmovn<mode>): Likewise.
841 (neon_vqmovn<mode>): Likewise.
842 (neon_vqmovun<mode>): Likewise.
843 (neon_vmovl<mode>): Likewise.
844 (neon_vmul_lane<mode>): Likewise.
845 (neon_vmul_lane<mode>): Likewise.
846 (neon_vmull_lane<mode>): Likewise.
847 (neon_vqdmull_lane<mode>): Likewise.
848 (neon_vqdmulh_lane<mode>): Likewise.
849 (neon_vqdmulh_lane<mode>): Likewise.
850 (neon_vmla_lane<mode>): Likewise.
851 (neon_vmla_lane<mode>): Likewise.
852 (neon_vmlal_lane<mode>): Likewise.
853 (neon_vqdmlal_lane<mode>): Likewise.
854 (neon_vmls_lane<mode>): Likewise.
855 (neon_vmls_lane<mode>): Likewise.
856 (neon_vmlsl_lane<mode>): Likewise.
857 (neon_vqdmlsl_lane<mode>): Likewise.
858 (neon_vext<mode>): Likewise.
859 (neon_vrev64<mode>): Likewise.
860 (neon_vrev32<mode>): Likewise.
861 (neon_vrev16<mode>): Likewise.
862 (neon_vbsl<mode>_internal): Likewise.
863 (neon_vshl<mode>): Likewise.
864 (neon_vqshl<mode>): Likewise.
865 (neon_vshr_n<mode>): Likewise.
866 (neon_vshrn_n<mode>): Likewise.
867 (neon_vqshrn_n<mode>): Likewise.
868 (neon_vqshrun_n<mode>): Likewise.
869 (neon_vshl_n<mode>): Likewise.
870 (neon_vqshl_n<mode>): Likewise.
871 (neon_vqshlu_n<mode>): Likewise.
872 (neon_vshll_n<mode>): Likewise.
873 (neon_vsra_n<mode>): Likewise.
874 (neon_vsri_n<mode>): Likewise.
875 (neon_vsli_n<mode>): Likewise.
876 (neon_vtbl1v8qi): Likewise.
877 (neon_vtbl2v8qi): Likewise.
878 (neon_vtbl3v8qi): Likewise.
879 (neon_vtbl4v8qi): Likewise.
880 (neon_vtbl1v16qi): Likewise.
881 (neon_vtbl2v16qi): Likewise.
882 (neon_vcombinev16qi): Likewise.
883 (neon_vtbx1v8qi): Likewise.
884 (neon_vtbx2v8qi): Likewise.
885 (neon_vtbx3v8qi): Likewise.
886 (neon_vtbx4v8qi): Likewise.
887 (*neon_vtrn<mode>_insn): Likewise.
888 (*neon_vzip<mode>_insn): Likewise.
889 (*neon_vuzp<mode>_insn): Likewise.
890 (neon_vld1<mode>): Likewise.
891 (neon_vld1_lane<mode>): Likewise.
892 (neon_vld1_lane<mode>): Likewise.
893 (neon_vld1_dup<mode>): Likewise.
894 (neon_vld1_dup<mode>): Likewise.
895 (neon_vld1_dupv2di): Likewise.
896 (neon_vst1<mode>): Likewise.
897 (neon_vst1_lane<mode>): Likewise.
898 (neon_vst1_lane<mode>): Likewise.
899 (neon_vld2<mode>): Likewise.
900 (neon_vld2<mode>): Likewise.
901 (neon_vld2_lane<mode>): Likewise.
902 (neon_vld2_lane<mode>): Likewise.
903 (neon_vld2_dup<mode>): Likewise.
904 (neon_vst2<mode>): Likewise.
905 (neon_vst2<mode>): Likewise.
906 (neon_vst2_lane<mode>): Likewise.
907 (neon_vst2_lane<mode>): Likewise.
908 (neon_vld3<mode>): Likewise.
909 (neon_vld3qa<mode>): Likewise.
910 (neon_vld3qb<mode>): Likewise.
911 (neon_vld3_lane<mode>): Likewise.
912 (neon_vld3_lane<mode>): Likewise.
913 (neon_vld3_dup<mode>): Likewise.
914 (neon_vst3<mode>): Likewise.
915 (neon_vst3qa<mode>): Likewise.
916 (neon_vst3qb<mode>): Likewise.
917 (neon_vst3_lane<mode>): Likewise.
918 (neon_vst3_lane<mode>): Likewise.
919 (neon_vld4<mode>): Likewise.
920 (neon_vld4qa<mode>): Likewise.
921 (neon_vld4qb<mode>): Likewise.
922 (neon_vld4_lane<mode>): Likewise.
923 (neon_vld4_lane<mode>): Likewise.
924 (neon_vld4_dup<mode>): Likewise.
925 (neon_vst4<mode>): Likewise.
926 (neon_vst4qa<mode>): Likewise.
927 (neon_vst4qb<mode>): Likewise.
928 (neon_vst4_lane<mode>): Likewise.
929 (neon_vst4_lane<mode>): Likewise.
930 (neon_vec_unpack<US>_lo_<mode>): Likewise.
931 (neon_vec_unpack<US>_hi_<mode>): Likewise.
932 (neon_vec_<US>mult_lo_<mode>): Likewise.
933 (neon_vec_<US>mult_hi_<mode>): Likewise.
934 (neon_vec_<US>shiftl_<mode>): Likewise.
935 (neon_unpack<US>_<mode>): Likewise.
936 (neon_vec_<US>mult_<mode>): Likewise.
937 (vec_pack_trunc_<mode>): Likewise.
938 (neon_vec_pack_trunc_<mode>): Likewise.
939 (neon_vabd<mode>_2): Likewise.
940 (neon_vabd<mode>_3): Likewise.
942 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
944 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
945 (load_pair): Update type attribute.
946 (store_pair): Update type attribute.
947 * config/aarch64/iterators.md (q): New.
949 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
951 * config/arm/types.md: Add new types for Neon insns.
953 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
954 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
955 Sergey Lega <sergey.s.lega@intel.com>
956 Anna Tikhonova <anna.tikhonova@intel.com>
957 Ilya Tocar <ilya.tocar@intel.com>
958 Andrey Turetskiy <andrey.turetskiy@intel.com>
959 Ilya Verbin <ilya.verbin@intel.com>
960 Kirill Yukhin <kirill.yukhin@intel.com>
961 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
963 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
964 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
965 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
967 (srcp14<mode>): Ditto.
968 (rsqrt14<mode>): Ditto.
969 (rsqrt14<mode>): Ditto.
970 (avx512f_vmscalef<mode>): Ditto.
971 (avx512f_scalef<mode>): Ditto.
972 (avx512f_getexp<mode>): Ditto.
973 (avx512f_sgetexp<mode>): Ditto.
974 (avx512f_fixupimm<mode>): Ditto.
975 (avx512f_sfixupimm<mode>): Ditto.
976 (avx512f_rndscale<mode>): Ditto.
977 (*avx512er_exp2<mode>): Ditto.
978 (*avx512er_rcp28<mode>): Ditto.
979 (avx512er_rsqrt28<mode>): Ditto.
980 (avx512f_getmant<mode>): Ditto.
981 (avx512f_getmant<mode>): Ditto.
982 (avx512f_rndscale<mode>): Fix formatting.
984 2013-10-15 Martin Jambor <mjambor@suse.cz>
986 * ipa-utils.h (ipa_edge_within_scc): Declare.
987 * ipa-cp.c (edge_within_scc): Moved...
988 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
990 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
991 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
992 Sergey Lega <sergey.s.lega@intel.com>
993 Anna Tikhonova <anna.tikhonova@intel.com>
994 Ilya Tocar <ilya.tocar@intel.com>
995 Andrey Turetskiy <andrey.turetskiy@intel.com>
996 Ilya Verbin <ilya.verbin@intel.com>
997 Kirill Yukhin <kirill.yukhin@intel.com>
998 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1000 * config/i386/predicates.md (const_8_to_15_operand): New.
1001 (const_16_to_31_operand): Ditto.
1002 * config/i386/sse.md (V8FI): New.
1004 (reduc_splus_v8df): Ditto.
1005 (reduc_splus_v16sf): Ditto.
1006 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1007 (vec_extract_hi_<mode>): Ditto.
1008 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1009 (vec_set_lo_<mode>): Ditto.
1010 (vec_set_hi_<mode>): Ditto.
1011 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1012 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1013 (avx512f_pshufd_1): Ditto.
1014 (avx512f_broadcast<mode>): Ditto.
1015 (avx512f_broadcast<mode>): Ditto.
1016 (define_split): Split vec_extract_lo into move.
1017 (ssequartermode): Ditto.
1018 (ssedoublemode): Extened with wider modes.
1019 (vec_extract_lo_<mode>): Ditto.
1021 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1022 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1023 Sergey Lega <sergey.s.lega@intel.com>
1024 Anna Tikhonova <anna.tikhonova@intel.com>
1025 Ilya Tocar <ilya.tocar@intel.com>
1026 Andrey Turetskiy <andrey.turetskiy@intel.com>
1027 Ilya Verbin <ilya.verbin@intel.com>
1028 Kirill Yukhin <kirill.yukhin@intel.com>
1029 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1031 * config/i386/predicates.md (register_or_constm1_operand): New.
1032 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1033 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1034 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1035 UNSPEC_SCATTER_PREFETCH
1037 (avx512f_ucmp<mode>3): Ditto.
1038 (avx512f_vternlog<mode>): Ditto.
1039 (avx512f_align<mode>): Ditto.
1040 (<shift_insn><mode>3): Ditto.
1041 (avx512f_<rotate>v<mode>): Ditto.
1042 (avx512f_<rotate><mode>): Ditto.
1043 (avx512f_eq<mode>3): Ditto.
1044 (avx512f_eq<mode>3_1): Ditto.
1045 (avx512f_gt<mode>3): Ditto.
1046 (avx512f_testm<mode>3): Ditto.
1047 (avx512f_testnm<mode>3): Ditto.
1048 (avx512pf_gatherpf<mode>): Ditto.
1049 (*avx512pf_gatherpf<mode>_mask): Ditto.
1050 (*avx512pf_gatherpf<mode>): Ditto.
1051 (avx512pf_scatterpf<mode>): Ditto.
1052 (*avx512pf_scatterpf<mode>_mask): Ditto.
1053 (*avx512pf_scatterpf<mode>): Ditto.
1054 (avx512f_vec_dup_gpr<mode>): Ditto.
1055 (clz<mode>2): Ditto.
1056 (conflict<mode>): Ditto.
1057 (REDUC_SMINMAX_MODE): Extened with wider modes.
1058 (reduc_<code>_<mode>): Ditto.
1059 (vlshr<mode>3): Ditto.
1060 (vashl<mode>3): Ditto.
1062 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1063 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1064 Sergey Lega <sergey.s.lega@intel.com>
1065 Anna Tikhonova <anna.tikhonova@intel.com>
1066 Ilya Tocar <ilya.tocar@intel.com>
1067 Andrey Turetskiy <andrey.turetskiy@intel.com>
1068 Ilya Verbin <ilya.verbin@intel.com>
1069 Kirill Yukhin <kirill.yukhin@intel.com>
1070 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1072 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1075 (avx512fmaskmode): Ditto.
1076 (bcstscalarsuff): Ditto.
1077 (avx512f_blendm<mode>): Ditto.
1078 (cmp_imm_predicate): Ditto.
1079 (avx512f_cmp<mode>3): Ditto.
1080 (avx512f_vec_dup<mode>): Ditto.
1081 (avx512f_vec_dup_mem<mode>): Ditto.
1082 (avx512f_vpermi2var<mode>3): Ditto.
1083 (avx512f_vpermt2var<mode>3): Ditto.
1084 (vec_init<mode>): Ditto.
1085 (avx512f_gathersi<mode>): Ditto.
1086 (*avx512f_gathersi<mode>): Ditto.
1087 (*avx512f_gathersi<mode>_2): Ditto.
1088 (avx512f_gatherdi<mode>): Ditto.
1089 (*avx512f_gatherdi<mode>): Ditto.
1090 (*avx512f_gatherdi<mode>_2): Ditto.
1091 (avx512f_scattersi<mode>): Ditto.
1092 (*avx512f_scattersi<mode>): Ditto.
1093 (avx512f_scatterdi<mode>): Ditto.
1094 (*avx512f_scatterdi<mode>): Ditto.
1095 (sseintprefix): Extened with wider modes.
1096 (VEC_GATHER_IDXSI): Ditto.
1097 (VEC_GATHER_IDXDI): Ditto.
1098 (VEC_GATHER_SRCDI): Ditto.
1100 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1101 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1103 * config/arm/t-aprofile: New file.
1104 * config.gcc: Handle --with-multilib-list option.
1106 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1108 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
1111 2013-10-15 Richard Biener <rguenther@suse.de>
1113 * tree-loop-distribution.c (build_empty_rdg): Inline into
1115 (rdg_flag_vertex): Inline into single user.
1116 (rdg_flag_vertex_and_dependent): Likewise.
1117 (build_rdg_partition_for_vertex): Remove processed bitmap.
1118 (rdg_build_partitions): Simplify.
1120 2013-10-15 Richard Biener <rguenther@suse.de>
1122 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1123 Restructure forwarding through conversions and copies to
1124 avoid performing copy-propagation the wrong way. Adjust
1125 recursion invocations.
1126 (forward_propagate_addr_expr): Add argument stating if we
1127 are recursing from a single-use.
1128 (ssa_forward_propagate_and_combine): Adjust.
1130 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1132 * dumpfile.h (gcc::dump_manager): New class, to hold state
1133 relating to dumpfile management.
1134 (get_dump_file_name): Remove in favor of method of dump_manager.
1135 (dump_initialized_p): Likewise.
1136 (dump_start): Likewise.
1137 (dump_finish): Likewise.
1138 (dump_switch_p): Likewise.
1139 (dump_register): Likewise.
1140 (get_dump_file_info): Likewise.
1141 * context.c (gcc::context::context): Construct the dump_manager
1143 * context.h (gcc::context::get_dumps): New.
1144 (gcc::context::m_dumps): New.
1145 * coverage.c (coverage_init): Port to dump_manager API.
1146 * dumpfile.c (extra_dump_files): Convert to field of
1148 (extra_dump_files_in_use): Likewise.
1149 (extra_dump_files_alloced): Likewise.
1150 (gcc::dump_manager::dump_manager): New.
1151 (dump_register): Convert to...
1152 (gcc::dump_manager::dump_register): ...method, replacing
1153 function-static next_dump with m_next_dump field.
1154 (get_dump_file_info): Convert to...
1155 (gcc::dump_manager::get_dump_file_info): ...method.
1156 (get_dump_file_name): Convert to...
1157 (gcc::dump_manager::get_dump_file_name): ...method.
1158 (dump_start): Convert to...
1159 (gcc::dump_manager::dump_start): ...method.
1160 (dump_finish): Convert to...
1161 (gcc::dump_manager::dump_finish): ...method.
1162 (dump_begin): Replace body with...
1163 (gcc::dump_manager::dump_begin): ...new method.
1164 (dump_phase_enabled_p): Convert to...
1165 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1166 (dump_phase_enabled_p): Convert to...
1167 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1168 (dump_initialized_p): Convert to...
1169 (gcc::dump_manager::dump_initialized_p): ...method.
1170 (dump_flag_name): Replace body with...
1171 (gcc::dump_manager::dump_flag_name): ...new method.
1172 (dump_enable_all): Convert to...
1173 (gcc::dump_manager::dump_enable_all): ...new method.
1174 (opt_info_enable_passes): Convert to...
1175 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1176 (dump_switch_p_1): Convert to...
1177 (gcc::dump_manager::dump_switch_p_1): ...new method.
1178 (dump_switch_p): Convert to...
1179 (gcc::dump_manager::dump_switch_p): ...new method.
1180 (opt_info_switch_p): Port to dump_manager API.
1181 (enable_rtl_dump_file): Likewise.
1182 * opts-global.c (handle_common_deferred_options): Port to new
1184 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1185 (pass_manager::register_one_dump_file): Likewise.
1186 (pass_manager::register_pass): Likewise.
1187 (pass_init_dump_file): Likewise.
1188 (pass_fini_dump_file): Likewise.
1189 * statistics.c (statistics_early_init): Likewise.
1191 2013-10-14 Richard Biener <rguenther@suse.de>
1193 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1194 iterative_hash_canonical_type, gimple_canonical_type_hash,
1195 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1196 gimple_register_canonical_type, print_gimple_types_stats,
1197 free_gimple_type_tables): Move to lto/lto.c
1198 (gt-gimple.h): Do not include.
1199 * gimple.h (gimple_register_canonical_type,
1200 print_gimple_types_stats, free_gimple_type_tables): Remove.
1201 * Makefile.in (GTFILES): Remove gimple.c.
1203 2013-10-14 Travis Snoozy <quandary@remstate.com>
1206 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1207 scanning for msp430x targets.
1209 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1212 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1213 (registers_ok_for_ldd_peep): Move around.
1214 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1215 * config/sparc/sparc.md (widening peepholes): Use it.
1217 2013-10-14 Richard Biener <rguenther@suse.de>
1221 * gimple.c (iterative_hash_canonical_type): Make sure to
1222 record the hash into the correct hashtable slot.
1224 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1226 PR rtl-optimization/58662
1227 * combine.c (try_combine): Take into account death nodes on I2 when
1228 splitting a PARALLEL of two independent SETs. Fix dump message.
1230 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1233 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1234 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1235 * config/sh/t-sh (sh_treg_combine.o): New entry.
1236 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1237 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1238 (register_sh_passes): New function. Register sh_treg_combine pass.
1239 (sh_option_override): Invoke it.
1240 (sh_canonicalize_comparison): Handle op0_preserve_value.
1241 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1242 opportunities. Canonicalize branch condition.
1243 (nott): Allow only if pseudos can be created for non-SH2A.
1245 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1248 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1249 (ix86_expand_movmem): Replace copy_addr_to_reg with
1250 ix86_copy_addr_to_reg.
1251 (ix86_expand_setmem): Likewise.
1253 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1255 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1256 provided by ix86_fp_compare_mode instead of CCFPUmode.
1258 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1260 * config/aarch64/arm_neon.h
1261 (vtbx<1,3>_<psu>8): Fix register constriants.
1263 2013-10-11 Jeff Law <law@redhat.com>
1265 PR tree-optimization/58640
1266 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1267 threading paths that cross over two loop entry points.
1269 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1271 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1272 handle vector float as well.
1273 (*vsx_le_perm_load_v4si): Likewise.
1274 (*vsx_le_perm_store_v2di): Likewise.
1275 (*vsx_le_perm_store_v4si): Likewise.
1277 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1279 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1280 directly to circumvent subtract from splat{31} workaround.
1281 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1283 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1284 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1285 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1286 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1287 separate big and little endian logic.
1288 (*altivec_vperm_<mode>_internal): New define_insn.
1289 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1290 separate big and little endian logic.
1291 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1292 (vec_permv16qi): Add little endian logic.
1294 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1296 * doc/extend.texi (returns_nonnull): Remove arguments.
1298 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1299 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1300 Sergey Lega <sergey.s.lega@intel.com>
1301 Anna Tikhonova <anna.tikhonova@intel.com>
1302 Ilya Tocar <ilya.tocar@intel.com>
1303 Andrey Turetskiy <andrey.turetskiy@intel.com>
1304 Ilya Verbin <ilya.verbin@intel.com>
1305 Kirill Yukhin <kirill.yukhin@intel.com>
1306 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1308 * config/i386/sse.md (VI48F_256_512): New.
1309 (avx2_permvar<mode>): Change to ...
1310 (<avx2_avx512f>_permvar<mode>): This.
1312 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1313 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1314 Sergey Lega <sergey.s.lega@intel.com>
1315 Anna Tikhonova <anna.tikhonova@intel.com>
1316 Ilya Tocar <ilya.tocar@intel.com>
1317 Andrey Turetskiy <andrey.turetskiy@intel.com>
1318 Ilya Verbin <ilya.verbin@intel.com>
1319 Kirill Yukhin <kirill.yukhin@intel.com>
1320 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1322 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1323 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1324 * config/i386/sse.md (VI4_AVX): New.
1325 (sf2simodelower): Ditto.
1326 (sse2_cvtps2dq): Change to ...
1327 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1329 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1330 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1331 Sergey Lega <sergey.s.lega@intel.com>
1332 Anna Tikhonova <anna.tikhonova@intel.com>
1333 Ilya Tocar <ilya.tocar@intel.com>
1334 Andrey Turetskiy <andrey.turetskiy@intel.com>
1335 Ilya Verbin <ilya.verbin@intel.com>
1336 Kirill Yukhin <kirill.yukhin@intel.com>
1337 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1339 * config/i386/sse.md (V_512): New.
1341 (vcond<V_512:mode><VF_512:mode>): Ditto.
1342 (vcond<V_512:mode><VI_512:mode>): Ditto.
1343 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1345 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1346 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1347 Sergey Lega <sergey.s.lega@intel.com>
1348 Anna Tikhonova <anna.tikhonova@intel.com>
1349 Ilya Tocar <ilya.tocar@intel.com>
1350 Andrey Turetskiy <andrey.turetskiy@intel.com>
1351 Ilya Verbin <ilya.verbin@intel.com>
1352 Kirill Yukhin <kirill.yukhin@intel.com>
1353 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1355 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1356 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1358 (fma<mode>4): Removed condition.
1359 (fms<mode>4): Ditto.
1360 (fnma<mode>4): Ditto.
1361 (fnms<mode>4): Ditto.
1362 (fma4i_fmadd_<mode>): Ditto.
1363 (*fma_fmadd_<mode>): Ditto.
1364 (*fma_fmsub_<mode>): Ditto.
1365 (*fma_fnmadd_<mode>): Ditto.
1366 (*fma_fnmsub_<mode>): Ditto.
1367 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1368 (*fma_fmaddsub_<mode>): Ditto.
1369 (*fma_fmsubadd_<mode>): Ditto.
1370 (*fmai_fmadd_<mode>): Ditto.
1371 (*fmai_fmsub_<mode>): Ditto.
1372 (*fmai_fnmadd_<mode>): Ditto.
1373 (*fmai_fnmsub_<mode>): Ditto.
1375 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1376 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1377 Sergey Lega <sergey.s.lega@intel.com>
1378 Anna Tikhonova <anna.tikhonova@intel.com>
1379 Ilya Tocar <ilya.tocar@intel.com>
1380 Andrey Turetskiy <andrey.turetskiy@intel.com>
1381 Ilya Verbin <ilya.verbin@intel.com>
1382 Kirill Yukhin <kirill.yukhin@intel.com>
1383 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1385 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1386 (VI124_256): Changed to ...
1387 (VI124_256_48_512): This.
1388 (ssepackmode): Extended with wider modes.
1389 (<code><mode>3): Changed iterator.
1390 (*avx2_<code><mode>3): Ditto.
1391 (vec_pack_trunc_<mode>): Ditto.
1393 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1394 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1395 Sergey Lega <sergey.s.lega@intel.com>
1396 Anna Tikhonova <anna.tikhonova@intel.com>
1397 Ilya Tocar <ilya.tocar@intel.com>
1398 Andrey Turetskiy <andrey.turetskiy@intel.com>
1399 Ilya Verbin <ilya.verbin@intel.com>
1400 Kirill Yukhin <kirill.yukhin@intel.com>
1401 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1403 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1404 (VI8F_256_512): Ditto.
1405 (abs<mode>2): Changed iterator.
1406 (avx2_perm<mode>): Changed to ...
1407 (<avx2_avx512f>_perm<mode>): This.
1408 (avx2_perm<mode>_1): Changed to ...
1409 (<avx2_avx512f>_perm<mode>_1): This.
1411 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1412 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1413 Sergey Lega <sergey.s.lega@intel.com>
1414 Anna Tikhonova <anna.tikhonova@intel.com>
1415 Ilya Tocar <ilya.tocar@intel.com>
1416 Andrey Turetskiy <andrey.turetskiy@intel.com>
1417 Ilya Verbin <ilya.verbin@intel.com>
1418 Kirill Yukhin <kirill.yukhin@intel.com>
1419 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1421 * config/i386/sse.md (VI48_AVX512F): New.
1422 (VI48_AVX2): Changed to ...
1423 (VI48_AVX2_48_AVX512F): This.
1424 (avx2_ashrv<mode>): Changed to ...
1425 (<avx2_avx512f>_ashrv<mode>): This.
1426 (avx2_<shift_insn>v<mode>): Changed to ...
1427 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1429 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1430 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1431 Sergey Lega <sergey.s.lega@intel.com>
1432 Anna Tikhonova <anna.tikhonova@intel.com>
1433 Ilya Tocar <ilya.tocar@intel.com>
1434 Andrey Turetskiy <andrey.turetskiy@intel.com>
1435 Ilya Verbin <ilya.verbin@intel.com>
1436 Kirill Yukhin <kirill.yukhin@intel.com>
1437 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1439 * config/i386/sse.md (VI4_AVX512F): New.
1440 (VI8_AVX2_AVX512F): Ditto.
1441 (mul<mode>3): Extended with wider modes.
1442 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1443 (mul<mode>3): Ditto.
1444 (vec_widen_<s>mult_odd_<mode>): Ditto.
1446 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1447 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1448 Sergey Lega <sergey.s.lega@intel.com>
1449 Anna Tikhonova <anna.tikhonova@intel.com>
1450 Ilya Tocar <ilya.tocar@intel.com>
1451 Andrey Turetskiy <andrey.turetskiy@intel.com>
1452 Ilya Verbin <ilya.verbin@intel.com>
1453 Kirill Yukhin <kirill.yukhin@intel.com>
1454 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1456 * config/i386/sse.md (VI2_AVX512F): New.
1457 (VI124_AVX512F): Ditto.
1458 (sseunpackmode): Extended with wider modes.
1459 (sseunpackfltmode): Ditto.
1460 (vec_unpacks_float_hi_<mode>): Ditto.
1461 (vec_unpacks_float_lo_<mode>): Ditto.
1462 (vec_unpacku_float_hi_<mode>): Ditto.
1463 (vec_unpacku_float_lo_<mode>): Ditto.
1464 (vec_unpacks_lo_<mode>): Ditto.
1465 (vec_unpacks_hi_<mode>): Ditto.
1466 (vec_unpacku_lo_<mode>): Ditto.
1467 (vec_unpacku_hi_<mode>): Ditto.
1469 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1470 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1471 Sergey Lega <sergey.s.lega@intel.com>
1472 Anna Tikhonova <anna.tikhonova@intel.com>
1473 Ilya Tocar <ilya.tocar@intel.com>
1474 Andrey Turetskiy <andrey.turetskiy@intel.com>
1475 Ilya Verbin <ilya.verbin@intel.com>
1476 Kirill Yukhin <kirill.yukhin@intel.com>
1477 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1479 * config/i386/i386.md (multdiv): New.
1480 (multdiv_mnemonic): Ditto.
1481 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1482 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1483 (<sse>_vmdiv<mode>3): Removed.
1485 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1486 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1487 Sergey Lega <sergey.s.lega@intel.com>
1488 Anna Tikhonova <anna.tikhonova@intel.com>
1489 Ilya Tocar <ilya.tocar@intel.com>
1490 Andrey Turetskiy <andrey.turetskiy@intel.com>
1491 Ilya Verbin <ilya.verbin@intel.com>
1492 Kirill Yukhin <kirill.yukhin@intel.com>
1493 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1495 * config/i386/sse.md (V): Extended with wider modes.
1497 (ssehalfvecmode): Ditto.
1499 (ssepackfltmode): Ditto.
1500 (avx_vec_concat<mode>): Ditto.
1501 (V_256_512): New iterator.
1502 (VF2_512_256): Ditto.
1503 (si2dfmode): New attribute.
1504 (si2dfmodelower): Ditto.
1506 (concat_tg_mode): Ditto.
1507 (floatv4siv4df2): Changed to ...
1508 (float<si2dfmodelower><mode>2): This.
1509 (avx_cvtps2pd256): Changed to ...
1510 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1511 (vec_pack_trunc_v4df): Changed to ...
1512 (vec_pack_trunc_<mode>): This.
1513 (avx_vpermil<mode>): Changed to ...
1514 (<sse2_avx_avx512f>_vpermil<mode>): This.
1515 (<fixsuffix>fix_truncv8dfv8si2): New.
1516 (vec_pack_sfix_trunc_v8df): Ditto.
1517 (avx512f_rndscale<mode>): Ditto.
1518 (avx512f_roundpd512): Ditto.
1519 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1521 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1522 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1523 Sergey Lega <sergey.s.lega@intel.com>
1524 Anna Tikhonova <anna.tikhonova@intel.com>
1525 Ilya Tocar <ilya.tocar@intel.com>
1526 Andrey Turetskiy <andrey.turetskiy@intel.com>
1527 Ilya Verbin <ilya.verbin@intel.com>
1528 Kirill Yukhin <kirill.yukhin@intel.com>
1529 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1531 * config/i386/i386.md (any_fix): New iterator.
1532 (fixsuffix): New attribute.
1533 * config/i386/sse.md (VF1): Extened with wider modes.
1537 (sseintvecmodelower): Ditto.
1538 (ssescalarmode): Ditto.
1539 (ssescalarnum): Ditto.
1541 (ssexmmmode): Ditto.
1542 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1543 (<sse>_rcp<mode>2): Change iterator.
1544 (rsqrt<mode>2): Ditto.
1545 (<sse>_rsqrt<mode>2): Ditto.
1546 (avx2_vec_dup<mode>): Ditto.
1547 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1548 (round<mode>2_sfix): Ditto.
1549 (avx2_pbroadcast<mode>): Ditto.
1550 (*andnot<mode>3): Handle XI mode.
1551 (*<code><mode>3): Ditto.
1552 (AVXTOSSEMODE): Removed.
1553 (avx_vpermil<mode>): Changed to ...
1554 (<sse2_avx_avx512f>_vpermil<mode>): This.
1556 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1557 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1558 Sergey Lega <sergey.s.lega@intel.com>
1559 Anna Tikhonova <anna.tikhonova@intel.com>
1560 Ilya Tocar <ilya.tocar@intel.com>
1561 Andrey Turetskiy <andrey.turetskiy@intel.com>
1562 Ilya Verbin <ilya.verbin@intel.com>
1563 Kirill Yukhin <kirill.yukhin@intel.com>
1564 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1566 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1567 (<sse>_comi): Ditto.
1568 (<sse>_ucomi): Ditto.
1569 (sse_cvtss2siq_2): Ditto.
1570 (sse2_cvtsd2si): Ditto.
1571 (sse2_cvtsd2siq): Ditto.
1572 (sse2_cvttsd2si): Ditto.
1573 (sse2_cvttsd2siq): Ditto.
1574 (<shift_insn><mode>3): Ditto.
1575 (sse2_cvtsi2sdq): Update constraint and prefix.
1576 (sse_cvtsi2ss): Update prefix.
1577 (sse_cvtsi2ssq): Ditto.
1579 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1581 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1582 ignore internal calls.
1584 2013-10-11 Richard Biener <rguenther@suse.de>
1586 * tree-pretty-print.c (dump_generic_node): Allow to dump
1587 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1588 (OVF) if TREE_OVERFLOW is set.
1590 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1592 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1594 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1595 * gimple.h: Likewise.
1597 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1599 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1602 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1604 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1605 * configure: Regenerate.
1607 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1609 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1610 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1611 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1612 similarly pp_right_paren.
1613 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1614 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1615 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1616 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1618 (omp_declare_simd_clauses_equal,
1619 omp_remove_redundant_declare_simd_attrs): New functions.
1620 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1621 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1622 * tree.h (OMP_LOOP_CHECK): Define.
1623 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1624 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1625 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1626 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1627 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1628 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1629 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1630 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1631 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1632 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1633 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1634 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1635 (omp_remove_redundant_declare_simd_attrs): New prototype.
1636 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1637 GIMPLE_OMP_TEAMS): New codes.
1638 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1639 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1641 (target_nesting_level): New variable.
1642 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1643 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1644 for collapse > 1 static schedule unless ordered.
1645 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1646 (determine_parallel_type): Adjust get_ws_args_for caller.
1647 (install_var_field): Handle mask & 4 for double indirection.
1648 (scan_sharing_clauses): Ignore shared clause on teams construct.
1649 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1650 (create_omp_child_function): If inside target or declare target
1651 constructs, set "omp declare target" attribute on the child function.
1652 (find_combined_for): New function.
1653 (scan_omp_parallel): Handle combined loops.
1654 (scan_omp_target, scan_omp_teams): New functions.
1655 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1656 restrictions and set ctx->cancellable for cancellable constructs.
1657 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1658 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1659 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1660 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1661 (omp_clause_aligned_alignment): New function.
1662 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1664 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1665 on teams constructs. Handle user defined reductions and new
1667 (lower_reduction_clauses): Don't set placeholder to address of ref
1668 if it has already the right type.
1669 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1670 (expand_parallel_call): Use the new non-_start suffixed builtins,
1671 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1672 and GOMP_parallel_end after the call.
1673 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1674 (expand_omp_for_init_counts): Handle combined loops.
1675 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1677 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1678 end of cancellable loops.
1679 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1680 Likewise. Handle collapse > 1 loops.
1681 (expand_omp_simd): Handle combined loops.
1682 (expand_omp_for): Add inner_stmt argument, adjust callers of
1683 expand_omp_for* functions, use expand_omp_for_static*chunk even
1684 for collapse > 1 unless ordered.
1685 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1686 of cancellable sections.
1687 (expand_omp_single): Remove need_barrier variable, just rely on
1688 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1689 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1690 (expand_omp_atomic_load, expand_omp_atomic_store,
1691 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1692 (expand_omp_target): New function.
1693 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1694 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1695 (build_omp_regions_1): Immediately close region for
1696 GF_OMP_TARGET_KIND_UPDATE.
1697 (maybe_add_implicit_barrier_cancel): New function.
1698 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1700 (lower_omp_single): Likewise. Add clobber after the barrier.
1701 (lower_omp_taskgroup): New function.
1702 (lower_omp_for): Handle combined loops. Adjust
1703 lower_rec_input_clauses caller. Handle cancellation.
1704 (lower_depend_clauses): New function.
1705 (lower_omp_taskreg): Lower depend clauses. Adjust
1706 lower_rec_input_clauses caller. Add clobber after the call. Handle
1708 (lower_omp_target, lower_omp_teams): New functions.
1709 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1710 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1711 and GOMP_cancellation_point calls.
1712 (lower_omp): Fold stmts inside of target region.
1713 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1714 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1715 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1716 (BT_FN_VOID_OMPFN_PTR_UINT,
1717 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1718 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1719 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1720 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1721 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1722 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1723 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1724 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1725 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1726 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1727 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1728 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1729 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1730 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1731 BUILT_IN_GOMP_PARALLEL_END.
1732 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1733 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1734 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1735 GF_OMP_FOR_KIND_DISTRIBUTE.
1736 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1737 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1738 (dump_gimple_omp_return): Print lhs if it has any.
1739 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1740 gimple_omp_atomic_seq_cst_p.
1741 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1742 and GIMPLE_OMP_TEAMS.
1743 * langhooks.c (lhd_omp_mappable_type): New function.
1744 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1745 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1747 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1748 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1749 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1751 (struct gimplify_omp_ctx): Add combined_loop field.
1752 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1753 on stmts inside of target region.
1754 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1755 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1756 ORT_TARGET and ORT_TARGET_DATA.
1757 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1759 (omp_notice_threadprivate_variable): Complain about threadprivate
1760 variables in target region.
1761 (omp_notice_variable): Complain about vars with non-mappable type
1762 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1763 (omp_check_private): Ignore ORT_TARGET* regions.
1764 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1765 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1766 (find_combined_omp_for): New function.
1767 (gimplify_omp_for): Handle gimplification of combined loops.
1768 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1770 (gimplify_omp_target_update): New function.
1771 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1772 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1773 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1774 (gimplify_body): If fndecl has "omp declare target" attribute, add
1775 implicit ORT_TARGET context around it.
1776 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1777 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1778 * tree-nested.c (convert_nonlocal_reference_stmt,
1779 convert_local_reference_stmt, convert_gimple_call): Handle
1780 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1781 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1782 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1783 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1784 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1785 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1786 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1787 BUILT_IN_GOMP_LOOP_END_CANCEL,
1788 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1789 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1790 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1791 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1792 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1793 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1794 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1795 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1796 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1797 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1798 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1799 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1800 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1801 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1802 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1803 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1804 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1805 gimple_build_omp_teams): New functions.
1806 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1807 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1808 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1809 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1810 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1811 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1812 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1813 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1814 GF_OMP_ATOMIC_SEQ_CST): New.
1815 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1816 gimple_build_omp_teams): New prototypes.
1817 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1818 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1819 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1820 GIMPLE_OMP_SINGLE as end of range.
1821 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1822 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1823 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1824 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1825 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1826 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1827 gimple_omp_target_kind, gimple_omp_target_set_kind,
1828 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1829 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1830 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1831 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1832 gimple_omp_teams_set_clauses): New inlines.
1833 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1834 and GIMPLE_OMP_TASKGROUP.
1835 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
1836 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1837 enum omp_clause_proc_bind_kind): New.
1838 (union omp_clause_subcode): Add depend_kind, map_kind and
1839 proc_bind_kind fields.
1840 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1841 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1842 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1843 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1844 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1846 2013-10-10 Teresa Johnson <tejohnson@google.com>
1848 * predict.c (tree_estimate_probability): Add new parameter
1849 for estimate_bb_frequencies.
1850 (estimate_bb_frequencies): Add new parameter to force estimation.
1851 (rebuild_frequencies): When max frequency in function is small,
1852 recompute counts from frequencies.
1853 * predict.h (estimate_bb_frequencies): New parameter.
1855 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1857 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1858 optimizations are disabled.
1860 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1862 * coverage.c (coverage_finish): Fix leak of da_file_name.
1864 2013-10-10 Jan Hubicka <jh@suse.cz>
1866 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1867 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1870 2013-10-10 Jakub Jelinek <jakub@redhat.com>
1873 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1874 if any labels are in FALLTHRU_BB, use a special label emitted
1875 immediately after the asm goto insn rather than label_rtx
1877 (expand_asm_stmt): Adjust caller.
1878 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1879 edge if the last insn in predecessor is a jump with single successor,
1880 but it isn't simplejump_p.
1882 2013-10-10 Richard Biener <rguenther@suse.de>
1884 PR tree-optimization/58656
1885 * tree-ssa-pre.c (phi_translate): Do not cache failed
1888 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1890 * gimplify.c: Include expr.h and tm_p.h for targets with special
1891 va-arg padding requirements.
1893 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1895 * tree-flow.h: Move some prototypes to gimple.h.
1896 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1897 * gimple.h: Relocate some prototypes from tree-flow.h
1898 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1900 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1901 (build_va_arg_indirect_ref): Relocate and make static.
1902 (std_gimplify_va_arg_expr): Relocate here.
1903 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1904 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1906 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1908 * doc/md.texi: Document the mnemonic attribute.
1910 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1913 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1914 (set_attr_alternative x ...) when searching for user defined
1917 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1919 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1921 2013-10-09 Easwaran Raman <eraman@google.com>
1923 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1924 * cfgexpand.c (defer_stack_allocation): ...use here
1925 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1927 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1929 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1930 (operand_equal_for_value_replacement): New function, extracted from
1931 value_replacement and enhanced to catch more cases.
1932 (value_replacement): Use operand_equal_for_value_replacement.
1934 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1936 * loop-doloop.c (doloop_modify, doloop_optimize): Use
1937 get_max_loop_iterations.
1939 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1941 * config/arm/aarch-common.c (arm_early_load_addr_dep):
1942 Place comment above function.
1944 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1946 * tree-flow.h: Remove all remaining prototypes, enums and structs that
1947 are not related to tree-cfg.c.
1948 * tree-ssa-address.h: New file. Relocate prototypes.
1949 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
1950 (addr_for_mem_ref): New. Combine call to get_address_description and
1951 return addr_for_mem_ref.
1952 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
1953 * tree-ssa-live.h: Adjust prototypes.
1954 * passes.c: Include tree-ssa-live.h.
1955 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
1956 * graphite.c (graphite_transform_loops): Make static.
1957 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1958 make_pass_graphite, pass_data_graphite_transforms,
1959 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
1960 * ipa-pure-const.c (warn_function_noreturn): Make static.
1961 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1962 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1963 Relocate from tree-cfg.c
1964 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
1966 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1967 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1968 Move to ipa-pure-const.c.
1969 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
1970 Relocate from tree-optimize.c.
1971 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
1972 make_pass_fixup_cfg): Move to tree-cfg.c.
1973 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
1974 Relocate some prototypes.
1975 * tree-data-ref.h (tree_check_data_deps) Add prototype.
1976 * tree-dump.c (dump_function_to_file): Remove prototype.
1977 Add tree-flow.h to the include file.
1978 * tree-dump.h: Remove prototype.
1979 * tree-parloops.h: New File. Add prototypes.
1980 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
1981 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
1982 from tree-ssa-loop.c.
1983 * tree-predcom.c (run_tree_predictive_commoning,
1984 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
1985 Relocate here from tree-ssa-loop.c.
1986 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
1987 ssa_name_values.release ().
1988 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
1989 (ssa_name_values): Relocate from tree-flow.h.
1990 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
1991 * tree-ssa-loop.c (run_tree_predictive_commoning,
1992 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
1993 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1994 make_pass_graphite, pass_data_graphite_transforms,
1995 make_pass_graphite_transforms, gate_tree_parallelize_loops,
1996 tree_parallelize_loops, pass_data_parallelize_loops,
1997 make_pass_parallelize_loops): Move to other files.
1998 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2000 * tree.h: Remove prototypes from tree-address.c.
2002 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2004 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2005 (struct int_tree_map): Move to tree-hasher.h
2006 (SCALE, LABEL, PERCENT): Move to gimple.h
2007 * tree-flow-inline.h: Delete. Move functions to other files.
2008 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2009 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2010 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2011 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2012 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2013 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2014 * tree-hasher.h: Don't include tree-flow.h.
2015 (struct int_tree_map): Relocate from tree-flow.h.
2016 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2017 tree-flow-inline.h and make static.
2018 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2020 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2021 tree-flow-inline.h and make static.
2022 * tree.h (is_global_var, may_be_aliased): Relocate from
2024 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2025 * value-prof.c: No longer include tree-flow-inline.h.
2026 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2028 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2030 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2031 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2032 (enum move_pos): Move to tree-ssa-loop-im.h
2033 * cfgloop.h: Move some prototypes.
2034 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2035 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2036 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2037 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2038 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2039 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2040 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2041 (enum move_pos): Relocate here.
2042 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2044 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2045 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2046 make_pass_lim): Relocate here from tree-ssa-loop.c.
2047 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2049 (loop_edge_to_cancel, unloop_loops): Make static.
2050 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2051 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2052 (tree_complete_unroll, gate_tree_complete_unroll,
2053 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2054 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2055 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2056 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2057 (stmt_invariant_in_loop_p): Remove unused function.
2058 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2059 * tree-ssa-loop-manip.h: New file. Add prototypes.
2060 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2061 (gcov_type_to_double_int): Move to cfgloop.h.
2062 (double_int_cmp, bound_index,
2063 estimate_numbers_of_iterations_loop): Make static.
2064 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2065 (max_loop_iterations): Factor out get_max_loop_iterations.
2066 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2068 * tree-ssa-loop-niter.h: New file. Add prototypes.
2069 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2070 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2071 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2072 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2073 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2074 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2075 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2076 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2077 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2078 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2079 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2080 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2081 pass_data_complete_unroll, make_pass_complete_unroll,
2082 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2083 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2084 tree-ssa-loop-ivcanon.c.
2085 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2086 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2087 tree-ssa-loop-prefetch.c.
2088 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2090 (get_lsm_tmp_name): Relocate and add suffix parameter.
2091 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2092 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2093 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2094 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2095 (get_estimated_loop_iterations): Factor out accessor from
2096 estimated_loop_iterations in tree-ssa-loop-niter.c.
2097 (get_max_loop_iterations): Factor out accessor from
2098 _max_loop_iterations in tree-ssa-niter.c.
2099 * loop-unroll.c (decide_unroll_constant_iterations,
2100 decide_unroll_runtime_iterations, decide_peel_simple,
2101 decide_unroll_stupid): Use new get_* accessors.
2103 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2105 PR tree-optimization/20318
2106 * doc/extend.texi (returns_nonnull): New function attribute.
2107 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2109 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2110 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2112 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2115 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2116 false if both components are bitfields.
2118 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2120 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2121 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2122 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2123 * config/aarch64/aarch64.h
2124 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2125 * config/aarch64/aarch64-simd-builtins.def
2126 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2128 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2130 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2132 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2134 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2136 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2138 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2139 (vneg_f64): New intrinsic.
2140 (vneg_s8): Asm replaced with C.
2141 (vneg_s16): Likewise.
2142 (vneg_s32): Likewise.
2143 (vneg_s64): New intrinsic.
2144 (vnegq_f32): Asm replaced with C.
2145 (vnegq_f64): Likewise.
2146 (vnegq_s8): Likewise.
2147 (vnegq_s16): Likewise.
2148 (vnegq_s32): Likewise.
2149 (vnegq_s64): Likewise.
2151 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2153 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2155 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2157 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2158 packed stack special handling.
2159 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2160 back to fixed stack slots for FPRs saved due to stdarg.
2162 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2164 * config/s390/s390.c (s390_frame_info): Restructure function.
2166 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2168 * config/s390/s390.c (struct s390_frame_layout): New field
2170 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2171 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2172 regs_ever_clobbered to char*.
2173 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2174 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2175 (s390_register_info): Enable FPR save slots. Move/Copy some
2176 functionality into ...
2177 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2178 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2180 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2181 not imply a stack frame.
2182 (s390_init_frame_layout): Remove variable clobbered_regs.
2183 (s390_update_register_info): Remove function.
2184 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2185 cannot be used for register renaming.
2186 (s390_hard_regno_scratch_ok): New function.
2187 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2188 (s390_initial_elimination_offset): Change offset calculation of
2189 the return address pointer.
2190 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2191 (restore_gprs): Set frame related flag.
2192 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2193 (s390_emit_prologue): Call s390_register_info instead of
2194 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2195 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2196 (s390_optimize_prologue): Call s390_optimize_register_info.
2197 Try to remove also FPR slot save/restore INSNs. Remove frame
2198 related flags from restore INSNs.
2200 2013-10-08 DJ Delorie <dj@redhat.com>
2202 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2205 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2206 avoid conflict with the MI use of %c.
2207 * config/rl78/rl78-real.md: change %c to %C throughout.
2208 * config/rl78/rl78-virt.md: Likewise.
2210 2013-10-08 Jan Hubicka <jh@suse.cz>
2212 * config/i386/i386.c (ix86_option_override_internal): Switch
2213 to SSE math for -ffast-math when target ISA supports SSE2.
2215 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2217 * tree-flow.h: Remove some prototypes.
2218 * tree.h: Remove some protypes, add a couple.
2219 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2220 using_eh_for_cleanups_p): Add interface routines for front ends.
2221 * tree-eh.h: New file. Add protoptyes.
2222 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2223 (add_stmt_to_eh_lp_fn): Make static.
2224 (lower_try_finally): Use new using_eh_for_cleanups_p.
2225 * emit-rtl.c: Include tree-eh.h.
2226 * gimple.h: Include tree-eh.h.
2228 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2230 PR tree-optimization/58480
2231 * tree-vrp.c (infer_nonnull_range): New function.
2232 (infer_value_range): Call infer_nonnull_range.
2234 2013-10-08 Dehao Chen <dehao@google.com>
2236 PR tree-optimization/58619
2237 * tree-inline.c (copy_phis_for_bb): Combine location data
2240 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2243 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2244 RTX_FRAME_RELATED_P on INSN.
2246 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2248 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2249 (altivec_expand_vec_perm_const): Call it.
2251 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2253 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2254 sequences for LE VSX loads and stores at expand time.
2255 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2257 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2258 (rs6000_gen_le_vsx_permute): New.
2259 (rs6000_gen_le_vsx_load): New.
2260 (rs6000_gen_le_vsx_store): New.
2261 (rs6000_gen_le_vsx_move): New.
2262 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2263 (*vsx_le_perm_load_v4si): New.
2264 (*vsx_le_perm_load_v8hi): New.
2265 (*vsx_le_perm_load_v16qi): New.
2266 (*vsx_le_perm_store_v2di): New.
2267 (*vsx_le_perm_store_v4si): New.
2268 (*vsx_le_perm_store_v8hi): New.
2269 (*vsx_le_perm_store_v16qi): New.
2270 (*vsx_xxpermdi2_le_<mode>): New.
2271 (*vsx_xxpermdi4_le_<mode>): New.
2272 (*vsx_xxpermdi8_le_V8HI): New.
2273 (*vsx_xxpermdi16_le_V16QI): New.
2274 (*vsx_lxvd2x2_le_<mode>): New.
2275 (*vsx_lxvd2x4_le_<mode>): New.
2276 (*vsx_lxvd2x8_le_V8HI): New.
2277 (*vsx_lxvd2x16_le_V16QI): New.
2278 (*vsx_stxvd2x2_le_<mode>): New.
2279 (*vsx_stxvd2x4_le_<mode>): New.
2280 (*vsx_stxvd2x8_le_V8HI): New.
2281 (*vsx_stxvd2x16_le_V16QI): New.
2283 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2285 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2287 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2289 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2290 loop to work also for 31bit ABI.
2291 Save the stack pointer for frame_size > 0.
2293 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2295 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2296 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2297 constraint letters from expanders.
2298 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2299 retry count to general_operand.
2300 ("tabort"): Give operand 0 a mode.
2301 ("tabort_1"): Add mode and constraint letter for operand 0.
2302 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2304 2013-10-04 Jeff Law <law@redhat.com>
2306 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2308 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2310 (thread_across_edge): Here. Call it.
2312 2013-10-04 Cary Coutant <ccoutant@google.com>
2314 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2315 discarding a location list expression (or a piece of one).
2317 2013-10-03 Jan Hubicka <jh@suse.cz>
2319 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2320 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2321 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2323 2013-10-03 Jan Hubicka <jh@suse.cz>
2325 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2326 accumulate-outgoing-args when producing unwind info.
2328 2013-10-03 Wei Mi <wmi@google.com>
2330 * lra-constraints.c (insert_move_for_subreg): New function
2331 extracted from simplify_operand_subreg.
2332 (simplify_operand_subreg): Add reload for paradoxical subreg.
2334 2013-10-03 Rong Xu <xur@google.com>
2336 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2337 the candidate of builtin_expect such that we should fix the
2338 size/time estimation.
2339 (estimate_function_body_sizes): Do the acutally size/time fix-up
2342 2013-10-03 Rong Xu <xur@google.com>
2344 * predict.c (tree_predict_by_opcode): Get the probability
2345 for builtin_expect from param builtin_expect_probability.
2346 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2347 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2348 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2350 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2353 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2354 * calls.c (alloca_call_p): Use get_callee_fndecl.
2355 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2356 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2358 (vrp_visit_stmt): Remove duplicated code.
2360 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2362 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2363 ceildf2, btruncdf2, instead of vsx_* name.
2365 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2366 iterators to only do V2DF and V4SF here. Move the DF code to
2367 rs6000.md where it is combined with SF mode. Replace <VSv> with
2368 just 'v' since only vector operations are handled with these insns
2369 after moving the DF support to rs6000.md.
2370 (vsx_sub<mode>3): Likewise.
2371 (vsx_mul<mode>3): Likewise.
2372 (vsx_div<mode>3): Likewise.
2373 (vsx_fre<mode>2): Likewise.
2374 (vsx_neg<mode>2): Likewise.
2375 (vsx_abs<mode>2): Likewise.
2376 (vsx_nabs<mode>2): Likewise.
2377 (vsx_smax<mode>3): Likewise.
2378 (vsx_smin<mode>3): Likewise.
2379 (vsx_sqrt<mode>2): Likewise.
2380 (vsx_rsqrte<mode>2): Likewise.
2381 (vsx_fms<mode>4): Likewise.
2382 (vsx_nfma<mode>4): Likewise.
2383 (vsx_copysign<mode>3): Likewise.
2384 (vsx_btrunc<mode>2): Likewise.
2385 (vsx_floor<mode>2): Likewise.
2386 (vsx_ceil<mode>2): Likewise.
2387 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2388 (vsx_sminsf3): Likewise.
2389 (vsx_fmadf4): Likewise.
2390 (vsx_fmsdf4): Likewise.
2391 (vsx_nfmadf4): Likewise.
2392 (vsx_nfmsdf4): Likewise.
2393 (vsx_cmpdf_internal1): Likewise.
2395 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2396 simpler to select whether a target has SPE or traditional floating
2397 point support in iterators.
2398 (TARGET_DF_SPE): Likewise.
2399 (TARGET_SF_FPR): Likewise.
2400 (TARGET_DF_FPR): Likewise.
2401 (TARGET_SF_INSN): Macros to say whether floating point support
2402 exists for a given operation for expanders.
2403 (TARGET_DF_INSN): Likewise.
2405 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2406 combining of SF/DF mode operations, using both traditional and VSX
2414 (abs<mode>2): Combine SF/DF modes using traditional floating point
2415 instructions. Add support for using the upper DF registers with
2416 VSX support, and SF registers with power8-vector support. Update
2417 expanders for operations supported by both the SPE and traditional
2418 floating point units.
2419 (abs<mode>2_fpr): Likewise.
2420 (nabs<mode>2): Likewise.
2421 (nabs<mode>2_fpr): Likewise.
2422 (neg<mode>2): Likewise.
2423 (neg<mode>2_fpr): Likewise.
2424 (add<mode>3): Likewise.
2425 (add<mode>3_fpr): Likewise.
2426 (sub<mode>3): Likewise.
2427 (sub<mode>3_fpr): Likewise.
2428 (mul<mode>3): Likewise.
2429 (mul<mode>3_fpr): Likewise.
2430 (div<mode>3): Likewise.
2431 (div<mode>3_fpr): Likewise.
2432 (sqrt<mode>3): Likewise.
2433 (sqrt<mode>3_fpr): Likewise.
2434 (fre<Fs>): Likewise.
2435 (rsqrt<mode>2): Likewise.
2436 (cmp<mode>_fpr): Likewise.
2437 (smax<mode>3): Likewise.
2438 (smin<mode>3): Likewise.
2439 (smax<mode>3_vsx): Likewise.
2440 (smin<mode>3_vsx): Likewise.
2441 (negsf2): Delete SF operations that are merged with DF.
2448 (fmasf4_fpr): Likewise.
2449 (fmssf4_fpr): Likewise.
2450 (nfmasf4_fpr): Likewise.
2451 (nfmssf4_fpr): Likewise.
2452 (sqrtsf2): Likewise.
2453 (rsqrtsf_internal1): Likewise.
2454 (smaxsf3): Likewise.
2455 (sminsf3): Likewise.
2456 (cmpsf_internal1): Likewise.
2457 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2458 (negdf2): Delete DF operations that are merged with SF.
2460 (nabsdf2): Likewise.
2466 (rsqrtdf_internal1): Likewise.
2467 (fmadf4_fpr): Likewise.
2468 (fmsdf4_fpr): Likewise.
2469 (nfmadf4_fpr): Likewise.
2470 (nfmsdf4_fpr): Likewise.
2471 (sqrtdf2): Likewise.
2472 (smaxdf3): Likewise.
2473 (smindf3): Likewise.
2474 (cmpdf_internal1): Likewise.
2475 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2476 (btrunc<mode>2): Delete separate expander, and combine with the
2477 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2478 (btrunc<mode>2_fpr): Likewise.
2479 (ceil<mode>2): Likewise.
2480 (ceil<mode>2_fpr): Likewise.
2481 (floor<mode>2): Likewise.
2482 (floor<mode>2_fpr): Likewise.
2483 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2484 Add support for using the upper registers with VSX and
2485 power8-vector. Move insns to be closer to the define_expands. On
2486 VSX systems, prefer the traditional form of FMA over the VSX
2487 version, since the traditional form allows the target not to
2488 overlap with the inputs.
2489 (fms<mode>4_fpr): Likewise.
2490 (nfma<mode>4_fpr): Likewise.
2491 (nfms<mode>4_fpr): Likewise.
2493 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2494 Richard Earnshaw <richard.earnshaw@arm.com>
2496 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2497 (struct mult_cost_table): Likewise.
2498 (struct mem_cost_table): Likewise.
2499 (struct fp_cost_table): Likewise.
2500 (struct vector_cost_table): Likewise.
2501 (cpu_cost_table): Likewise.
2502 * config/arm/arm.opt (mold-rts-costs): New option.
2503 (mnew-generic-costs): Likewise.
2504 * config/arm/arm.c (generic_extra_costs): New table.
2505 (cortexa15_extra_costs): Likewise.
2506 (arm_slowmul_tune): Use NULL as new costs.
2507 (arm_fastmul_tune): Likewise.
2508 (arm_strongarm_tune): Likewise.
2509 (arm_xscale_tune): Likewise.
2510 (arm_9e_tune): Likewise.
2511 (arm_v6t2_tune): Likewise.
2512 (arm_cortex_a5_tune): Likewise.
2513 (arm_cortex_a9_tune): Likewise.
2514 (arm_v6m_tune): Likewise.
2515 (arm_fa726te_tune): Likewise.
2516 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2517 (arm_cortex_tune): Use generict_extra_costs.
2518 (shifter_op_p): New function.
2519 (arm_unspec_cost): Likewise.
2520 (LIBCALL_COST): Define.
2521 (arm_new_rtx_costs): New function.
2522 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2523 table is available. Use old costs otherwise unless mnew-generic-costs
2525 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2526 (cpu_cost_table): Declare.
2528 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2531 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2532 (*subs_mul_imm_<mode>)
2533 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2534 (*sub_<shift>_<mode>)
2535 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2536 Remove k constraint.
2538 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2540 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2542 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2544 2013-10-02 Teresa Johnson <tejohnson@google.com>
2546 * predict.c (probably_never_executed): New function.
2547 (probably_never_executed_bb_p): Invoke probably_never_executed.
2548 (probably_never_executed_edge_p): Ditto.
2549 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2550 Treat profile insanities conservatively.
2552 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2554 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2556 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2558 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2559 use it. Use smaller increase for scratch. Don't increase reject
2560 for early clobber scratch.
2561 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2562 setting eliminated regs except setting fp from hfp.
2563 (lra_eliminate): Check lra_insn_recog_data on NULL.
2565 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2568 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2569 setting -mvsx-timode by default until the underlying problem is fixed.
2570 (RS6000_CPU, power7 defaults): Likewise.
2572 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2574 * config/x-linux (host-linux.o): Remove header dependencies.
2575 Use $(COMPILE) and $(POSTCOMPILE).
2576 * config/t-linux-android (linux-android.o): Ditto.
2578 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2580 * Makefile.in (expmed.o-warn): Remove.
2582 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2584 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2585 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2587 2013-10-02 Teresa Johnson <tejohnson@google.com>
2589 * dojump.c (do_jump_1): Divide probability between
2590 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2592 2013-10-02 Tom Tromey <tromey@redhat.com>
2594 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2596 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2598 * tree-flow.h: Remove some prototypes.
2599 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2600 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2601 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2602 mark_virtual_phi_result_for_renaming): Relocate here.
2603 * tree-into-ssa.h: Add prototypes.
2604 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2605 single_pred_before_succ_order.
2606 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2607 (nonfreeing_call_p) Move to gimple.c.
2608 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2610 * basic-block.h (single_pred_before_succ_order): Add prototype.
2611 * gimple.c (nonfreeing_call_p): Relocate here.
2612 * gimple.h: Add prototype.
2613 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2614 * tree-ssa-dom.h: New file. Relocate prototypes here.
2615 * tree-ssa.h: Include tree-ssa-dom.h.
2617 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2619 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2620 Use $(COMPILE) and $(POSTCOMPILE).
2622 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2624 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2626 * tree-flow.h: Remove some prototypes.
2627 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2628 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2629 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2630 propagate_tree_value*): Move from here to...
2631 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2632 propagate_tree_value*): Relocate here.
2633 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2634 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2635 * gimple-fold.c: Remove gimple-fold.h from include list.
2636 * tree-vrp.c: Remove gimple-fold.h from include list.
2637 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2638 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2639 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2640 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2641 * sese.c: Add tree-ssa-propagate.h to include list.
2643 2013-10-02 Richard Biener <rguenther@suse.de>
2645 * tree-loop-distribution.c: Include tree-vectorizer.h for
2647 (enum partition_kind): Remove PKIND_REDUCTION.
2648 (struct partition_s): Remove has_writes member, add reduction_p member.
2649 (partition_alloc): Adjust.
2650 (partition_builtin_p): Likewise.
2651 (partition_has_writes): Remove.
2652 (partition_reduction_p): New function.
2653 (partition_merge_into): Likewise.
2654 (generate_code_for_partition): Commonize builtin partition
2656 (rdg_cannot_recompute_vertex_p): Remove.
2657 (already_processed_vertex_p): Likewise.
2658 (rdg_flag_vertex): Do not set has_writes.
2659 (classify_partition): Adjust.
2660 (rdg_build_partitions): Do not set has_writes, treat all
2661 partitions as useful.
2662 (distribute_loop): Record number of library calls generated. Adjust.
2663 (tree_loop_distribution): Report number of loops and library
2664 calls generated as opt-info.
2666 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2668 * tree-flow.h: Include new .h files. Move prototypes.
2669 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2670 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2671 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2672 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2673 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2674 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2675 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2676 * tree.h (get_ref_base_and_extent): Move prototype out.
2677 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2679 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2680 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2681 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2682 * tree-scalar-evolution.c: Include tree.h.
2683 * sese.c: Include tree.h.
2684 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2685 * dwarf2out.c: Include tree-dfa.h.
2686 * tree-chrec.c: Include tree.h.
2687 * tree-data-ref.c: Include tree.h.
2689 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2691 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2694 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2696 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2697 Use $(COMPILE) and $(POSTCOMPILE).
2698 (sol2-cxx.o): Likewise.
2699 (sol2-stubs.o): Likewise.
2701 * config/x-solaris (host-solaris.o): Likewise.
2703 * config/sparc/t-sparc (sparc.o): Remove.
2704 (sparc-c.o): Remove header dependencies.
2705 Use $(COMPILE) and $(POSTCOMPILE).
2706 * config/sparc/x-sparc: Likewise.
2708 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2710 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2711 * config/arc/arc700.md: Likewise.
2712 * config/arc/arc-modes.def: Likewise.
2713 * config/arc/arc-simd.h: Likewise.
2714 * config/arc/t-arc-uClibc: Likewise.
2715 * config/arc/t-arc-newlib: Likewise.
2717 2013-10-02 Renlin Li <renlin.li@arm.com>
2719 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2721 (aarch64_expand_epilogue): Likewise.
2723 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2724 Yufeng Zhang <yufeng.zhang@arm.com>
2726 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2728 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2729 'base_in' represent a conversion and legal_cast_p_1 holds; set
2730 'base_in' with the returned value from get_unwidened.
2732 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2734 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2735 plus_constant is not used.
2737 2013-10-01 Wei Mi <wmi@google.com>
2739 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2740 * config/i386/i386.md: Add define_peephole2 to
2741 break partial reg stall for cvtss2sd/cvtsd2ss.
2743 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2745 * config/arc/arc.c (pass_arc_ifcvt::clone):
2746 Update for ctxt_ -> m_ctxt change.
2748 2013-10-01 Jeff Law <law@redhat.com>
2750 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2751 outgoing_edge and intermediate_edge fields. Instead store the path.
2752 (redirection_data::hash): Hash on the last edge's destination index.
2753 (redirection_data::equal): Check the entire thread path.
2754 (lookup_redirectio_data): Corresponding changes.
2755 (create_edge_and_update_destination_phis): Likewise.
2756 (thread_single_edge): Likewise.
2758 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2759 Diego Novillo <dnovillo@google.com>
2761 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2762 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2763 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2764 versions of these patterns.
2766 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2767 (__builtin_arc_aligned): Likewise.
2769 * config/arc/arc.md: Expand adc_0 comment stating the intended
2770 purpose and why it isn't ready.
2771 Replace commented out call_value_via_label_mixed with a
2772 plain comment about bl_s.
2774 * config/arc/arc.c (stdio.h): Don't include directly.
2775 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2776 Remove if (1) condition.
2777 (arc_encode_section_info): Fix comment.
2779 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2781 * config/arc/arc.c (arc_conditional_register_usage):
2782 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2783 Also set reg_alloc_order for DMA config regs.
2785 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2786 Jeremy Bennett <jeremy.bennett@embecosm.com>
2788 * doc/install.texi (--with-cpu): Mention ARC.
2789 (arc-*-elf32): New paragraph.
2790 (arc-linux-uclibc): Likewise.
2791 * doc/md.texi (Machine Constraints): Add ARC part.
2792 * doc/invoke.texi: (menu): Add ARC Options.
2793 (Machine Dependent Options) <ARC Options>: Add synopsis.
2794 (node ARC Options): Add.
2795 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2796 (ARC Built-in Functions): New section defining
2797 generic ARC built-in functions.
2798 (ARC SIMD Built-in Functions): New section defining SIMD specific
2800 (Declaring Attributes of Functions): Extended
2801 description of short_call and long_call attributes for ARC and
2802 added index entries.
2804 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2805 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2806 Joern Rennecke <joern.rennecke@embecosm.com>
2807 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2808 Brendan Kehoe <brendan@zen.org>
2809 Michael Eager <eager@eagercon.com>
2810 Simon Cook <simon.cook@embecosm.com>
2811 Jeremy Bennett <jeremy.bennett@embecosm.com>
2813 * config/arc, common/config/arc: New directories.
2815 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2816 Brendan Kehoe <brendan@zen.org>
2817 Simon Cook <simon.cook@embecosm.com>
2819 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2821 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2823 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2824 * tree-ssa-coalesce.h: New. Move prototype to here.
2825 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2826 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2827 (gimple_can_coalesce_p): Move to...
2828 * gimple.c (gimple_can_coalesce_p): Here.
2830 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2832 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2833 (dump_decl_set): Move to gimple.c.
2834 * gimple.h: Don't include tree-ssa-operands.h.
2835 (dump_decl_set): Add prototype.
2836 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2837 Move to gimple-ssa.h.
2838 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2839 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2840 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2841 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2842 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2844 (dump_decl_set): Relocate here.
2845 * gimple-ssa.h: New file.
2846 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2847 Relocate from gimple.h.
2848 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2849 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2850 swap_tree_operands and remove non-ssa path.
2851 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2852 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2854 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2855 vect_is_simple_reduction_1): Use swap_ssa_operands.
2856 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2857 (enum need_phi_state): Move to tree-into-ssa.c.
2858 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2859 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2860 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2861 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2862 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2863 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2864 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2865 num_imm_uses): Move to ssa-iterators.h.
2866 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2867 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2869 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2870 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2871 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2872 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2873 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2874 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2875 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2876 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2877 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2878 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2879 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2880 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2881 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2882 (set_phi_nodes): Move to tree-phinodes.h.
2883 * tree-ssa-operands.h (enum ssa_op_iter_type,
2884 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2885 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2886 (dump_decl_set): Remove prototype.
2887 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2888 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2889 (set_phi_nodes): Relocate from tree-flow-inline.h.
2890 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2892 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2893 include list. Temporarily add gimple.h to include list.
2894 * ssa-iterators.h: New file.
2895 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2896 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2897 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2898 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2899 Relocate from tree-ssa-operands.h.
2900 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2901 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2902 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2903 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2904 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2905 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2906 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2907 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2908 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2909 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2910 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2911 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2912 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2913 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2914 Relocate from tree-flow-inline.h.
2915 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2917 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2920 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2921 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2922 Insert '\t' to output template.
2923 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2924 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2925 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2926 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2928 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
2930 * doc/install.texi (Host/target specific installation notes for GCC):
2931 Put @anchor before @heading.
2932 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2933 Use @email for email addresses.
2935 2013-10-01 Jeff Law <law@redhat.com>
2937 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
2938 a vec. Only delete the path if we create one without successfully
2939 registering a jump thread.
2940 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
2942 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
2943 (paths): New vector of jump threading paths.
2944 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
2945 (THREAD_PATH): New accessor macro for the entire thread path.
2946 (lookup_redirection_data): Get intermediate and final outgoing edge
2947 from the thread path.
2948 (create_edge_and_update_destination_phis): Copy the threading path.
2949 (ssa_fix_duplicate_block_edges): Get edges and block types from the
2950 jump threading path.
2951 (ssa_redirect_edges): Get edges and block types from the jump threading
2952 path. Free the path vector.
2953 (thread_block): Get edges from the jump threading path. Look at the
2954 entire path to see if we thread to a loop exit. If we cancel a jump
2955 thread request, then free the path vector.
2956 (thread_single_edge): Get edges and block types from the jump threading
2957 path. Free the path vector.
2958 (thread_through_loop_header): Get edges and block types from the jump
2959 threading path. Free the path vector.
2960 (mark_threaded_blocks): Iterate over the vector of paths and store
2961 the path on the appropriate edge. Get edges and block types from the
2962 jump threading path.
2963 (mark_threaded_blocks): Get edges and block types from the jump
2964 threading path. Free the path vector.
2965 (thread_through_all_blocks): Use the vector of paths rather than
2966 a vector of 3-edge sets.
2967 (register_jump_thread): Accept pointer to a path vector rather
2968 than the path vector itself. Store the path vector for later use.
2971 2013-10-01 Jakub Jelinek <jakub@redhat.com>
2972 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2975 * config/s390/s390.c (s390_split_branches): Modify check for table
2977 (s390_chunkify_start): Rearrange table jump insn check in order to
2978 deal with compare and branch insns correctly.
2980 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
2984 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
2985 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
2986 define_insn_and_split.
2987 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
2988 (shiftsi3_compare): New pattern.
2990 * config/arm/unspecs.md (UNSPEC_RRX): New.
2992 2013-10-01 Alan Modra <amodra@gmail.com>
2994 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
2995 casing inout operands.
2997 2013-10-01 Richard Biener <rguenther@suse.de>
2999 PR tree-optimization/58553
3000 * tree-loop-distribution.c (struct partition_s): Add niter member.
3001 (classify_partition): Populate niter member for the partition
3002 and properly identify whether the relevant store happens before
3003 or after the loop exit.
3004 (generate_memset_builtin): Use niter member from the partition.
3005 (generate_memcpy_builtin): Likewise.
3007 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3009 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3010 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3012 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3014 * basic-block.h (edge_list): Prefix member names with "m_".
3015 * context.h (context): Likewise.
3016 * domwalk.h (dom_walker): Likewise.
3017 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3018 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3019 * hash-table.h (hash_table): Likewise.
3020 * machmode.h (bit_field_mode_iterator): Likewise.
3021 * pass_manager.h (pass_list): Likewise.
3022 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3023 * tree-pass.h (pass_data): Likewise.
3024 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3025 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3026 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3027 * asan.c (pass_data_asan): Update accordingly.
3028 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3029 (control_dependences::control_dependences): Likewise.
3030 (control_dependences::~control_dependences): Likewise.
3031 (control_dependences::~control_dependences): Likewise.
3032 (control_dependences::get_edges_dependent_on): Likewise.
3033 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3034 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3035 * context.c (gcc::context::context): Likewise.
3036 * cprop.c (pass_rtl_cprop::clone): Likewise.
3037 * domwalk.c (dom_walker::walk): Likewise.
3038 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3039 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3040 * mode-switching.c (pass_mode_switching::clone): Likewise.
3041 * passes.c (opt_pass::opt_pass): Likewise.
3042 (pass_manager::pass_manager): Likewise.
3043 * predict.c (pass_strip_predict_hints::clone): Likewise.
3044 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3045 (pass_split_all_insns::clone): Likewise.
3046 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3048 (bit_field_mode_iterator::next_mode): Likewise.
3049 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3050 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3051 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3052 * tree-complex.c (pass_lower_complex::clone): Likewise.
3053 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3054 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3055 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3056 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3057 (pass_fold_builtins::clone): Likewise.
3058 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3059 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3060 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3061 (pass_cd_dce::clone): Likewise.
3062 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3063 (pass_phi_only_cprop::clone): Likewise.
3064 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3065 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3066 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3067 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3068 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3069 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3070 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3071 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3072 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3073 * tree-vrp.c (pass_vrp::clone): Likewise.
3074 * tsan.c (pass_tsan::clone): Likewise.
3076 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3079 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3080 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3083 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3084 optimization, punt if sign_bit_p looked through any zero extension.
3086 2013-09-30 Teresa Johnson <tejohnson@google.com>
3088 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3089 Update redirected out edge count in joiner case.
3090 (ssa_redirect_edges): Common the joiner and non-joiner cases
3091 so that joiner case gets profile updates.
3093 2013-09-30 Richard Biener <rguenther@suse.de>
3095 PR tree-optimization/58554
3096 * tree-loop-distribution.c (classify_partition): Require
3097 unconditionally executed stores for memcpy and memset recognition.
3098 (tree_loop_distribution): Calculate dominance info.
3100 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3102 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3103 (NO_PROFILE_COUNTERS): Likewise.
3104 (PROFILE_HOOK): Likewise.
3105 (FUNCTION_PROFILER): Likewise.
3106 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3108 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3110 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3111 calls and defines in TARGET_MACHO conditional.
3112 (load_macho_picbase_di): Likewise.
3113 (reload_macho_picbase): Likewise.
3114 (reload_macho_picbase_si): Likewise.
3115 (reload_macho_picbase_di): Likewise.
3116 (nonlocal_goto_receiver): Likewise.
3118 2013-09-30 Nick Clifton <nickc@redhat.com>
3120 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3121 that use the MSP430X ISA.
3122 (msp430_option_override): Scan -mmcu command line option for any
3123 MCU name that supports the MSP430X ISA.
3124 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3125 -mmcu options which enable the MSP430X ISA.
3127 2013-09-30 Richard Biener <rguenther@suse.de>
3130 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3131 before looking for setjmp-like calls.
3133 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3136 * config/darwin-protos.h (machopic_get_function_picbase): New.
3137 * config/darwin.c (machopic_get_function_picbase): New.
3138 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3139 label for a new func. (load_macho_picbase_di): Likewise.
3140 (reload_macho_picbase): New expand.
3141 (reload_macho_picbase_si): New insn.
3142 (reload_macho_picbase_di): New insn.
3143 (nonlocal_goto_receiver): New define and split.
3144 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3145 (unspecv enum): Add UNSPECV_NLGR.
3147 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3149 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3150 that altivec registers are correctly sized on Darwin.
3152 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3154 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3155 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3156 * config/x-darwin (host-darwin.o): Likewise.
3157 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3158 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3159 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3161 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3163 * doc/invoke.texi: Fix usage of @tie{} command.
3165 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3167 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3169 2013-09-28 Jan Hubicka <jh@suse.cz>
3171 * config/i386/x86-tune.def: Add documentation for each of the options;
3174 2013-09-28 Jan Hubicka <jh@suse.cz>
3176 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3178 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3179 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3180 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3181 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3182 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3184 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3186 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3187 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3188 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3189 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3190 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3191 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3192 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3193 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3194 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3195 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3196 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3197 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3198 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3199 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3200 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3201 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3202 gimple.h, godump.c, graphite-clast-to-gimple.c,
3203 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3204 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3205 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3206 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3207 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3208 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3209 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3210 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3211 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3212 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3213 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3214 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3215 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3216 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3217 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3218 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3219 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3220 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3221 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3222 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3223 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3224 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3225 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3226 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3227 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3228 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3229 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3230 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3231 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3232 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3233 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3234 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3235 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3236 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3237 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3238 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3239 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3240 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3241 whitespace before "(".
3243 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3245 * expr.h (extract_bit_field): Remove packedp parameter.
3246 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3247 from forward declaration.
3248 (store_split_bit_field): Remove packedp arg from calls to
3249 extract_fixed_bit_field.
3250 (extract_bit_field_1): Remove packedp parameter and packedp
3251 argument from recursive calls and calls to extract_fixed_bit_field.
3252 (extract_bit_field): Remove packedp parameter and corresponding
3253 arg to extract_bit_field_1.
3254 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3256 (extract_split_bit_field): Remove packedp arg from call to
3257 extract_fixed_bit_field.
3258 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3259 (copy_blkmode_from_reg): Likewise.
3260 (copy_blkmode_to_reg): Likewise.
3261 (read_complex_part): Likewise.
3262 (store_field): Likewise.
3263 (expand_expr_real_1): Likewise.
3264 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3265 to extract_bit_field.
3266 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3267 call to extract_bit_field.
3268 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3269 call to extract_bit_field.
3270 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3271 and special packedp behavior from -fstrict-volatile-bitfields
3274 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3276 * lra-eliminations.c (init_elim_table): Guard value_p.
3278 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3280 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3281 DFmode, DImode, and SFmode in the upper VSX registers based on the
3282 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3283 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3284 of the VSX register handling.
3286 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3288 (zero_extendsidi2_lfiwzx): Add support for loading into the
3289 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3290 only do VSX memory options on Altivec registers.
3291 (extendsidi2_lfiwax): Likewise.
3292 (extendsfdf2_fpr): Likewise.
3293 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3294 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3295 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3296 (movdi_internal64): Likewise.
3298 2013-09-27 Xinliang David Li <davidxl@google.com>
3300 * opts.c (finish_options): Adjust parameters
3301 according to vect cost model.
3302 (common_handle_option): Set dynamic vect cost
3304 targhooks.c (default_add_stmt_cost): Compute stmt cost
3306 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3307 Use helper function.
3308 * tree-vectorizer.h (unlimited_cost_model): New function.
3309 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3310 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3312 (vect_enhance_data_refs_alignment): Ditto.
3313 * flag-types.h: New enum.
3314 * common/config/i386/i386-common.c (ix86_option_init_struct):
3315 No need to initialize vect_cost_model flag.
3316 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3319 2013-09-27 Diego Novillo <dnovillo@google.com>
3321 * gimple.h (enum ssa_mode): Remove.
3323 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3325 * cfgloop.h (number_of_loops): Fix typo in check for null.
3327 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3330 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3331 are children of outermost saved_cfun's loop, and set it up to
3332 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3333 if loop != loop0 and SESE region contains bbs that belong to loop0.
3335 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3337 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3338 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3339 (set_address_segment, set_address_base, set_address_index)
3340 (set_address_disp): Accept the argument unconditionally.
3341 (baseness): Remove must_be_base_p and must_be_index_p checks.
3342 (decompose_normal_address): Classify as much as possible in the
3345 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3347 * cse.c (count_reg_usage): Handle INT_LIST.
3348 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3349 * reginfo.c (reg_scan_mark_refs): Likewise.
3350 * reload1.c (eliminate_regs_1): Likewise.
3352 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3355 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3358 2013-09-27 Richard Biener <rguenther@suse.de>
3360 PR tree-optimization/58459
3361 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3362 restriction not propagating into loops.
3364 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3366 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3367 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3368 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3370 2013-09-26 Richard Biener <rguenther@suse.de>
3372 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3374 2013-09-26 Richard Biener <rguenther@suse.de>
3376 * alias.h (component_uses_parent_alias_set): Rename to ...
3377 (component_uses_parent_alias_set_from): ... this.
3378 * alias.c (component_uses_parent_alias_set): Rename to ...
3379 (component_uses_parent_alias_set_from): ... this and return
3381 (reference_alias_ptr_type_1): Use the result from
3382 component_uses_parent_alias_set_from instead of stripping
3383 components one at a time.
3384 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3386 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3388 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3389 Move prototypes to...
3390 * tree-ssa-ter.h: New File. Move prototypes here.
3391 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3392 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3393 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3394 from is_replaceable_p.
3395 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3396 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3397 (process_replaceable): Use ter_is_replaceable_p.
3398 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3399 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3400 newly refactored ssa_is_replaceable_p.
3401 * cfgexpand.c: Include tree-outof-ssa.h.
3402 * ssaexpand.h: Delete.
3404 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3406 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3407 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3409 (create_gimple_tmp): Delete.
3410 (get_expr_type, build_assign, build_type_cast): Move to...
3411 * gimple-builder.c: New File.
3412 (get_expr_type): Relocate from gimple.c.
3413 (build_assign, build_type_cast): Change to only create ssanames.
3414 * gimple.h: Move prototypes to...
3415 * gimple-builder.h: New File. Here.
3416 * tree-ssa.h: And here.
3417 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3418 count_uses_and_derefs): Relocate from gimple.c.
3419 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3420 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3421 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3422 gimple_replace_ssa_lhs.
3423 * asan.c: Include gimple-builder.h.
3424 * Makefile.in: Add gimple-builder.o.
3426 2013-09-26 Richard Biener <rguenther@suse.de>
3428 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3429 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3430 (loe_visit_block): Use gcc_checking_assert.
3431 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3432 gimple_assign_ssa_name_copy_p.
3433 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3436 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3438 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3439 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3441 2013-09-26 Richard Biener <rguenther@suse.de>
3443 PR tree-optimization/58539
3444 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3445 the fact that debug statements are not taking part in loop-closed
3448 2013-09-26 Nick Clifton <nickc@redhat.com>
3450 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3451 time warning message.
3452 (msp430_print_operand_raw): Delete unused letter parameter.
3453 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3454 (msp430_print_operand_address): New function.
3455 (msp430_print_operand): Move address printing code from here to
3457 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3459 (zero_extendpsisi2): Likewise.
3460 (extendpsisi2): New pattern.
3461 (andneghi3): New pattern.
3463 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3465 * config/aarch64/aarch64.opt (mlra): New option.
3466 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3467 (TARGET_LRA_P): Define.
3469 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3471 * expr.c (expand_assignment): Remove obsolete comment.
3473 2013-09-25 Jeff Law <law@redhat.com>
3475 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3476 tree-ssa-threadupdate.h.
3477 (register_jump_thread): Similarly.
3478 * tree-ssa-threadupdate.h: New header file.
3479 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3480 * tree-vrp.c: Likewise.
3481 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3482 (thread_around_empty_blocks): Change type of path vector argument to
3483 an edge,type pair from just an edge. Initialize both elements when
3484 appending to a jump threading path. Tweak references to elements
3486 (thread_across_edge): Similarly. Release memory for the elements
3488 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3489 (dump_jump_thread_path): New function broken out from
3490 register_jump_thread.
3491 (register_jump_thread): Use dump_jump_thread_path. Change type of
3492 path vector entries. Search the path for NULL edges and dump
3493 the path if one is found. Tweak the conversion of path to 3-edge
3494 form to use the block copy type information embedded in the path.
3496 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3498 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3500 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3501 Vladimir Makarov <vmakarov@redhat.com>
3503 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3504 from the least significant bit.
3505 (strip_address_mutations): Add bitfield operations handling.
3506 (must_be_index_p): Add shifting and rotate operations handling.
3507 (set_address_base): Use must_be_base_p predicate.
3508 (set_address_index): Use must_be_index_p predicate.
3510 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3511 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3512 Sergey Lega <sergey.s.lega@intel.com>
3513 Anna Tikhonova <anna.tikhonova@intel.com>
3514 Ilya Tocar <ilya.tocar@intel.com>
3515 Andrey Turetskiy <andrey.turetskiy@intel.com>
3516 Ilya Verbin <ilya.verbin@intel.com>
3517 Kirill Yukhin <kirill.yukhin@intel.com>
3518 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3520 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3522 (ix86_expand_vector_move_misalign): Support new unaligned load and
3523 stores and use new names.
3524 (CODE_FOR_sse2_storedqu): Rename to ...
3525 (CODE_FOR_sse2_storedquv16qi): ... this.
3526 (CODE_FOR_sse2_loaddqu): Rename to ...
3527 (CODE_FOR_sse2_loaddquv16qi): ... this.
3528 (CODE_FOR_avx_loaddqu256): Rename to ...
3529 (CODE_FOR_avx_loaddquv32qi): ... this.
3530 (CODE_FOR_avx_storedqu256): Rename to ...
3531 (CODE_FOR_avx_storedquv32qi): ... this.
3532 * config/i386/i386.md (fpint_logic): New.
3533 * config/i386/sse.md (VMOVE): Extend for AVX512.
3537 (VI_UNALIGNED_LOADSTORE): Ditto.
3538 (sse2_avx_avx512f): Ditto.
3539 (sse2_avx2): Extend for AVX512.
3540 (sse4_1_avx2): Ditto.
3541 (avx2_avx512f): New.
3542 (sse): Extend for AVX512.
3545 (avxsizesuffix): Ditto.
3546 (sseintvecmode): Ditto.
3548 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3549 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3550 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3551 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3552 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3553 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3554 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3555 (STORENT_MODE): Extend for AVX512.
3556 (*absneg<mode>2): Replace constraint "x" with "v".
3557 (*mul<mode>3): Ditto.
3558 (*ieee_smin<mode>3): Ditto.
3559 (*ieee_smax<mode>3): Ditto.
3560 (avx_cmp<mode>3): Replace VF with VF_128_256.
3561 (*<sse>_maskcmp<mode>3_comm): Ditto.
3562 (<sse>_maskcmp<mode>3): Ditto.
3563 (<sse>_andnot<mode>3): Extend for AVX512.
3564 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3565 (<code><mode>3, fpint_logic): New.
3566 (*<code><mode>3): Extend for AVX512.
3567 (avx512flogicsuff): New.
3568 (avx512f_<logic><mode>): Ditto.
3569 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3571 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3572 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3573 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3574 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3575 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3576 (xop_vpermil2<mode>3): Ditto.
3577 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3578 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3579 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3580 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3582 2013-09-25 Tom Tromey <tromey@redhat.com>
3584 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3585 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3586 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3587 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3588 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3589 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3590 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3591 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3592 (GRAPHITE_HTAB_H): Remove.
3594 2013-09-25 Tom Tromey <tromey@redhat.com>
3596 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3598 2013-09-25 Tom Tromey <tromey@redhat.com>
3600 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3602 2013-09-25 Tom Tromey <tromey@redhat.com>
3604 * config/i386/t-i386 (i386.o): Remove.
3605 (i386-c.o): Use COMPILE and POSTCOMPILE.
3607 2013-09-25 Tom Tromey <tromey@redhat.com>
3609 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3611 2013-09-25 Tom Tromey <tromey@redhat.com>
3613 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3614 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3615 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3616 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3617 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3618 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3619 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3620 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3621 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3622 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3623 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3624 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3625 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3626 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3627 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3628 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3629 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3630 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3631 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3632 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3633 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3634 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3635 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3636 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3637 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3638 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3639 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3640 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3641 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3642 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3643 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3644 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3645 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3646 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3647 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3648 (graphite-clast-to-gimple.o, graphite-dependences.o)
3649 (graphite-interchange.o, graphite-poly.o)
3650 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3651 (graphite-optimize-isl.o, tree-vect-loop.o)
3652 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3653 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3654 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3655 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3656 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3657 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3658 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3659 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3660 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3661 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3662 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3663 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3664 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3665 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3666 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3667 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3668 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3669 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3670 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3671 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3672 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3673 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3674 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3675 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3676 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3677 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3678 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3679 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3680 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3681 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3682 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3683 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3684 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3685 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3686 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3687 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3688 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3689 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3690 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3691 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3692 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3693 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3694 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3695 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3696 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3697 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3698 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3699 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3700 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3701 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3702 (gcov-dump.o): Remove.
3703 (default-c.o): Use COMPILE and POSTCOMPILE.
3704 (CFLAGS-gcc.o): New variable.
3705 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3707 2013-09-25 Tom Tromey <tromey@redhat.com>
3709 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3710 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3711 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3712 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3713 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3714 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3715 (c-family/array-notation-common.o, c-family/stub-objc.o)
3716 (c-family/c-ubsan.o): Remove.
3718 2013-09-25 Tom Tromey <tromey@redhat.com>
3720 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3722 2013-09-25 Tom Tromey <tromey@redhat.com>
3724 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3725 to add -DENABLE_SHARED_LIBGCC.
3726 (gcc.o): Don't use subshell.
3728 2013-09-25 Tom Tromey <tromey@redhat.com>
3730 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3731 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3732 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3733 * configure, config.in: Rebuild.
3735 2013-09-25 Tom Tromey <tromey@redhat.com>
3737 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3738 (COMPILE, POSTCOMPILE): New variables.
3739 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3740 (DEPFILES): New variable.
3741 Include ".Po" files.
3742 * configure.ac: Add checks for dependency checking.
3743 * configure, aclocal.m4: Regenerate.
3745 2013-09-25 Tom Tromey <tromey@redhat.com>
3747 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3748 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3750 2013-09-25 Tom Tromey <tromey@redhat.com>
3752 * Makefile.in (generated_files): Add options.h,
3753 target-hooks-def.h, insn-opinit.h,
3754 common/common-target-hooks-def.h, pass-instances.def,
3755 c-family/c-target-hooks-def.h.
3757 2013-09-25 Jeff Law <law@redhat.com>
3759 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3760 than foo[foo.length () - 1] to access last member in a vec.
3761 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3763 2013-09-25 Richard Biener <rguenther@suse.de>
3766 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3768 2013-09-25 Jan Hubicka <jh@suse.cz>
3770 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3773 2013-09-25 Marek Polacek <polacek@redhat.com>
3776 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3777 when determining the type name.
3779 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3781 * config/sh/sh.md: Fix formatting.
3783 2013-09-24 Xinliang David Li <davidxl@google.com>
3785 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3786 max peel iterations parameter.
3787 * param.def: New parameter.
3788 * doc/invoke.texi: Document New parameter.
3790 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3792 * gimple-pretty-print.c: Various whitespace tweaks.
3793 * tree-core.h: Likewise.
3794 * tree-pretty-print.c: Likewise.
3795 * tree-ssa-alias.c: Likewise.
3796 * tree-ssa-copy.c: Likewise.
3797 * tree-ssanames.c: Likewise.
3798 * tree-ssanames.h: Likewise.
3799 * tree-vrp.c: Likewise.
3801 2013-09-24 Alan Modra <amodra@gmail.com>
3805 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3806 for output operands that disallow regs. Don't use EXPAND_WRITE on
3809 2013-09-24 Richard Biener <rguenther@suse.de>
3812 * tree.c (reference_alias_ptr_type): Move ...
3813 * alias.c (reference_alias_ptr_type): ... here and implement
3814 in terms of the new reference_alias_ptr_type_1.
3815 (ref_all_alias_ptr_type_p): New helper.
3816 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3817 use ref_all_alias_ptr_type_p.
3818 (get_deref_alias_set): Add flag_strict_aliasing check here.
3819 (reference_alias_ptr_type_1): New function, split out from ...
3820 (get_alias_set): ... here.
3821 (alias_ptr_types_compatible_p): New function.
3822 * alias.h (reference_alias_ptr_type): Declare.
3823 (alias_ptr_types_compatible_p): Likewise.
3824 * tree.h (reference_alias_ptr_type): Remove.
3825 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3826 to compare MEM_REF alias types.
3828 2013-09-24 Richard Biener <rguenther@suse.de>
3830 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3832 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3834 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3835 reload helper function arrays into a single array reg_addr.
3836 (reload_fpr_gpr): Likewise.
3837 (reload_gpr_vsx): Likewise.
3838 (reload_vsx_gpr): Likewise.
3839 (struct rs6000_reg_addr): Likewise.
3840 (reg_addr): Likewise.
3841 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3842 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3843 (rs6000_init_hard_regno_mode_ok): Likewise.
3844 (rs6000_secondary_reload_direct_move): Likewise.
3845 (rs6000_secondary_reload): Likewise.
3847 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3848 constraints: wu, ww, and wy. Repurpose wv constraint added during
3849 power8 changes. Put wg constraint in alphabetical order.
3851 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3852 for future work to add ISA 2.07 VSX single precision support.
3853 (-mvsx-scalar-double): Change default from -1 to 1, update
3854 documentation comment.
3855 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3856 (-mupper-regs-df): New debug switch to control whether DF values
3857 can go in the traditional Altivec registers.
3858 (-mupper-regs-sf): New debug switch to control whether SF values
3859 can go in the traditional Altivec registers.
3861 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3863 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3864 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3865 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3866 (rs6000_debug_legitimate_address_p): Print if we are running
3867 before, during, or after reload.
3868 (rs6000_secondary_reload): Add a comment.
3869 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3871 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3872 constraints. Update documentation string.
3873 (wd constraint): Likewise.
3874 (wf constraint): Likewise.
3875 (wg constraint): Likewise.
3876 (wn constraint): Likewise.
3877 (ws constraint): Likewise.
3878 (wt constraint): Likewise.
3879 (wx constraint): Likewise.
3880 (wz constraint): Likewise.
3881 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3883 (ww constraint): Likewise.
3884 (wy constraint): Likewise.
3885 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3886 the previous submissions.
3887 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3889 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3891 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3892 (REG_BR_PROB): Say that the probability is stored in an int_list.
3893 * reg-notes.def: Update commentary to mention INT_LIST.
3894 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3895 (INT_LIST): New rtx.
3896 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3897 * rtlanal.c (int_reg_note_p): New function.
3898 (alloc_reg_note): Assert that the note does not have an int argument.
3899 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3900 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3901 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3902 rather than an INSN_LIST. Handle INT_LIST.
3903 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3904 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3905 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3906 Manipulate them as ints rather than rtxes.
3907 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3908 * regmove.c (copy_src_to_dest): Likewise.
3909 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3911 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3912 into the cases that need it.
3913 * config/arm/arm.c (arm_unwind_emit): Likewise.
3915 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3916 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3917 * loop-doloop.c (add_test, doloop_modify): Likewise.
3918 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3919 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3920 * predict.c (combine_predictions_for_insn): Likewise.
3921 * print-rtl.c (print_rtx): Handle INT_LIST.
3922 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3923 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3924 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3925 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3926 (ix86_split_fp_branch, predict_jump): Likewise.
3927 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3928 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3929 * config/spu/spu.c (ea_load_store_inline): Likewise.
3931 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3932 value of a REG_BR_PROB note.
3933 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3934 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3935 * emit-rtl.c (try_split): Likewise.
3936 * predict.c (br_prob_note_reliable_p): Likewise.
3937 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
3938 * reorg.c (mostly_true_jump): Likewise.
3939 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
3940 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
3941 * config/i386/i386.c (ix86_print_operand): Likewise.
3942 * config/ia64/ia64.c (ia64_print_operand): Likewise.
3943 * config/mmix/mmix.c (mmix_print_operand): Likewise.
3944 * config/rs6000/rs6000.c (output_cbranch): Likewise.
3945 * config/s390/s390.c (s390_expand_tbegin): Likewise.
3946 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
3947 * config/sparc/sparc.c (output_cbranch): Likewise.
3948 * config/spu/spu.c (get_branch_target): Likewise.
3949 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
3950 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
3952 2013-09-23 Jan Hubicka <jh@suse.cz>
3954 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
3956 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
3957 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
3959 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
3960 (gimple_fold_call): Dump inconsistent devirtualizations; add
3961 sanity check for type based devirtualizations.
3962 * ipa-prop.c: Include ipa-utils.h
3963 (ipa_intraprocedural_devirtualization): Add sanity check.
3964 (try_make_edge_direct_virtual_call): Likewise.
3966 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
3968 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
3969 assignment statements.
3971 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
3973 * gimple-pretty-print.c (dump_ssaname_info): New function.
3974 (dump_gimple_phi): Call it.
3975 (pp_gimple_stmt_1): Likewise.
3976 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
3977 (range_info_def): Declare.
3978 * tree-pretty-print.c (pp_double_int): New function.
3979 (dump_generic_node): Call it.
3980 * tree-pretty-print.h (pp_double_int): Declare.
3981 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
3982 * tree-ssanames.h (range_info_def): New structure.
3983 (value_range_type): Move definition here.
3984 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
3986 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
3988 (set_range_info): New function.
3989 (get_range_info): Likewise.
3990 (duplicate_ssa_name_range_info): Likewise.
3991 (duplicate_ssa_name_fn): Check pointer type and call
3992 duplicate_ssa_name_range_info.
3993 * tree-ssa-copy.c (fini_copy_prop): Likewise.
3994 * tree-vrp.c (value_range_type): Remove definition, now in
3996 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
3997 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
3998 (SSA_NAME_RANGE_INFO): New macro.
4000 2013-09-23 Richard Biener <rguenther@suse.de>
4002 PR tree-optimization/58464
4003 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4004 (phi_trans_add): Change to add conditionally on being not present.
4005 (phi_translate_1): Remove recursion detection here.
4006 (phi_translate): Pre-seed the cache with NULL to catch
4007 recursion here in a more generic way.
4008 (bitmap_find_leader): Adjust comment.
4009 (get_representative_for): Dump value-numbers.
4010 (create_expression_by_pieces): Likewise.
4011 (insert_into_preds_of_block): Likewise.
4013 2013-09-23 Christian Bruel <christian.bruel@st.com>
4016 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4017 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4019 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4022 2013-09-20 Renlin Li <renlin.li@arm.com>
4024 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4026 (aarch64_expand_epilogue): Likewise.
4027 (aarch64_legitimize_reload_address): Likewise.
4029 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4031 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4032 NULL_TREE before pushing them onto the vector. Likewise for labels.
4034 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4036 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4037 comparison operators when -fno-trapping-math is in effect.
4038 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4039 comparison operators in TFmode and assert that unsupported operators
4041 (ia64_print_operand): Likewise.
4043 2013-09-21 Jan Hubicka <jh@suse.cz>
4045 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4046 (sse_typeless_stores): Enable for core
4047 (sse_load0_by_pxor): Likewise.
4048 (four_jump_limit): Disable for core.
4049 (pad_returns): Likewise.
4050 (avoid_vector_decode): Likewise.
4051 (fuse_cmp_and_branch): Enable for cores.
4052 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4054 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4057 * config/pa/pa.c (pa_option_override): Disable auto increment and
4058 decrement instructions until reload is completed.
4060 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4061 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4062 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4064 2013-09-20 DJ Delorie <dj@redhat.com>
4065 Nick Clifton <nickc@redhat.com>
4067 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4068 (need_to_save): Save bank 0 on interrupts.
4069 (characterize_address): Strip far address wrappers.
4070 (rl78_as_legitimate_address): Likewise.
4071 (transcode_memory_rtx): Likewise.
4072 (rl78_peep_movhi_p): Disable this peephole after devirt.
4073 (rl78_propogate_register_origins): Forget all origins when a
4075 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4076 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4078 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4079 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4080 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4082 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4084 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4086 * config/pa/pa.md: In "scc" insn patterns, change output template to
4087 handle const0_rtx in reg_or_0_operand operands.
4089 2013-09-20 Martin Husemann <martin@NetBSD.org>
4092 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4093 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4095 2013-09-20 Richard Biener <rguenther@suse.de>
4098 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4099 remembering SSA name version and block index.
4100 (new_scev_info_str): Adjust.
4101 (hash_scev_info): Likewise. Also hash the block index.
4102 (eq_scev_info): Adjust.
4103 (find_var_scev_info): Likewise.
4104 (struct instantiate_cache_entry): Remove.
4105 (struct instantiate_cache_type): Use a htab to map name, block
4107 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4108 (get_instantiated_value_entry): Likewise.
4109 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4110 (instantiate_scev_name): Adjust.
4112 2013-09-20 Jeff Law <law@redhat.com>
4114 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4116 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4118 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4119 Call aarch64_simd_expand_args to update op[argc].
4121 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4123 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4126 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4128 * gengtype.c (file_rules): Added rule for *.cc files.
4129 (get_output_file_with_visibility): Give fatal message when no
4132 2013-09-20 Renlin Li <renlin.li@arm.com>
4134 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4135 (aarch64_expand_epilogue): Likewise.
4136 (aarch64_legitimize_reload_address): Likewise.
4138 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4141 * expr.c (expand_assignment): Remove misalignp code path.
4143 2013-09-20 Marek Polacek <polacek@redhat.com>
4146 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4147 TYPE_PRECISION. Add asserts.
4149 2013-09-20 Richard Biener <rguenther@suse.de>
4151 PR tree-optimization/58453
4152 * tree-loop-distribution.c (distribute_loop): Apply the cost
4153 model for -ftree-loop-distribute-patterns, too.
4155 2013-09-20 Richard Biener <rguenther@suse.de>
4158 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4159 make type comparison less strict.
4161 2013-09-20 Alan Modra <amodra@gmail.com>
4163 * configure: Regenerate.
4164 * aclocal.m4: Regenerate.
4166 2013-09-20 Marek Polacek <polacek@redhat.com>
4169 * doc/extend.texi: Document that attribute used is meant to be used
4170 on variables with static storage duration.
4172 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4174 PR tree-optimization/58472
4175 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4176 simd_lane_access set inv_p = false.
4177 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4178 the simduid magic VAR_DECL.
4180 2013-09-19 Jan Hubicka <jh@suse.cz>
4182 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4184 2013-09-17 Jeff Law <law@redhat.com>
4186 * tree-ssa-dom.c (record_temporary_equivalences): New function
4187 split out of dom_opt_dom_walker::after_dom_children.
4188 (dom_opt_dom_walker::thread_across_edge): Move common code
4189 in here from dom_opt_dom_walker::after_dom_children.
4190 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4192 2013-09-19 Jan Hubicka <jh@suse.cz>
4194 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4195 (TARGET_GENERIC): Use PROCESOR_GENERIC
4196 (enum processor_type): Unify generic32 and 64.
4197 * i386.md (cpu): Likewise.
4198 * x86-tune.def (use_leave): Enable for generic32.
4199 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4200 * athlon.md: Change generic64 to generic in all occurences.
4201 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4202 (ix86_target_macros_internal): Likewise.
4203 * driver-i386.c (host_detect_local_cpu): Likewise.
4204 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4206 (generic_memcpy, generic_memset, generic_cost): This one.
4207 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4208 (m_GENERIC32, m_GENERIC64): Remove.
4209 (m_GENERIC): Turn into one flag.
4210 (processor_target): Unify generic tunnings.
4211 (ix86_option_override_internal): Replace generic32/64 by generic.
4212 (ix86_issue_rate): Likewise.
4213 (ix86_adjust_cost): Likewise.
4215 2013-09-19 Jan Hubicka <jh@suse.cz>
4217 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4218 of speculative flag.
4220 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4222 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4223 GOMP_sections_start, even if !exit_reachable.
4225 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4227 * lra-constraints.c (need_for_all_save_p): Use macro
4228 HARD_REGNO_CALL_PART_CLOBBERED.
4229 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4230 set up pseudo conflict hard regs.
4232 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4235 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4238 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4240 PR rtl-optimization/58438
4241 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4242 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4243 from previous subpasses.
4245 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4247 * arm.c (arm_get_frame_offsets): Validate architecture supports
4248 LDRD/STRD before accepting the tuning preference.
4249 (arm_expand_prologue): Likewise.
4250 (arm_expand_epilogue): Likewise.
4252 2013-09-18 Richard Biener <rguenther@suse.de>
4254 PR tree-optimization/58417
4255 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4256 have chrecs with symbols defined in the loop as operands.
4257 (chrec_fold_multiply): Likewise.
4258 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4259 parameters before folding binary operations.
4260 (struct instantiate_cache_entry_hasher): Remove.
4261 (struct instantiate_cache_type): Use a pointer-map.
4262 (instantiate_cache_type::instantiate_cache_type): New function.
4263 (instantiate_cache_type::get): Likewise.
4264 (instantiate_cache_type::set): Likewise.
4265 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4266 (get_instantiated_value_entry): Likewise.
4267 (global_cache): New global.
4268 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4269 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4270 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4271 (instantiate_scev_name): Adjust.
4272 (instantiate_scev): Construct global instead of local cache.
4273 (resolve_mixers): Likewise.
4275 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4276 Paolo Carlini <paolo.carlini@oracle.com>
4279 * doc/implement-cxx.texi: Fix references to the C++ standards.
4281 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4283 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4284 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4285 simd_lane_access drs, update also DR_ALIGNED_TO.
4287 2013-09-18 Marek Polacek <polacek@redhat.com>
4290 * doc/extend.texi: Document no_sanitize_undefined attribute.
4291 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4292 no_sanitize_undefined attribute.
4294 2013-09-18 Nick Clifton <nickc@redhat.com>
4296 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4297 (ASM_DECLARE_FUNCTION_NAME): Define.
4299 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4301 * compare-elim.c (find_comparison_dom_walker): New class
4302 (find_comparisons_in_bb): Rename to
4303 find_comparison_dom_walker::before_dom_children
4304 (find_comparisons): Adjust
4305 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4307 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4308 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4309 (init_walk_dominator_tree): Remove declaration.
4310 (fini_walk_dominator_tree): Remove declaration.
4311 * fwprop.c (single_def_use_dom_walker): New class
4312 (single_def_use_enter_block): Convert to
4313 single_def_use_dom_walker::before_dom_children.
4314 (single_def_use_leave_block): Convert to
4315 single_def_use_dom_walker::after_dom_children.
4316 (build_single_def_use_links): Adjust.
4317 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4319 (find_candidates_in_block): Convert to
4320 find_candidates_dom_walker::before_dom_children.
4321 (execute_strength_reduction): Adjust.
4322 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4323 (sese_dom_walker): New class.
4324 (sese_dom_walker::sese_dom_walker): New constructor.
4325 (sese_dom_walker::~sese_dom_walker): New destructor.
4326 (build_sese_conditions_before): Convert to
4327 sese_dom_walker::before_dom_children.
4328 (build_sese_conditions_after): Convert to
4329 sese_dom_walker::after_dom_children.
4330 (build_poly_scop): Adjust
4331 * tree-into-ssa.c (rewrite_dom_walker): New class
4332 (rewrite_enter_block): Convert to
4333 rewrite_dom_walker::before_dom_children.
4334 (rewrite_leave_block): Convert to
4335 rewrite_dom_walker::after_dom_children.
4336 (rewrite_update_dom_walker): New class.
4337 (rewrite_update_enter_block): Convert to
4338 rewrite_update_dom_walker::before_dom_children.
4339 (rewrite_update_leave_block): Convert to
4340 rewrite_update_dom_walker::after_dom_children.
4341 (rewrite_blocks, rewrite_into_ssa): Adjust.
4342 (mark_def_dom_walker): New class.
4343 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4344 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4345 (mark_def_sites_blocks): Convert to
4346 mark_def_dom_walker::before_dom_children.
4347 (mark_def_site_blocks): Remove.
4348 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4349 (tree_ssa_dominator_optimize): Adjust.
4350 (dom_thread_across_edge): Convert to method
4351 dom_opt_dom_walker::thread_across_edge.
4352 (dom_opt_enter_block): Convert to member function
4353 dom_opt_dom_walker::before_dom_children.
4354 (dom_opt_leave_block): Convert to member function
4355 dom_opt_dom_walker::after_dom_children.
4356 * tree-ssa-dse.c (dse_dom_walker): New class.
4357 (dse_enter_block): Convert to member function
4358 dse_dom_walker::before_dom_children.
4359 (tree_ssa_dse): Adjust.
4360 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4361 (determine_invariantness_stmt): Convert to method
4362 invariantness_dom_walker::before_dom_children.
4363 (determine_invariantness): Remove
4364 (move_computations_dom_walker): New class.
4365 (move_computations_stmt): Convert to method
4366 move_computations_dom_walker::before_dom_children.
4367 (move_computations, tree_ssa_lim): Adjust.
4368 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4369 (nt_init_block): Make method
4370 notrappping_dom_walker::before_dom_children.
4371 (nt_fini_block): Make
4372 method nontrapping_dom_walker::after_dom_children.
4373 (get_non_trapping): Adjust.
4374 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4375 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4376 (eliminate_leave_block): Make method.
4377 eliminate_dom_walker::after_dom_children.
4379 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4380 (strlen_enter_block): Make method
4381 strlen_dom_walker::before_dom_children.
4382 (strlen_leave_block): Make
4383 method strlen_dom_walker::after_dom_children.
4384 (tree_ssa_strlen): Adjust.
4385 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4386 (tree_ssa_uncprop): Adjust.
4387 (uncprop_leave_block): Make method
4388 uncprop_dom_walker::after_dom_children.
4389 (uncprop_leave_block): Make method
4390 uncprop_dom_walker::before_dom_children.
4392 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4394 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4395 branch in same basic block. Check both src and dest of the move insn.
4397 2013-09-17 Nick Clifton <nickc@redhat.com>
4399 * config/rl78/rl78-real.md (bf): New pattern.
4401 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4402 (rl78_print_operand): Do not put a # before a %B.
4403 * config/rl78/rl78.opt: Tweak doc strings.
4405 2013-09-17 DJ Delorie <dj@redhat.com>
4407 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4408 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4409 (rl78_option_override): Likewise, if -mallregs.
4410 (is_virtual_register): Likewise.
4411 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4412 (REGNO_OK_FOR_BASE_P): Likewise.
4413 * config/rl78/rl78.opt (-mallregs): New.
4415 2013-09-17 Nick Clifton <nickc@redhat.com>
4417 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4418 For interrupt functions: save all call clobbered registers if the
4419 interrupt handler is not a leaf function.
4420 (rl78_expand_prologue): Always recompute the frame information.
4421 For interrupt functions: only select bank 0 if one of the bank 0
4422 registers is going to be psuhed.
4424 2013-09-17 DJ Delorie <dj@redhat.com>
4426 * config/rl78/constraints.md: For each W* constraint, rename to C*
4427 and create a W* constraint that checks for an optional ES: prefix
4429 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4430 (es_addr): New. Used to wrap far addresses.
4431 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4432 (rl78_es_base): New.
4433 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4434 wrapped far addresses.
4435 (rl78_print_operand_1): Unwrap far addresses before processing.
4436 (rl78_lo16): Wrap far addresses in unspecs.
4437 (rl78_es_addr): New.
4438 (rl78_es_base): New.
4439 (insn_ok_now): Check for not-yet-wrapped far addresses.
4440 (transcode_memory_rtx): Properly re-wrap far addresses.
4442 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4444 * config/sparc/t-rtems: Add leon3 multilibs.
4446 2013-09-17 Cong Hou <congh@google.com>
4448 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4449 when checking the dot production pattern. The type of rhs operand
4450 of multiply is now checked correctly.
4452 2013-09-17 Jeff Law <law@redhat.com>
4454 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4455 edge implied equivalences into successor phis.
4456 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4457 here from tree-ssa-threadedge.c.
4458 (mark_threaded_blocks): When threading through a joiner, if both
4459 successors of the joiner's clone reach the same block, verify the
4460 PHI arguments are equal. If not, cancel the jump threading request.
4461 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4462 tree-ssa-threadupdate.c
4463 (thread_across_edge): Don't check PHI argument equality when
4464 threading through joiner block here.
4466 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4468 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4469 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4470 (warn_uninit, warn_uninitialized_vars,
4471 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4472 Move to tree-ssa-uninit.c.
4473 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4474 (has_undefined_value_p): New. Pass dependant parts of
4475 ssa_undefined_value_p.
4476 (uninit_undefined_value_p): Use has_undefined_value_p.
4477 (warn_uninit, warn_uninitialized_vars,
4478 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4479 Move from tree-ssa.c.
4480 * tree-ssa.h: Adjust prototypes.
4482 2013-09-17 Jan Hubicka <jh@suse.cz>
4485 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4486 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4487 FUNCTION_NOT_OPTIMIZED.
4488 * ipa-inline-analysis.c (compute_inline_parameters): Function
4489 not optimized is not inlinable unless it is alwaysinline.
4490 (inline_analyze_function): Force calls in not optimized
4491 function not inlinable.
4493 2013-09-17 Jan Hubicka <jh@suse.cz>
4496 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4498 * ipa.c (function_and_variable_visibility): Likewise.
4499 * ipa-profile.c (ipa_profile): Likewise.
4501 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4504 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4505 attribute, and don't inline the resolver in this case.
4507 2013-09-17 Teresa Johnson <tejohnson@google.com>
4509 * coverage.c (get_coverage_counts): Add missing newline.
4511 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4513 PR tree-optimization/58088
4514 * fold-const.c (mask_with_trailing_zeros): New function.
4515 (fold_binary_loc): Make sure we don't recurse infinitely
4516 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4517 Use mask_with_trailing_zeros where appropriate.
4519 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4521 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4522 of 'prev' var to get better distance estimation.
4524 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4526 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4527 * tree-inline.c (is_parameter_of): New predicate.
4528 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4529 a parameter has been remapped.
4530 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4531 (optimize_inline_calls): Initialize transform_parameter.
4532 (copy_gimple_seq_and_replace_locals): Likewise.
4533 (tree_function_versioning): Likewise.
4534 (maybe_inline_call_in_expr): Likewise.
4536 2013-09-17 Nick Clifton <nickc@redhat.com>
4538 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4539 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4541 (is_attr_func): New function.
4542 (msp430_is_interrupt_func): New function.
4543 (is_naked_func): New function.
4544 (is_reentrant_func): New function.
4545 (is_critical_func): New function.
4546 (msp430_start_function): Add annotations for function attributes.
4547 (msp430_attr): New function.
4548 (msp430_attribute_table): New.
4549 (msp430_function_section): New function.
4550 (TARGET_ASM_FUNCTION_SECTION): Define.
4551 (msp430_builtin): New enum.
4552 (msp430_init_builtins): New function.
4553 (msp430_builtin_devl): New function.
4554 (msp430_expand_builtin): New function.
4555 (TARGET_INIT_BUILTINS): Define.
4556 (TARGET_EXPAND_BUILTINS): Define.
4557 (TARGET_BUILTIN_DECL): Define.
4558 (msp430_expand_prologue): Add support for naked, interrupt,
4559 critical and reentrant functions.
4560 (msp430_expand_epilogue): Likewise.
4561 (msp430_print_operand): Handle 'O' character.
4562 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4564 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4565 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4566 (pushm): Use a 'n' rather than an 'i' constraint.
4567 (msp_return): Add generation of the interrupt return instruction.
4568 (disable_interrupts): New pattern.
4569 (enable_interrupts): New pattern.
4570 (push_intr_state): New pattern.
4571 (pop_intr_state): New pattern.
4572 (bic_SR): New pattern.
4573 (bis_SR): New pattern.
4574 * doc/extend.texi: Document MSP430 function attributes and builtin
4577 2013-09-17 Richard Biener <rguenther@suse.de>
4579 PR tree-optimization/58432
4580 * tree-loop-distribution.c (tree_loop_distribution): Also
4581 scan PHIs for outside loop uses and seed a partition from them.
4583 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4585 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4586 (restructure_reference): Call backtrace_base_for_ref.
4588 2013-09-17 Alan Modra <amodra@gmail.com>
4591 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4594 2013-09-16 DJ Delorie <dj@redhat.com>
4596 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4597 vregs location for RL78/G10.
4598 (rl78_expand_prologue): Avoid SEL on G10.
4599 (rl78_expand_epilogue): Likewise.
4600 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4601 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4602 __RL78_G10__ when appropriate.
4603 (ASM_SPEC): Pass -mg10 along to the assembler.
4604 * config/rl78/rl78.md (sel_rb): Disable for G10.
4605 * config/rl78/rl78.opt: Add -mg10 option.
4606 * config/rl78/t-rl78: Add -mg10 multilib.
4608 2013-09-16 Xinliang David Li <davidxl@google.com>
4610 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4611 * omp-low.c (omp_max_vf): Ditto.
4612 (expand_omp_simd): Ditto.
4613 * tree-vectorizer.c (vectorize_loops): Ditto.
4614 (gate_vect_slp): Ditto.
4615 (gate_increase_alignment): Ditto.
4616 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4617 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4618 (gate_tree_vectorize): Name change.
4619 (tree_vectorize): Ditto.
4620 (pass_vectorize::gate): Call new function.
4621 (pass_vectorize::execute): Ditto.
4622 * opts.c: O3 default setting change.
4623 (finish_options): Check new flag.
4624 * doc/invoke.texi: Document new flags.
4625 * common.opt: New flags.
4627 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4629 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4630 (Condition Code): Don't reference it.
4631 * doc/tm.texi: Regenerate.
4633 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4636 * lra-constraints.c (undo_optional_reloads): Consider all optional
4637 reload even if it did not get a hard reg.
4639 2013-09-16 Teresa Johnson <tejohnson@google.com>
4641 * dumpfile.c (dump_loc): Remove newline emission.
4642 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4643 emission to dump_printf_loc calls where missing.
4644 (vect_mark_for_runtime_alias_test): Ditto.
4645 (vect_analyze_data_ref_dependence): Ditto.
4646 (vect_analyze_data_ref_dependences): Ditto.
4647 (vect_slp_analyze_data_ref_dependence): Ditto.
4648 (vect_slp_analyze_data_ref_dependences): Ditto.
4649 (vect_compute_data_ref_alignment): Ditto.
4650 (vect_update_misalignment_for_peel): Ditto.
4651 (vect_verify_datarefs_alignment): Ditto.
4652 (vector_alignment_reachable_p): Ditto.
4653 (vect_get_data_access_cost): Ditto.
4654 (vect_enhance_data_refs_alignment): Ditto.
4655 (vect_find_same_alignment_drs): Ditto.
4656 (vect_analyze_data_refs_alignment): Ditto.
4657 (vect_analyze_group_access): Ditto.
4658 (vect_analyze_data_ref_access): Ditto.
4659 (vect_analyze_data_ref_accesses): Ditto.
4660 (vect_prune_runtime_alias_test_list): Ditto.
4661 (vect_analyze_data_refs): Ditto.
4662 (vect_create_addr_base_for_vector_ref): Ditto.
4663 (vect_create_data_ref_ptr): Ditto.
4664 (vect_grouped_store_supported): Ditto.
4665 (vect_grouped_load_supported): Ditto.
4666 * value-prof.c (check_counter): Ditto.
4667 (check_ic_target): Ditto.
4668 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4669 (vect_recog_widen_mult_pattern): Ditto.
4670 (vect_recog_widen_sum_pattern): Ditto.
4671 (vect_recog_over_widening_pattern): Ditto.
4672 (vect_recog_widen_shift_pattern): Ditto.
4673 (vect_recog_rotate_pattern): Ditto.
4674 (vect_recog_vector_vector_shift_pattern): Ditto.
4675 (vect_recog_divmod_pattern): Ditto.
4676 (vect_recog_mixed_size_cond_pattern): Ditto.
4677 (vect_recog_bool_pattern): Ditto.
4678 (vect_pattern_recog_1): Ditto.
4679 (vect_pattern_recog): Ditto.
4680 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4681 (vect_is_simple_iv_evolution): Ditto.
4682 (vect_analyze_scalar_cycles_1): Ditto.
4683 (vect_get_loop_niters): Ditto.
4684 (vect_analyze_loop_1): Ditto.
4685 (vect_analyze_loop_form): Ditto.
4686 (vect_analyze_loop_operations): Ditto.
4687 (vect_analyze_loop_2): Ditto.
4688 (vect_analyze_loop): Ditto.
4689 (report_vect_op): Ditto.
4690 (vect_is_slp_reduction): Ditto.
4691 (vect_is_simple_reduction_1): Ditto.
4692 (vect_get_known_peeling_cost): Ditto.
4693 (vect_estimate_min_profitable_iters): Ditto.
4694 (vect_model_reduction_cost): Ditto.
4695 (vect_model_induction_cost): Ditto.
4696 (get_initial_def_for_induction): Ditto.
4697 (vect_create_epilog_for_reduction): Ditto.
4698 (vectorizable_reduction): Ditto.
4699 (vectorizable_induction): Ditto.
4700 (vectorizable_live_operation): Ditto.
4701 (vect_loop_kill_debug_uses): Ditto.
4702 (vect_transform_loop): Ditto.
4703 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4704 (vect_stmt_relevant_p): Ditto.
4705 (process_use): Ditto.
4706 (vect_mark_stmts_to_be_vectorized): Ditto.
4707 (vect_model_simple_cost): Ditto.
4708 (vect_model_promotion_demotion_cost): Ditto.
4709 (vect_model_store_cost): Ditto.
4710 (vect_get_store_cost): Ditto.
4711 (vect_model_load_cost): Ditto.
4712 (vect_get_load_cost): Ditto.
4713 (vect_init_vector_1): Ditto.
4714 (vect_get_vec_def_for_operand): Ditto.
4715 (vect_finish_stmt_generation): Ditto.
4716 (vectorizable_call): Ditto.
4717 (vectorizable_conversion): Ditto.
4718 (vectorizable_assignment): Ditto.
4719 (vectorizable_shift): Ditto.
4720 (vectorizable_operation): Ditto.
4721 (vectorizable_store): Ditto.
4722 (vectorizable_load): Ditto.
4723 (vectorizable_condition): Ditto.
4724 (vect_analyze_stmt): Ditto.
4725 (vect_transform_stmt): Ditto.
4726 (vect_is_simple_use): Ditto.
4727 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4728 (vect_can_advance_ivs_p): Ditto.
4729 (vect_update_ivs_after_vectorizer): Ditto.
4730 (vect_do_peeling_for_loop_bound): Ditto.
4731 (vect_gen_niters_for_prolog_loop): Ditto.
4732 (vect_update_inits_of_drs): Ditto.
4733 (vect_create_cond_for_alias_checks): Ditto.
4734 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4735 (vect_build_slp_tree_1): Ditto.
4736 (vect_supported_load_permutation_p): Ditto.
4737 (vect_analyze_slp_instance): Ditto.
4738 (vect_analyze_slp): Ditto.
4739 (vect_make_slp_decision): Ditto.
4740 (vect_detect_hybrid_slp): Ditto.
4741 (vect_bb_vectorization_profitable_p): Ditto.
4742 (vect_slp_analyze_bb_1): Ditto.
4743 (vect_update_slp_costs_according_to_vf): Ditto.
4744 (vect_get_mask_element): Ditto.
4745 (vect_transform_slp_perm_load): Ditto.
4746 (vect_schedule_slp_instance): Ditto.
4747 (vect_schedule_slp): Ditto.
4748 (vect_slp_transform_bb): Ditto.
4749 * profile.c (read_profile_edge_counts): Ditto.
4750 (compute_branch_probabilities): Ditto.
4751 * coverage.c (get_coverage_counts): Ditto.
4753 2013-09-16 Diego Novillo <dnovillo@google.com>
4755 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4757 2013-09-16 Jan Hubicka <jh@suse.cz>
4759 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4760 abstract functions; for static functions check the presence of body.
4762 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4764 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4765 * config/aarch64/aarch64-simd.md
4766 (aarch64_mla_elt<mode>): New.
4767 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4768 (aarch64_mls_elt<mode>): Likewise.
4769 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4770 (aarch64_fma4_elt<mode>): Likewise.
4771 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4772 (aarch64_fma4_elt_to_128v2df): Likewise.
4773 (aarch64_fma4_elt_to_64df): Likewise.
4774 (fnma<mode>4): Likewise.
4775 (aarch64_fnma4_elt<mode>): Likewise.
4776 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4777 (aarch64_fnma4_elt_to_128v2df): Likewise.
4778 (aarch64_fnma4_elt_to_64df): Likewise.
4779 * config/aarch64/iterators.md (VDQSF): New.
4780 * config/aarch64/arm_neon.h
4781 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4782 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4784 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4786 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4787 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4788 (aarch64_mul3_elt_to_128df): Likewise.
4789 (aarch64_mul3_elt_to_64v2df): Likewise.
4790 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4792 (VMUL_CHANGE_NLANES) Likewise.
4795 * config/aarch64/arm_neon.h
4796 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4798 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4800 * config/aarch64/arm_neon.h
4801 (vcvtx_high_f32_f64): Fix parameters.
4803 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4804 Uros Bizjak <ubizjak@gmail.com>
4806 * config/alpha.c: Include tree-ssanames.h.
4808 2013-09-16 Richard Biener <rguenther@suse.de>
4810 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4811 (dot_rdg_1): Handle control_dd.
4812 (create_edge_for_control_dependence): New function.
4813 (create_rdg_edges): Add control dependences if asked for.
4814 (build_rdg): Likewise.
4815 (generate_loops_for_partition): If there are not necessary
4816 control stmts remove all their dependencies.
4817 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4818 (distribute_loop): Pass on control dependences.
4819 (tree_loop_distribution): Compute control dependences and remove
4820 restriction on number of loop nodes.
4822 2013-09-16 Jakub Jelinek <jakub@redhat.com>
4824 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4827 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4829 * cse.c (try_const_anchors): Punt on CC modes.
4831 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4833 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4835 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
4838 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4841 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4844 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
4846 (darwin_objc1_section): Likewise.
4847 (darwin_file_end): Emit Image Info section when required.
4849 2013-09-14 Jan Hubicka <jh@suse.cz>
4851 * tree-into-ssa.c (gate_into_ssa): New.
4852 (pass_data_build_ssa): Use it.
4853 * cgraph.h (expand_thunk): Update prototype.
4854 * cgraphunit.c (analyze_function): Expand thunks early.
4855 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4856 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4857 set lowered flag; do not add new function.
4858 (assemble_thunks_and_aliases): Update.
4859 * tree-ssa.c (gate_init_datastructures): New gate.
4860 (pass_data_init_datastructures): Use it.
4862 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4865 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
4866 xmm register set described in the psABI.
4868 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4870 * dwarf2out.c (should_emit_struct_debug): Add check
4871 for type_decl variable is not NULL.
4873 2013-09-13 Jacek Caban <jacek@codeweavers.com>
4875 * config.gcc: Use new winnt-c.c target hooks
4876 * config/t-winnt: New file
4877 * config/winnt-c.c: New file
4878 * doc/tm.texi.in: Document new hook
4879 * doc/tm.texi: Regenerated
4881 2013-09-13 Jan Hubicka <jh@suse.cz>
4884 * ipa-inline.c (check_callers): New function.
4885 (check_caller_edge): Remove.
4886 (want_inline_function_to_all_callers_p): Also permit alises that are
4888 (inline_to_all_callers): Terminate the walk when devirtualization
4889 introduce new calls.
4891 2013-09-13 Jan Hubicka <jh@suse.cz>
4893 * ipa-inline-analysis.c (struct growth_data): Add node.
4894 (do_estimate_growth_1): Fix detection of recursion.
4896 2013-09-13 Jakub Jelinek <jakub@redhat.com>
4898 PR tree-optimization/58392
4899 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4900 to avoid shadowing of outer loop variable. If
4901 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4902 replace_by_duplicate_decl simduid of loops that have it set and
4903 set dest_cfun->has_simduid_loops and/or
4904 dest_cfun->has_force_vect_loops.
4905 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4906 instead of maybe_lookup_decl.
4907 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4908 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4909 call. Copy over force_vect and copy and remap simduid. Set
4910 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4911 (copy_cfg_body): Remove blocks_to_copy argument. Use
4912 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4913 caller. Don't set cfun->has_simduid_loops and/or
4914 cfun->has_force_vect_loops here.
4915 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4917 (expand_call_inline, tree_function_versioning): Adjust copy_body
4920 2013-09-13 Martin Jambor <mjambor@suse.cz>
4923 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4924 the assert if the edge was a speculative one.
4926 2013-09-13 Richard Biener <rguenther@suse.de>
4928 * tree-data-ref.h (known_dependences_p): Move here ...
4929 * tree-loop-distribution.c (known_dependences_p): ... from here.
4930 (dump_rdg_component, debug_rdg_component): Remove.
4932 (generate_loops_for_partition): Use gimple_uid instead of
4933 relying on matching stmt visit order.
4934 (rdg_build_partitions): Take starting stmt vector.
4935 (ldist_gen): Merge into ...
4936 (distribute_loop): ... this function. Do not compute starting
4938 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
4940 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4942 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
4943 Set type attribute correctly. Set predicable_short_it attribute.
4944 (cmpsi_shiftsi): Remove %? from output template.
4946 2013-09-13 Richard Biener <rguenther@suse.de>
4948 * tree-loop-distribution.c (struct rdg_component,
4949 rdg_defs_used_in_other_loops_p, free_rdg_components,
4950 rdg_build_components): Remove.
4951 (stmts_from_loop): Do not record virtual PHIs.
4952 (generate_loops_for_partition): Skip virtual PHIs.
4953 (build_rdg_partition_for_component): Rename to ...
4954 (build_rdg_partition_for_vertex): ... this and adjust.
4955 (rdg_build_partitions): Take a vector of starting vertices
4956 instead of components. Remove unnecessary leftover handling.
4957 (ldist_gen): Do not build components or record other stores.
4958 (distribute_loop): Do not distribute loops containing stmts
4961 2013-09-13 Christian Bruel <christian.bruel@st.com>
4964 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
4966 2013-09-13 Kai Tietz <ktietz@redhat.com>
4968 * config.gcc: Separate cases for mingw and cygwin targets,
4969 and add 64-bit cygwin target case.
4971 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
4972 dll-export inline-functions.
4973 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
4976 2013-09-13 Jeff Law <law@redhat.com>
4980 2013-09-06 Jeff Law <law@redhat.com>
4982 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4983 edge implied equivalences into successor phis.
4985 2013-09-12 DJ Delorie <dj@redhat.com>
4987 * config/rl78/rl78-virt.md: Change from | to \; for asm line
4990 2013-09-12 Brooks Moses <bmoses@google.com>
4993 * Makefile.in: Do not install driver binaries in $(target)/bin.
4995 2013-09-12 DJ Delorie <dj@redhat.com>
4997 * config/rl78/rl78.opt (mrelax): New.
4998 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
4999 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5001 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5003 (rl78_expand_epilogue): Likewise.
5005 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5008 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5009 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5010 pointer elimination with using remove_reg_equal_offset_note.
5012 2013-09-12 DJ Delorie <dj@redhat.com>
5014 * config/msp430/: New port.
5015 * config.gcc (msp430): Added.
5016 * doc/invoke.texi: Document MSP430 options.
5017 * doc/install.texi: Document msp430-elf
5018 * doc/md.texi: Document msp430-elf
5019 * doc/contrib.texi: Document msp430-elf
5021 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5024 2013-09-12 Martin Jambor <mjambor@suse.cz>
5027 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5028 reference descriptor is NULL.
5029 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5032 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5034 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5035 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5036 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5037 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5038 make_temp_ssa_name): move to tree-ssanames.h
5039 * tree-ssa-alias.h: Move prototype.
5040 * tree-ssa.h: Include tree-ssanames.h.
5041 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5042 * tree-ssanames.h: New. Move items from tree-flow*.h
5043 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5045 2013-09-12 Richard Biener <rguenther@suse.de>
5047 PR tree-optimization/58404
5048 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5049 propagate non-invariant addresses into dereferences wrapped
5050 in component references.
5052 2013-09-12 Richard Biener <rguenther@suse.de>
5054 PR tree-optimization/58402
5055 * passes.def: Move pass_late_warn_uninitialized later.
5057 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5059 * tree-ssa.h: New. Move content from tree-flow.h and
5061 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5062 Move prototypes belonging to tree-ssa.c.
5063 * tree-flow-inline.h (redirect_edge_var_map_def,
5064 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5066 * gimple.h: Adjust prototypes.
5067 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5069 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5070 * tree.h: Move prototype to tree-ssa.h.
5071 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5072 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5073 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5074 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5075 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5076 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5077 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5078 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5079 graphite-blocking.c, graphite-clast-to-gimple.c,
5080 graphite-dependences.c, graphite-interchange.c,
5081 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5082 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5083 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5084 ipa-reference.c, ipa-split.c, ipa-utils.c,
5085 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5086 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5087 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5088 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5089 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5090 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5091 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5092 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5093 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5094 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5095 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5096 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5097 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5098 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5099 value-prof.c, var-tracking.c,
5100 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5102 2013-09-12 Richard Biener <rguenther@suse.de>
5104 PR tree-optimization/58396
5105 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5106 (build_rdg): Take a loop-nest parameter, fix memleaks.
5107 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5109 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5111 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5114 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5116 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5117 The FMA instruction names should have a 'v' prefix.
5119 2013-09-12 Richard Biener <rguenther@suse.de>
5121 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5122 (dot_rdg): Use popen instead of system in optional code.
5123 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5124 (already_processed_vertex_p): Adjust.
5125 (has_anti_or_output_dependence, predecessor_has_mem_write,
5126 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5127 rdg_flag_uses): Remove.
5128 (rdg_flag_vertex): Simplify.
5129 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5131 (build_rdg_partition_for_component): Process the first vertex
5132 of a component only.
5133 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5135 2013-09-12 Alan Modra <amodra@gmail.com>
5137 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5139 2013-09-11 DJ Delorie <dj@redhat.com>
5140 Nick Clifton <nickc@redhat.com>
5142 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5143 (rl78_stack_based_mem): New.
5144 * config/rl78/constraints.md (Iv08): New.
5153 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5155 (movsi): Change from expand to insn-and-split.
5156 (ashrsi3): Clobber AX.
5160 * config/rl78/rl78.md (CC_REG): Fix.
5161 (addsi3): Allow memory and immediate operands.
5162 (addsi3_internal): Split into...
5163 (addsi3_internal_virt): ...new, and ...
5164 (addsi3_internal_real): ...new.
5166 (subsi3_internal_virt): New.
5167 (subsi3_internal_real): New.
5168 (mulsi3): Add memory operand.
5169 (mulsi3_rl78): Likewise.
5170 (mulsi3_g13): Likewise.
5171 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5172 (cbranchqi4_real): Add more constraint options.
5173 (cbranchhi4_real): Expand pattern.
5174 (cbranchhi4_real_signed): New.
5175 (cbranchhi4_real_inverted): New.
5176 (cbranchsi4_real_lt): New.
5177 (cbranchsi4_real_ge): New.
5178 (cbranchsi4_real_signed): New.
5179 (cbranchsi4_real): New.
5181 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5183 (lshrsi3_virt): Likewise.
5184 (ashlsi3_virt): Likewise.
5185 (cbranchqi4_virt_signed): New.
5186 (cbranchhi4_virt_signed): New.
5187 (cbranchsi4_virt): New.
5188 * config/rl78/rl78.c: Whitespace fixes throughout.
5189 (move_elim_pass): New.
5190 (pass_data_rl78_move_elim): New.
5191 (pass_rl78_move_elim): New.
5192 (make_pass_rl78_move_elim): New.
5193 (rl78_devirt_info): Run devirt earlier.
5194 (rl78_move_elim_info): New.
5195 (rl78_asm_file_start): Register it.
5196 (rl78_split_movsi): New.
5197 (rl78_as_legitimate_address): Allow virtual base registers when
5199 (rl78_addr_space_convert): Remove spurious debug stuff.
5200 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5201 (rl78_print_operand): More cases for not printing '#'.
5202 (rl78_expand_compare): Remove most of the logic.
5203 (content_memory): New.
5204 (clear_content_memory): New.
5205 (get_content_index): New.
5206 (get_content_name): New.
5207 (display_content_memory): New.
5208 (update_content): New.
5209 (record_content): New.
5210 (already_contains): New.
5211 (insn_ok_now): Re-recog insns with virtual registers.
5212 (add_postponed_content_update): New.
5213 (process_postponed_content_update): New.
5214 (gen_and_emit_move): New.
5215 (transcode_memory_rtx): Record new location content.
5216 Use gen_and_emit_move.
5217 (force_into_acc): New.
5218 (move_to_acc): Use gen_and_emit_move.
5219 (move_from_acc): Likewise.
5220 (move_acc_to_reg): Likewise.
5221 (move_to_x): Likewise.
5222 (move_to_hl): Likewise.
5223 (move_to_de): Likewise.
5224 (rl78_alloc_physical_registers_op1): Record location content.
5225 (has_constraint): New.
5226 (rl78_alloc_physical_registers_op2): Record location content.
5228 (rl78_alloc_physical_registers_ro1): Likewise.
5229 (rl78_alloc_physical_registers_cmp): Likewise.
5230 (rl78_alloc_physical_registers_umul): Likewise.
5231 (rl78_alloc_address_registers_macax): New.
5232 (rl78_alloc_physical_registers): Initialize and set location
5233 content memory as needed.
5234 (rl78_reorg): Make sure split2 is called.
5235 (rl78_rtx_costs): New.
5237 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5239 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5240 for (not (neg ...)) and (neg (not ...)) cases.
5242 2013-09-11 Richard Biener <rguenther@suse.de>
5245 * passes.def: Split critical edges before late uninit warning passes.
5246 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5248 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5250 PR tree-optimization/58385
5251 * fold-const.c (build_range_check): If both low and high are NULL,
5252 use omit_one_operand_loc to preserve exp side-effects.
5254 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5256 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5258 2013-09-11 Richard Biener <rguenther@suse.de>
5260 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5261 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5262 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5263 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5264 stmts_from_loop, known_dependences_p, build_empty_rdg,
5265 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5266 * tree-loop-distribution.c: ... here.
5267 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5268 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5269 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5270 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5271 * tree-loop-distribution.c: ... here.
5272 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5273 (struct partition_s): Add loops member.
5274 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5275 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5276 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5278 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5279 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5280 Sergey Lega <sergey.s.lega@intel.com>
5281 Anna Tikhonova <anna.tikhonova@intel.com>
5282 Ilya Tocar <ilya.tocar@intel.com>
5283 Andrey Turetskiy <andrey.turetskiy@intel.com>
5284 Ilya Verbin <ilya.verbin@intel.com>
5285 Kirill Yukhin <kirill.yukhin@intel.com>
5286 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5288 * config/i386/constraints.md (k): New.
5290 * config/i386/i386.c (const regclass_map): Add new mask registers.
5291 (dbx_register_map): Ditto.
5292 (dbx64_register_map): Ditto.
5293 (svr4_dbx_register_map): Ditto.
5294 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5296 (ix86_preferred_reload_class): Disable constants for mask registers.
5297 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5298 (ix86_hard_regno_mode_ok): Support new mask registers.
5299 (x86_order_regs_for_local_alloc): Ditto.
5300 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5301 (FIXED_REGISTERS): Add new mask registers.
5302 (CALL_USED_REGISTERS): Ditto.
5303 (REG_ALLOC_ORDER): Ditto.
5304 (VALID_MASK_REG_MODE): New.
5305 (FIRST_MASK_REG): Ditto.
5306 (LAST_MASK_REG): Ditto.
5307 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5308 (MAYBE_MASK_CLASS_P): New.
5309 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5310 (REG_CLASS_CONTENTS): Ditto.
5311 (MASK_REGNO_P): New.
5312 (ANY_MASK_REG_P): Ditto.
5313 (HI_REGISTER_NAMES): Add new mask registers.
5314 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5315 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5317 (attribute "type"): Add mskmov, msklog.
5318 (attribute "length_immediate"): Support them.
5319 (attribute "memory"): Ditto.
5320 (attribute "prefix_0f"): Ditto.
5321 (*movhi_internal): Support new mask registers.
5322 (*movqi_internal): Ditto.
5323 (define_split): Split out clobber pattern is a logic
5324 insn on mask registers.
5325 (*k<logic><mode>): New.
5326 (*andhi_1): Extend to support mask regs.
5327 (*andqi_1): Extend to support mask regs.
5329 (define_split): Split and-not to and and not if operands
5331 (*<code><mode>_1): Separate HI mode to new pattern...
5332 (*<code>hi_1): This.
5333 (*<code>qi_1): Extend to support mask regs.
5335 (kortestzhi): Ditto.
5336 (kortestchi): Ditto.
5338 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5339 (*one_cmplhi2_1): ...Here, now with mask registers support.
5340 (*one_cmplqi2_1): Support new mask registers.
5341 (HI/QImode arithmetics splitter): Don't split if mask registers
5343 (HI/QImode not splitter): Ditto.
5344 * config/i386/predicated.md (mask_reg_operand): New.
5345 (general_reg_operand): Ditto.
5347 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5349 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5350 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5352 2013-09-10 Jeff Law <law@redhat.com>
5354 PR tree-optimization/58380
5355 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5356 of threading through a buried loop header.
5358 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5359 return value for single successor case.
5361 2013-09-10 Jan Hubicka <jh@suse.cz>
5363 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5365 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5368 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5369 support conditional execution.
5370 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5372 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5374 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5376 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5377 Check destination too to check move insn.
5378 (undo_optional_reloads): Add check that the original peudo did not
5379 changed its allocation and the optional reload was inherited on last
5380 inheritance pass. Break loop after deciding to keep optional reload.
5381 (lra_undo_inheritance): Add check that inherited pseudo still in
5384 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5386 * config/aarch64/aarch64.md (generic_sched): New.
5387 * config/aarch64/aarch64-generic.md (load): Make conditional
5388 on generic_sched attribute.
5389 (nonload): Likewise.
5391 2013-09-10 Jan Hubicka <jh@suse.cz>
5393 * lto-cgraph.c: Include ipa-utils.h.
5394 (compute_ltrans_boundary): Also add possible targets into the boundary.
5396 2013-09-10 Jan Hubicka <jh@suse.cz>
5398 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5399 VAR_DECL of vtable rather than full expression.
5401 2013-09-10 Jan Hubicka <jh@suse.cz>
5402 Paolo Carlini <paolo.carlini@oracle.com>
5404 * cgraphunit.c (analyze_functions): Save input_location, set it
5405 to UNKNOWN_LOCATION and restore it at the end.
5407 2013-09-10 Martin Jambor <mjambor@suse.cz>
5409 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5410 represented by a thunk.
5412 2013-09-10 Jeff Law <law@redhat.com>
5414 PR tree-optimization/58343
5415 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5416 jump threading requests through loop headers buried in the middle
5417 of a jump threading path.
5419 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5420 in return value/type.
5422 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5424 PR rtl-optimization/58365
5425 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5426 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5429 2013-09-10 Richard Biener <rguenther@suse.de>
5431 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5432 * tree-data-ref.c (create_rdg_vertices): Collect all data
5433 references, signal failure to the caller, use data-ref API.
5434 (build_rdg): Compute data references only once. Maintain lifetime
5435 of data references and data dependences from within RDG.
5436 (free_rdg): Free dependence relations.
5437 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5438 inventing extra dependences.
5439 (distribute_loop): Update for RDG API changes.
5441 2013-09-10 Kai Tietz <ktietz@redhat.com>
5443 * doc/invoke.texi (fms-extensions): Document changed
5444 behavior for ms-abi targets.
5445 * config/i386/i386.c (ix86_option_override_internal):
5446 Set default value of option -fms-extension for ms-abi targets.
5448 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5450 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5452 2013-09-10 Alan Modra <amodra@gmail.com>
5455 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5457 2013-09-10 Alan Modra <amodra@gmail.com>
5459 * config/rs6000/predicates.md (add_cint_operand): New.
5460 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5461 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5462 using add_cint_operand.
5463 (largetoc_high_plus_aix): Likewise.
5465 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5467 PR tree-optimization/58364
5468 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5469 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5470 the current range can't be an unconditional true or false.
5472 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5474 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5476 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5478 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5480 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5483 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5485 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5487 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5488 NO_REGS for immediate that can't be moved directly into FP_REGS.
5490 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5492 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5493 comparison with negated operand.
5494 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5497 2013-09-09 Richard Biener <rguenther@suse.de>
5500 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5501 of a subloop record all its block as affecting loop-closed SSA form.
5503 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5505 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5506 of an rtx/bitpos pair.
5507 (store_fixed_bit_field): Update accordingly.
5509 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5511 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5513 * builtins.c (expand_errno_check): Likewise.
5514 * dwarf2cfi.c (init_return_column_size): Likewise.
5515 * except.c (sjlj_mark_call_sites): Likewise.
5516 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5517 * lra-constraints.c (emit_inc): Likewise.
5518 * ree.c (combine_set_extension): Likewise.
5519 * regmove.c (fixup_match_2): Likewise.
5520 * reload1.c (inc_for_reload): Likewise.
5522 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5524 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5525 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5526 Use gen_int_mode with the mode of the associated simplify_* call.
5527 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5528 * expmed.c (expand_shift_1): Likewise.
5529 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5530 * loop-iv.c (iv_number_of_iterations): Likewise.
5531 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5532 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5534 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5536 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5537 of the associated expand_* call.
5538 (asan_emit_stack_protection): Likewise.
5539 * builtins.c (round_trampoline_addr): Likewise.
5540 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5541 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5542 (emit_store_flag): Likewise.
5543 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5544 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5546 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5547 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5548 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5549 * modulo-sched.c (generate_prolog_epilog): Likewise.
5550 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5551 (expand_ctz, expand_ffs, expand_unop): Likewise.
5553 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5555 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5556 of the associated gen_rtx_* call.
5557 * caller-save.c (init_caller_save): Likewise.
5558 * combine.c (find_split_point, make_extraction): Likewise.
5559 (make_compound_operation): Likewise.
5560 * dwarf2out.c (mem_loc_descriptor): Likewise.
5561 * explow.c (plus_constant, probe_stack_range): Likewise.
5562 * expmed.c (expand_mult_const): Likewise.
5563 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5564 * reload1.c (init_reload): Likewise.
5565 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5566 * var-tracking.c (adjust_mems): Likewise.
5567 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5568 rather than gen_rtx_fmt_ee.
5570 2013-09-09 Jan Hubicka <jh@suse.cz>
5573 * value-prof.c (gimple_ic): Copy also abnormal edges.
5575 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5577 * asan.c (asan_shadow_cst): Use gen_int_mode.
5579 2013-09-08 Jan Hubicka <jh@suse.cz>
5581 * ipa-profile.c: Add toplevel comment.
5582 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5583 (contains_hot_call_p): New function.
5584 (ipa_propagate_frequency): Set frequencies based on counts when
5586 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5587 profile; do not tamper with profile after inlining if it is read.
5589 2013-09-08 Jan Hubicka <jh@suse.cz>
5591 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5594 2013-09-08 Jan Hubicka <jh@suse.cz>
5596 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5599 2013-09-08 Jeff Law <law@redhat.com>
5602 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5605 2013-09-08 Andi Kleen <ak@linux.intel.com>
5607 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5609 2013-09-08 Jan Hubicka <jh@suse.cz>
5611 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5613 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5615 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5616 for non-debug insns.
5617 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5618 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5619 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5622 2013-09-08 Jan Hubicka <jh@suse.cz>
5624 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5625 targets and devirtualize to BUILT_IN_UNREACHABLE.
5626 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5627 * ipa.c (walk_polymorphic_call_targets): New function.
5628 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5629 functions; use the new timevar.
5630 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5631 was removed from the program.
5632 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5633 not consider it in the walk when its vtable is dead.
5634 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5636 (devirt_variable_node_removal_hook): New function.
5637 (possible_polymorphic_call_targets): Also register
5638 devirt_variable_node_removal_hook.
5639 (ipa_devirt): Do not do non-speculative devirtualization.
5640 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5642 2013-09-08 Jan Hubicka <jh@suse.cz>
5644 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5645 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5646 varpool_remove_variable_insertion_hook): Declare.
5647 * varpool.c (varpool_node_hook_list): New structure.
5648 (first_varpool_node_removal_hook,
5649 first_varpool_variable_insertion_hook): New variables.
5650 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5651 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5652 varpool_remove_variable_insertion_hook,
5653 varpool_call_variable_insertion_hooks): New functions.
5654 (varpool_remove_node): Use it.
5656 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5659 * diagnostic.c (diagnostic_build_prefix): When s.file is
5660 "<built-in>" don't output line and column numbers.
5662 2013-09-06 Jan Hubicka <jh@suse.cz>
5664 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5665 * lto-streamer-out.c: Stream thunks, too.
5666 * lto-streamer-in.c (input_function): Pop cfun here
5667 (lto_read_body): Instead of here.
5669 2013-09-06 Caroline Tice <cmtice@google.com>
5671 * doc/install.texi: Add documentation for the --enable-vtable-verify
5672 and the --disable-libvtv configure options.
5674 2013-09-06 Jeff Law <law@redhat.com>
5676 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5677 edge implied equivalences into successor phis.
5679 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5681 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5683 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5685 * resource.c (mark_target_live_regs): Compute resources taking
5686 into account if a call is predicated or not.
5688 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5690 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5691 the compiler in the function names.
5693 2013-09-06 Jan Hubicka <jh@suse.cz>
5696 * ipa-inline.c (has_caller_p): New function.
5697 (want_inline_function_to_all_callers_p): Use it.
5698 (sum_callers, inline_to_all_callers): Break out from ...
5699 (ipa_inline): ... here.
5701 2013-09-06 Jan Hubicka <jh@suse.cz>
5703 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5704 only when AVX is enabled.
5706 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5708 * config/aarch64/aarch64.md
5709 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5710 is fpsimd_<load/store>2.
5711 (load_pair<mode>): Likewise.
5712 (store_pair<mode>): Likewise.
5714 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5716 * config/arm/types.md (type): Add "mrs" type.
5717 * config/aarch64/aarch64.md
5718 (aarch64_load_tp_hard): Make type "mrs".
5720 (load_tp_hard): Make type "mrs".
5721 * config/arm/cortex-a15.md: Update with new attributes.
5722 * config/arm/cortex-a5.md: Update with new attributes.
5723 * config/arm/cortex-a53.md: Update with new attributes.
5724 * config/arm/cortex-a7.md: Update with new attributes.
5725 * config/arm/cortex-a8.md: Update with new attributes.
5726 * config/arm/cortex-a9.md: Update with new attributes.
5727 * config/arm/cortex-m4.md: Update with new attributes.
5728 * config/arm/cortex-r4.md: Update with new attributes.
5729 * config/arm/fa526.md: Update with new attributes.
5730 * config/arm/fa606te.md: Update with new attributes.
5731 * config/arm/fa626te.md: Update with new attributes.
5732 * config/arm/fa726te.md: Update with new attributes.
5734 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5736 * config/aarch64/aarch64.md
5737 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5738 (*movtf_aarch64): Likewise.
5740 (thumb1_movdi_insn): Use "multiple" for type where more than one
5741 instruction is used for a move.
5742 (*arm32_movhf): Likewise.
5743 (*thumb_movdf_insn): Likewise.
5745 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5747 * config/arm/types.md (type): Rename fcpys to fmov.
5749 (*arm_movsi_vfp): Rename type fcpys as fmov.
5750 (*thumb2_movsi_vfp): Likewise
5751 (*movhf_vfp_neon): Likewise
5752 (*movhf_vfp): Likewise
5753 (*movsf_vfp): Likewise
5754 (*thumb2_movsf_vfp): Likewise
5755 (*movsfcc_vfp): Likewise
5756 (*thumb2_movsfcc_vfp): Likewise
5757 * config/aarch64/aarch64-simd.md
5758 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5759 * config/aarch64/aarch64.md
5760 (*movsi_aarch64): Replace type mov_reg with fmovs.
5761 (*movdi_aarch64): Likewise
5762 (*movsf_aarch64): Likewise
5763 (*movdf_aarch64): Likewise
5765 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5766 * config/arm/iwmmxt.md
5767 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5768 * config/arm/arm1020e.md: Update with new attributes.
5769 * config/arm/cortex-a15-neon.md: Update with new attributes.
5770 * config/arm/cortex-a5.md: Update with new attributes.
5771 * config/arm/cortex-a53.md: Update with new attributes.
5772 * config/arm/cortex-a7.md: Update with new attributes.
5773 * config/arm/cortex-a8-neon.md: Update with new attributes.
5774 * config/arm/cortex-a9.md: Update with new attributes.
5775 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5776 * config/arm/cortex-r4f.md: Update with new attributes.
5777 * config/arm/marvell-pj4.md: Update with new attributes.
5778 * config/arm/vfp11.md: Update with new attributes.
5780 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5782 * config/aarch64/aarch64.md
5783 (*madd<mode>): Fix type attribute.
5784 (*maddsi_uxtw): Likewise.
5785 (*msub<mode>): Likewise.
5786 (*msubsi_uxtw): Likewise.
5787 (<su_optab>maddsidi4): Likewise.
5788 (<su_optab>msubsidi4): Likewise.
5790 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5792 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5793 * config/arm/arm.md (core_cycles): Remove fdiv.
5794 * config/arm/vfp.md:
5795 (*sqrtsf2_vfp): Update for attribute changes.
5796 (*sqrtdf2_vfp): Likewise.
5797 * config/aarch64/aarch64.md:
5798 (sqrt<mode>2): Update for attribute changes.
5799 * config/arm/arm1020e.md: Update with new attributes.
5800 * config/arm/cortex-a15-neon.md: Update with new attributes.
5801 * config/arm/cortex-a5.md: Update with new attributes.
5802 * config/arm/cortex-a53.md: Update with new attributes.
5803 * config/arm/cortex-a7.md: Update with new attributes.
5804 * config/arm/cortex-a8-neon.md: Update with new attributes.
5805 * config/arm/cortex-a9.md: Update with new attributes.
5806 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5807 * config/arm/cortex-r4f.md: Update with new attributes.
5808 * config/arm/marvell-pj4.md: Update with new attributes.
5809 * config/arm/vfp11.md: Update with new attributes.
5811 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5813 * config/arm/types.md
5814 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5815 * config/aarch64/aarch64.md
5816 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5818 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5819 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5820 (float<GPI:mode><GPF:mode>2): Likewise.
5822 (*truncsisf2_vfp): Update with new attributes.
5823 (*truncsidf2_vfp): Likewise.
5824 (fixuns_truncsfsi2): Likewise.
5825 (fixuns_truncdfsi2): Likewise.
5826 (*floatsisf2_vfp): Likewise.
5827 (*floatsidf2_vfp): Likewise.
5828 (floatunssisf2): Likewise.
5829 (floatunssidf2): Likewise.
5830 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5831 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5832 * config/arm/arm1020e.md: Update with new attributes.
5833 * config/arm/cortex-a15-neon.md: Update with new attributes.
5834 * config/arm/cortex-a5.md: Update with new attributes.
5835 * config/arm/cortex-a53.md: Update with new attributes.
5836 * config/arm/cortex-a7.md: Update with new attributes.
5837 * config/arm/cortex-a8-neon.md: Update with new attributes.
5838 * config/arm/cortex-a9.md: Update with new attributes.
5839 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5840 * config/arm/cortex-r4f.md: Update with new attributes.
5841 * config/arm/marvell-pj4.md: Update with new attributes.
5842 * config/arm/vfp11.md: Update with new attributes.
5844 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5846 * config/aarch64/arm_neon.h
5847 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5848 (vqtbx<1,2,3,4><q>_s8): Likewise.
5850 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5852 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5853 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
5854 (add<mode>3): Add type attribute.
5855 (add<mode>3): Likewise.
5856 (usadd<mode>3): Likewise.
5857 (ssadd<mode>3): Likewise.
5858 (sub<mode>3): Likewise.
5859 (sub<mode>3): Likewise.
5860 (ussub<mode>3): Likewise.
5861 (sssub<mode>3): Likewise.
5862 (ssmulsa3): Likewise.
5863 (usmulusa3): Likewise.
5864 (arm_usatsihi): Likewise.
5866 (*movdi_vfp): Add types for all instructions.
5867 (*movdi_vfp_cortexa8): Likewise.
5868 (*movhf_vfp_neon): Likewise.
5869 (*movhf_vfp): Likewise.
5870 (*movdf_vfp): Likewise.
5871 (*thumb2_movdf_vfp): Likewise.
5872 (*thumb2_movdfcc_vfp): Likewise.
5873 * config/arm/arm.md: Add type attribute to all insn patterns.
5874 (*thumb1_adddi3): Add type attribute.
5875 (*arm_adddi3): Likewise.
5876 (*adddi_sesidi_di): Likewise.
5877 (*adddi_zesidi_di): Likewise.
5878 (*thumb1_addsi3): Likewise.
5879 (addsi3_compare0): Likewise.
5880 (*addsi3_compare0_scratch): Likewise.
5881 (*compare_negsi_si): Likewise.
5882 (cmpsi2_addneg): Likewise.
5883 (*addsi3_carryin_<optab>): Likewise.
5884 (*addsi3_carryin_alt2_<optab>): Likewise.
5885 (*addsi3_carryin_clobercc_<optab>): Likewise.
5886 (*subsi3_carryin): Likewise.
5887 (*subsi3_carryin_const): Likewise.
5888 (*subsi3_carryin_compare): Likewise.
5889 (*subsi3_carryin_compare_const): Likewise.
5890 (*arm_subdi3): Likewise.
5891 (*thumb_subdi3): Likewise.
5892 (*subdi_di_zesidi): Likewise.
5893 (*subdi_di_sesidi): Likewise.
5894 (*subdi_zesidi_di): Likewise.
5895 (*subdi_sesidi_di): Likewise.
5896 (*subdi_zesidi_ze): Likewise.
5897 (thumb1_subsi3_insn): Likewise.
5898 (*arm_subsi3_insn): Likewise.
5899 (*anddi3_insn): Likewise.
5900 (*anddi_zesidi_di): Likewise.
5901 (*anddi_sesdi_di): Likewise.
5902 (*ne_zeroextracts): Likewise.
5903 (*ne_zeroextracts): Likewise.
5904 (*ite_ne_zeroextr): Likewise.
5905 (*ite_ne_zeroextr): Likewise.
5906 (*anddi_notdi_di): Likewise.
5907 (*anddi_notzesidi): Likewise.
5908 (*anddi_notsesidi): Likewise.
5909 (andsi_notsi_si): Likewise.
5910 (thumb1_bicsi3): Likewise.
5911 (*iordi3_insn): Likewise.
5912 (*iordi_zesidi_di): Likewise.
5913 (*iordi_sesidi_di): Likewise.
5914 (*thumb1_iorsi3_insn): Likewise.
5915 (*xordi3_insn): Likewise.
5916 (*xordi_zesidi_di): Likewise.
5917 (*xordi_sesidi_di): Likewise.
5918 (*arm_xorsi3): Likewise.
5919 (*andsi_iorsi3_no): Likewise.
5920 (*smax_0): Likewise.
5921 (*smax_m1): Likewise.
5922 (*arm_smax_insn): Likewise.
5923 (*smin_0): Likewise.
5924 (*arm_smin_insn): Likewise.
5925 (*arm_umaxsi3): Likewise.
5926 (*arm_uminsi3): Likewise.
5927 (*minmax_arithsi): Likewise.
5928 (*minmax_arithsi_): Likewise.
5929 (*satsi_<SAT:code>): Likewise.
5930 (arm_ashldi3_1bit): Likewise.
5931 (arm_ashrdi3_1bit): Likewise.
5932 (arm_lshrdi3_1bit): Likewise.
5933 (*arm_negdi2): Likewise.
5934 (*thumb1_negdi2): Likewise.
5935 (*arm_negsi2): Likewise.
5936 (*thumb1_negsi2): Likewise.
5937 (*negdi_extendsid): Likewise.
5938 (*negdi_zero_extend): Likewise.
5939 (*arm_abssi2): Likewise.
5940 (*thumb1_abssi2): Likewise.
5941 (*arm_neg_abssi2): Likewise.
5942 (*thumb1_neg_abss): Likewise.
5943 (one_cmpldi2): Likewise.
5944 (extend<mode>di2): Likewise.
5945 (*compareqi_eq0): Likewise.
5946 (*arm_extendhisi2addsi): Likewise.
5947 (*arm_movdi): Likewise.
5948 (*thumb1_movdi_insn): Likewise.
5949 (*arm_movt): Likewise.
5950 (*thumb1_movsi_insn): Likewise.
5951 (pic_add_dot_plus_four): Likewise.
5952 (pic_add_dot_plus_eight): Likewise.
5953 (tls_load_dot_plus_eight): Likewise.
5954 (*thumb1_movhi_insn): Likewise.
5955 (*thumb1_movsf_insn): Likewise.
5956 (*movdf_soft_insn): Likewise.
5957 (*thumb_movdf_insn): Likewise.
5958 (cbranchsi4_insn): Likewise.
5959 (cbranchsi4_scratch): Likewise.
5960 (*negated_cbranchsi4): Likewise.
5961 (*tbit_cbranch): Likewise.
5962 (*tlobits_cbranch): Likewise.
5963 (*tstsi3_cbranch): Likewise.
5964 (*cbranchne_decr1): Likewise.
5965 (*addsi3_cbranch): Likewise.
5966 (*addsi3_cbranch_scratch): Likewise.
5967 (*arm_cmpdi_insn): Likewise.
5968 (*arm_cmpdi_unsig): Likewise.
5969 (*arm_cmpdi_zero): Likewise.
5970 (*thumb_cmpdi_zero): Likewise.
5971 (*deleted_compare): Likewise.
5972 (*mov_scc): Likewise.
5973 (*mov_negscc): Likewise.
5974 (*mov_notscc): Likewise.
5975 (*cstoresi_eq0_thumb1_insn): Likewise.
5976 (cstoresi_nltu_thumb1): Likewise.
5977 (cstoresi_ltu_thu): Likewise.
5978 (thumb1_addsi3_addgeu): Likewise.
5979 (*arm_jump): Likewise.
5980 (*thumb_jump): Likewise.
5981 (*check_arch2): Likewise.
5982 (arm_casesi_internal): Likewise.
5983 (thumb1_casesi_dispatch): Likewise.
5984 (*arm_indirect_jump): Likewise.
5985 (*thumb1_indirect_jump): Likewise.
5987 (*and_scc): Likewise.
5988 (*ior_scc): Likewise.
5989 (*compare_scc): Likewise.
5990 (*cond_move): Likewise.
5991 (*cond_arith): Likewise.
5992 (*cond_sub): Likewise.
5993 (*cmp_ite0): Likewise.
5994 (*cmp_ite1): Likewise.
5995 (*cmp_and): Likewise.
5996 (*cmp_ior): Likewise.
5997 (*ior_scc_scc): Likewise.
5998 (*ior_scc_scc_cmp): Likewise.
5999 (*and_scc_scc): Likewise.
6000 (*and_scc_scc_cmp): Likewise.
6001 (*and_scc_scc_nod): Likewise.
6002 (*negscc): Likewise.
6003 (movcond_addsi): Likewise.
6004 (movcond): Likewise.
6005 (*ifcompare_plus_move): Likewise.
6006 (*if_plus_move): Likewise.
6007 (*ifcompare_move_plus): Likewise.
6008 (*if_move_plus): Likewise.
6009 (*ifcompare_arith_arith): Likewise.
6010 (*if_arith_arith): Likewise.
6011 (*ifcompare_arith_move): Likewise.
6012 (*if_arith_move): Likewise.
6013 (*ifcompare_move_arith): Likewise.
6014 (*if_move_arith): Likewise.
6015 (*ifcompare_move_not): Likewise.
6016 (*if_move_not): Likewise.
6017 (*ifcompare_not_move): Likewise.
6018 (*if_not_move): Likewise.
6019 (*ifcompare_shift_move): Likewise.
6020 (*if_shift_move): Likewise.
6021 (*ifcompare_move_shift): Likewise.
6022 (*if_move_shift): Likewise.
6023 (*ifcompare_shift_shift): Likewise.
6024 (*ifcompare_not_arith): Likewise.
6025 (*ifcompare_arith_not): Likewise.
6026 (*if_arith_not): Likewise.
6027 (*ifcompare_neg_move): Likewise.
6028 (*if_neg_move): Likewise.
6029 (*ifcompare_move_neg): Likewise.
6030 (*if_move_neg): Likewise.
6031 (prologue_thumb1_interwork): Likewise.
6032 (*cond_move_not): Likewise.
6033 (*sign_extract_onebit): Likewise.
6034 (*not_signextract_onebit): Likewise.
6035 (stack_tie): Likewise.
6036 (align_4): Likewise.
6037 (align_8): Likewise.
6038 (consttable_end): Likewise.
6039 (consttable_1): Likewise.
6040 (consttable_2): Likewise.
6041 (consttable_4): Likewise.
6042 (consttable_8): Likewise.
6043 (consttable_16): Likewise.
6044 (*thumb1_tablejump): Likewise.
6045 (prefetch): Likewise.
6046 (force_register_use): Likewise.
6047 (thumb_eh_return): Likewise.
6048 (load_tp_hard): Likewise.
6049 (load_tp_soft): Likewise.
6050 (tlscall): Likewise.
6051 (*arm_movtas_ze): Likewise.
6052 (*arm_rev): Likewise.
6053 (*arm_revsh): Likewise.
6054 (*arm_rev16): Likewise.
6055 * config/arm/thumb2.md
6056 (*thumb2_smaxsi3): Likewise.
6057 (*thumb2_sminsi3): Likewise.
6058 (*thumb32_umaxsi3): Likewise.
6059 (*thumb2_uminsi3): Likewise.
6060 (*thumb2_negdi2): Likewise.
6061 (*thumb2_abssi2): Likewise.
6062 (*thumb2_neg_abss): Likewise.
6063 (*thumb2_movsi_insn): Likewise.
6064 (tls_load_dot_plus_four): Likewise.
6065 (*thumb2_movhi_insn): Likewise.
6066 (*thumb2_mov_scc): Likewise.
6067 (*thumb2_mov_negs): Likewise.
6068 (*thumb2_mov_negs): Likewise.
6069 (*thumb2_mov_nots): Likewise.
6070 (*thumb2_mov_nots): Likewise.
6071 (*thumb2_movsicc_): Likewise.
6072 (*thumb2_movsfcc_soft_insn): Likewise.
6073 (*thumb2_indirect_jump): Likewise.
6074 (*thumb2_and_scc): Likewise.
6075 (*thumb2_ior_scc): Likewise.
6076 (*thumb2_ior_scc_strict_it): Likewise.
6077 (*thumb2_cond_move): Likewise.
6078 (*thumb2_cond_arith): Likewise.
6079 (*thumb2_cond_ari): Likewise.
6080 (*thumb2_cond_sub): Likewise.
6081 (*thumb2_negscc): Likewise.
6082 (*thumb2_movcond): Likewise.
6083 (thumb2_casesi_internal): Likewise.
6084 (thumb2_casesi_internal_pic): Likewise.
6085 (*thumb2_alusi3_short): Likewise.
6086 (*thumb2_mov<mode>_shortim): Likewise.
6087 (*thumb2_addsi_short): Likewise.
6088 (*thumb2_subsi_short): Likewise.
6089 (thumb2_addsi3_compare0): Likewise.
6090 (*thumb2_cbz): Likewise.
6091 (*thumb2_cbnz): Likewise.
6092 (*thumb2_one_cmplsi2_short): Likewise.
6093 (*thumb2_negsi2_short): Likewise.
6094 (*orsi_notsi_si): Likewise.
6095 * config/arm/arm1020e.md: Update with new attributes.
6096 * config/arm/arm1026ejs.md: Update with new attributes.
6097 * config/arm/arm1136jfs.md: Update with new attributes.
6098 * config/arm/arm926ejs.md: Update with new attributes.
6099 * config/arm/cortex-a15.md: Update with new attributes.
6100 * config/arm/cortex-a5.md: Update with new attributes.
6101 * config/arm/cortex-a53.md: Update with new attributes.
6102 * config/arm/cortex-a7.md: Update with new attributes.
6103 * config/arm/cortex-a8.md: Update with new attributes.
6104 * config/arm/cortex-a9.md: Update with new attributes.
6105 * config/arm/cortex-m4.md: Update with new attributes.
6106 * config/arm/cortex-r4.md: Update with new attributes.
6107 * config/arm/fa526.md: Update with new attributes.
6108 * config/arm/fa606te.md: Update with new attributes.
6109 * config/arm/fa626te.md: Update with new attributes.
6110 * config/arm/fa726te.md: Update with new attributes.
6112 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6114 * config/aarch64/aarch64-simd.md
6115 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6116 <vwx> iterator to ensure correct register choice.
6117 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6118 (aarch64_sqdmull_n<mode>): Likewise.
6119 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6120 * config/aarch64/arm_neon.h
6121 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6122 (vml<as><q>_n_<su>16): Likewise.
6123 (vml<as>l_high_lane<q>_<su>16): Likewise.
6124 (vml<as>l_high_n_<su>16): Likewise.
6125 (vml<as>l_lane<q>_<su>16): Likewise.
6126 (vml<as>l_n_<su>16): Likewise.
6127 (vmul<q>_lane<q>_<su>16): Likewise.
6128 (vmul<q>_n_<su>16): Likewise.
6129 (vmull_lane<q>_<su>16): Likewise.
6130 (vmull_n_<su>16): Likewise.
6131 (vmull_high_lane<q>_<su>16): Likewise.
6132 (vmull_high_n_<su>16): Likewise.
6133 (vqrdmulh<q>_n_s16): Likewise.
6135 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6137 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6138 have the correct lane parameter.
6140 2013-09-06 Richard Biener <rguenther@suse.de>
6142 * cfganal.c (control_dependences::~control_dependences):
6143 Properly free all of the vector.
6145 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6148 * config/i386/i386.c (ix86_conditional_register_usage):
6149 Proper initialize extended SSE registers.
6151 2013-09-06 Jan Hubicka <jh@suse.cz>
6153 PR tree-optimization/58311
6154 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6156 2013-09-06 Jan Hubicka <jh@suse.cz>
6158 * Makefile.in (tree-sra.o): Update dependencies.
6159 * tree-sra.c: Include ipa-utils.h
6160 (scan_function): Use recursive_call_p.
6161 (has_caller_p): New function.
6162 (cgraph_for_node_and_aliases): Count also callers of aliases.
6164 2013-09-06 Jan Hubicka <jh@suse.cz>
6167 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6168 * tree-tailcall.c: Include ipa-utils.h.
6169 (find_tail_calls): Use it.
6170 * ipa-pure-const.c (check_call): Likewise.
6171 * ipa-utils.c (recursive_call_p): New function.
6172 * ipa-utils.h (recursive_call_p): Dclare.
6173 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6174 (symtab_semantically_equivalent_p): New function.
6175 * Makefile.in (tree-tailcall.o): Update dependencies.
6177 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6179 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6182 2013-09-06 Richard Biener <rguenther@suse.de>
6184 * lto-streamer.h (lto_global_var_decls): Remove.
6185 * Makefile.in (OBJS): Remove lto-symtab.o.
6186 (lto-symtab.o): Remove.
6187 (GTFILES): Remove lto-symtab.c
6188 * lto-symtab.c: Move to lto/
6190 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6192 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6193 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6194 (UNSPEC_FPINT_RINT): New constant definitions.
6195 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6196 definition with 2 attributes.
6197 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6198 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6201 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6203 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6204 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6205 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6206 attribute to "z196_alone".
6207 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6210 2013-09-06 Richard Biener <rguenther@suse.de>
6212 * basic-block.h (class control_dependences): New.
6213 * tree-ssa-dce.c (control_dependence_map): Remove.
6215 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6216 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6217 find_pdom, find_control_dependence, find_all_control_dependences):
6219 (mark_control_dependent_edges_necessary,
6220 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6221 tree_dce_done, perform_tree_ssa_dce): Adjust.
6222 * cfganal.c (set_control_dependence_map_bit,
6223 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6224 find_all_control_dependences): Move from tree-ssa-dce.c and
6225 implement as methods of control_dependences class.
6226 (control_dependences::control_dependences): New.
6227 (control_dependences::~control_dependences): Likewise.
6228 (control_dependences::get_edges_dependent_on): Likewise.
6229 (control_dependences::get_edge): Likewise.
6231 2013-09-04 Jan Hubicka <jh@suse.cz>
6233 * tree.c (types_same_for_odr): Drop overactive check.
6234 * ipa-devirt.c (hash_type_name): Likewise.
6236 2013-09-04 Jan Hubicka <jh@suse.cz>
6238 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6239 (analyze_functions): ... here.
6241 2013-09-04 Jan Hubicka <jh@suse.cz>
6244 * cgraphunit.c (analyze_functions): Clear AUX fields
6245 after processing; initialize assembler name has.
6247 2013-09-05 Jeff Law <law@redhat.com>
6249 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6250 from thread_around_empty_block. Record threading path into PATH.
6251 Recurse if threading through the initial block is successful.
6252 (thread_across_edge): Corresponding changes to slightly simplify.
6254 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6256 * config/aarch64/aarch64.md
6257 (type): Remove frecpe, frecps, frecpx.
6258 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6259 fix to be a TARGET_SIMD instruction.
6260 (aarch64_frecps): Remove.
6261 * config/aarch64/aarch64-simd.md
6262 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6263 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6265 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6266 Sofiane Naci <sofiane.naci@arm.com>
6268 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6269 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6270 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6271 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6272 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6273 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6274 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6275 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6276 into "clz, "rbit". Rename "shift" to "shift_imm".
6277 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6278 changes. Update for attribute changes all occurrences of arlo_* and
6280 * config/arm/arm-fixed.md: Update for attribute changes
6281 all occurrences of arlo_* types.
6282 * config/arm/thumb2.md: Update for attribute changes all occurrences
6284 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6285 (cortexa7_older_only): Likewise.
6286 (cortexa7_younger): Likewise.
6287 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6288 (1020alu_shift_op): Likewise.
6289 (1020alu_shift_reg_op): Likewise.
6290 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6291 (alu_shift_op): Likewise.
6292 (alu_shift_reg_op): Likewise.
6293 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6294 (11_alu_shift_op): Likewise.
6295 (11_alu_shift_reg_op): Likewise.
6296 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6297 (9_alu_shift_reg_op): Likewise.
6298 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6300 (cortex_a15_alu_shift): Likewise.
6301 (cortex_a15_alu_shift_reg): Likewise.
6302 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6304 (cortex_a5_alu_shift): Likewise.
6305 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6307 (cortex_a53_alu_shift): Likewise.
6308 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6310 (cortex_a7_alu_reg): Likewise.
6311 (cortex_a7_alu_shift): Likewise.
6312 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6314 (cortex_a8_alu_shift): Likewise.
6315 (cortex_a8_alu_shift_reg): Likewise.
6316 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6317 (cortex_a9_dp_shift): Likewise.
6318 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6320 * config/arm/cortex-r4.md
6321 (cortex_r4_alu): Update for attribute changes.
6322 (cortex_r4_mov): Likewise.
6323 (cortex_r4_alu_shift_reg): Likewise.
6324 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6325 (526_alu_shift_op): Likewise.
6326 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6327 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6328 (626te_alu_shift_op): Likewise.
6329 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6330 (726te_alu_shift_op): Likewise.
6331 (726te_alu_shift_reg_op): Likewise.
6332 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6333 (mp626_alu_shift_op): Likewise.
6334 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6335 (pj4_alu_conds): Likewise.
6336 (pj4_shift): Likewise.
6337 (pj4_shift_conds): Likewise.
6338 (pj4_alu_shift): Likewise.
6339 (pj4_alu_shift_conds): Likewise.
6340 * config/aarch64/aarch64.md: Update for attribute change
6341 all occurrences of arlo_* and shift* types.
6343 2013-09-05 Mike Stump <mikestump@comcast.net>
6345 * tree.h: Move documentation for tree_function_decl to tree-core.h
6346 with the declaration.
6348 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6351 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6352 looking for widest mode.
6354 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6356 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6358 2013-09-05 Richard Biener <rguenther@suse.de>
6360 PR tree-optimization/58137
6361 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6362 Do not create vectors of pointers.
6363 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6364 types for the components of the vector initializer.
6365 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6366 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6368 2013-09-05 Martin Jambor <mjambor@suse.cz>
6370 * ipa-prop.c (remove_described_reference): Accept missing references,
6371 return false if that hppens, otherwise return true.
6372 (cgraph_node_for_jfunc): New function.
6373 (try_decrement_rdesc_refcount): Likewise.
6374 (try_make_edge_direct_simple_call): Use them.
6375 (ipa_edge_removal_hook): Remove references from rdescs.
6376 (ipa_edge_duplication_hook): Clone rdescs and their references
6377 when the new edge has the same caller as the old one.
6378 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6379 reference before removing any edges.
6381 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6383 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6385 * thumb2.md (thumb2_storewb_parisi): New pattern.
6387 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6389 * config/aarch64/aarch64-option-extensions.def: Add
6390 AARCH64_OPT_EXTENSION of 'crc'.
6391 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6392 (AARCH64_ISA_CRC): Ditto.
6393 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6394 description of the CRC extension.
6396 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6398 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6399 * config/rs6000/linux.h: Ditto.
6400 * alpha/linux.h: Ditto.
6401 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6402 no_c99_libc_has_function.
6403 * config/c6x/uclinux-elf.h: Ditto.
6404 * config/lm32/uclinux-elf.h: Ditto.
6405 * config/m68k/uclinux.h: Ditto.
6406 * config/moxie/uclinux.h: Ditto.
6407 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6408 (crisv32-*-linux*, cris-*-linux*): Ditto.
6409 * config/bfin/bfin.c: Include "tm_p.h".
6411 2013-09-05 Richard Biener <rguenther@suse.de>
6413 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6414 check for a definition without a basic-block.
6416 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6417 Sofiane Naci <sofiane.naci@arm.com>
6419 * config/aarch64/aarch64.md
6420 (*movti_aarch64): Rename r_2_f and f_2_r.
6421 (*movsf_aarch64): Likewise.
6422 (*movdf_aarch64): Likewise.
6423 (*movtf_aarch64): Likewise.
6424 (aarch64_movdi_<mode>low): Likewise.
6425 (aarch64_movdi_<mode>high): Likewise.
6426 (aarch64_mov<mode>high_di): Likewise.
6427 (aarch64_mov<mode>low_di): Likewise.
6428 (aarch64_movtilow_tilow): Likewise.
6429 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6430 values to config/arm/types.md
6431 (attribute "conds"): Update for attribute change.
6432 (anddi3_insn): Likewise.
6433 (iordi3_insn): Likewise.
6434 (xordi3_insn): Likewise.
6435 (one_cmpldi2): Likewise.
6436 * config/arm/types.md (type): Add Neon types.
6437 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6438 use "type" attribute.
6439 (movmisalign<mode>_neon_store): Likewise.
6440 (movmisalign<mode>_neon_load): Likewise.
6441 (vec_set<mode>_internal): Likewise.
6442 (vec_setv2di_internal): Likewise.
6443 (vec_extract<mode>): Likewise.
6444 (vec_extractv2di): Likewise.
6445 (add<mode>3_neon): Likewise.
6446 (adddi3_neon): Likewise.
6447 (sub<mode>3_neon): Likewise.
6448 (subdi3_neon): Likewise.
6449 (mul<mode>3_neon): Likewise.
6450 (mul<mode>3add<mode>_neon): Likewise.
6451 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6452 (fma<VCVTF:mode>4)): Likewise.
6453 (fma<VCVTF:mode>4_intrinsic): Likewise.
6454 (fmsub<VCVTF:mode>4)): Likewise.
6455 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6456 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6457 (ior<mode>3): Likewise.
6458 (and<mode>3): Likewise.
6459 (anddi3_neon): Likewise.
6460 (orn<mode>3_neon): Likewise.
6461 (orndi3_neon): Likewise.
6462 (bic<mode>3_neon): Likewise.
6463 (bicdi3_neon): Likewise.
6464 (xor<mode>3): Likewise.
6465 (one_cmpl<mode>2): Likewise.
6466 (abs<mode>2): Likewise.
6467 (neg<mode>2): Likewise.
6468 (umin<mode>3_neon): Likewise.
6469 (umax<mode>3_neon): Likewise.
6470 (smin<mode>3_neon): Likewise.
6471 (smax<mode>3_neon): Likewise.
6472 (vashl<mode>3): Likewise.
6473 (vashr<mode>3_imm): Likewise.
6474 (vlshr<mode>3_imm): Likewise.
6475 (ashl<mode>3_signed): Likewise.
6476 (ashl<mode>3_unsigned): Likewise.
6477 (neon_load_count): Likewise.
6478 (ashldi3_neon_noclobber): Likewise.
6479 (signed_shift_di3_neon): Likewise.
6480 (unsigned_shift_di3_neon): Likewise.
6481 (ashrdi3_neon_imm_noclobber): Likewise.
6482 (lshrdi3_neon_imm_noclobber): Likewise.
6483 (widen_ssum<mode>3): Likewise.
6484 (widen_usum<mode>3): Likewise.
6485 (quad_halves_<code>v4si): Likewise.
6486 (quad_halves_<code>v4sf): Likewise.
6487 (quad_halves_<code>v8hi): Likewise.
6488 (quad_halves_<code>v16qi): Likewise.
6489 (reduc_splus_v2di): Likewise.
6490 (neon_vpadd_internal<mode>): Likewise.
6491 (neon_vpsmin<mode>): Likewise.
6492 (neon_vpsmax<mode>): Likewise.
6493 (neon_vpumin<mode>): Likewise.
6494 (neon_vpumax<mode>): Likewise.
6495 (ss_add<mode>_neon): Likewise.
6496 (us_add<mode>_neon): Likewise.
6497 (ss_sub<mode>_neon): Likewise.
6498 (us_sub<mode>_neon): Likewise.
6499 (neon_vadd<mode>_unspec): Likewise.
6500 (neon_vaddl<mode>): Likewise.
6501 (neon_vaddw<mode>): Likewise.
6502 (neon_vhadd<mode>): Likewise.
6503 (neon_vqadd<mode>): Likewise.
6504 (neon_vaddhn<mode>): Likewise.
6505 (neon_vmul<mode>): Likewise.
6506 (neon_vmla<mode>): Likewise.
6507 (neon_vmlal<mode>): Likewise.
6508 (neon_vmls<mode>): Likewise.
6509 (neon_vmlsl<mode>): Likewise.
6510 (neon_vqdmulh<mode>): Likewise.
6511 (neon_vqdmlal<mode>): Likewise.
6512 (neon_vqdmlsl<mode>): Likewise.
6513 (neon_vmull<mode>): Likewise.
6514 (neon_vqdmull<mode>): Likewise.
6515 (neon_vsub<mode>_unspec): Likewise.
6516 (neon_vsubl<mode>): Likewise.
6517 (neon_vsubw<mode>): Likewise.
6518 (neon_vqsub<mode>): Likewise.
6519 (neon_vhsub<mode>): Likewise.
6520 (neon_vsubhn<mode>): Likewise.
6521 (neon_vceq<mode>): Likewise.
6522 (neon_vcge<mode>): Likewise.
6523 (neon_vcgeu<mode>): Likewise.
6524 (neon_vcgt<mode>): Likewise.
6525 (neon_vcgtu<mode>): Likewise.
6526 (neon_vcle<mode>): Likewise.
6527 (neon_vclt<mode>): Likewise.
6528 (neon_vcage<mode>): Likewise.
6529 (neon_vcagt<mode>): Likewise.
6530 (neon_vtst<mode>): Likewise.
6531 (neon_vabd<mode>): Likewise.
6532 (neon_vabdl<mode>): Likewise.
6533 (neon_vaba<mode>): Likewise.
6534 (neon_vabal<mode>): Likewise.
6535 (neon_vmax<mode>): Likewise.
6536 (neon_vmin<mode>): Likewise.
6537 (neon_vpaddl<mode>): Likewise.
6538 (neon_vpadal<mode>): Likewise.
6539 (neon_vpmax<mode>): Likewise.
6540 (neon_vpmin<mode>): Likewise.
6541 (neon_vrecps<mode>): Likewise.
6542 (neon_vrsqrts<mode>): Likewise.
6543 (neon_vqabs<mode>): Likewise.
6544 (neon_vqneg<mode>): Likewise.
6545 (neon_vcls<mode>): Likewise.
6546 (clz<mode>2): Likewise.
6547 (popcount<mode>2): Likewise.
6548 (neon_vrecpe): Likewise.
6549 (neon_vrsqrte): Likewise.
6550 (neon_vget_lane<mode>_sext_internal): Likewise.
6551 (neon_vget_lane<mode>_zext_internal): Likewise.
6552 (neon_vdup_n<mode>): Likewise.
6553 (neon_vdup_nv2di): Likewise.
6554 (neon_vdpu_lane<mode>_internal): Likewise.
6555 (neon_vswp<mode>): Likewise.
6556 (float<mode><V_cvtto>2): Likewise.
6557 (floatuns<mode><V_cvtto>2): Likewise.
6558 (fix_trunc<mode><V_cvtto>)2): Likewise
6559 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6560 (neon_vcvt<mode>): Likewise.
6561 (neon_vcvtv4sfv4hf): Likewise.
6562 (neon_vcvtv4hfv4sf): Likewise.
6563 (neon_vcvt_n<mode>): Likewise.
6564 (neon_vmovn<mode>): Likewise.
6565 (neon_vqmovn<mode>): Likewise.
6566 (neon_vqmovun<mode>): Likewise.
6567 (neon_vmovl<mode>): Likewise.
6568 (neon_vmul_lane<mode>): Likewise.
6569 (neon_vmull_lane<mode>): Likewise.
6570 (neon_vqdmull_lane<mode>): Likewise.
6571 (neon_vqdmulh_lane<mode>): Likewise.
6572 (neon_vmla_lane<mode>): Likewise.
6573 (neon_vmlal_lane<mode>): Likewise.
6574 (neon_vqdmlal_lane<mode>): Likewise.
6575 (neon_vmls_lane<mode>): Likewise.
6576 (neon_vmlsl_lane<mode>): Likewise.
6577 (neon_vqdmlsl_lane<mode>): Likewise.
6578 (neon_vext<mode>): Likewise.
6579 (neon_vrev64<mode>): Likewise.
6580 (neon_vrev32<mode>): Likewise.
6581 (neon_vrev16<mode>): Likewise.
6582 (neon_vbsl<mode>_internal): Likewise.
6583 (neon_vshl<mode>): Likewise.
6584 (neon_vqshl<mode>): Likewise.
6585 (neon_vshr_n<mode>): Likewise.
6586 (neon_vshrn_n<mode>): Likewise.
6587 (neon_vqshrn_n<mode>): Likewise.
6588 (neon_vqshrun_n<mode>): Likewise.
6589 (neon_vshl_n<mode>): Likewise.
6590 (neon_vqshl_n<mode>): Likewise.
6591 (neon_vqshlu_n<mode>): Likewise.
6592 (neon_vshll_n<mode>): Likewise.
6593 (neon_vsra_n<mode>): Likewise.
6594 (neon_vsri_n<mode>): Likewise.
6595 (neon_vsli_n<mode>): Likewise.
6596 (neon_vtbl1v8qi): Likewise.
6597 (neon_vtbl2v8qi): Likewise.
6598 (neon_vtbl3v8qi): Likewise.
6599 (neon_vtbl4v8qi): Likewise.
6600 (neon_vtbx1v8qi): Likewise.
6601 (neon_vtbx2v8qi): Likewise.
6602 (neon_vtbx3v8qi): Likewise.
6603 (neon_vtbx4v8qi): Likewise.
6604 (neon_vtrn<mode>_internal): Likewise.
6605 (neon_vzip<mode>_internal): Likewise.
6606 (neon_vuzp<mode>_internal): Likewise.
6607 (neon_vld1<mode>): Likewise.
6608 (neon_vld1_lane<mode>): Likewise.
6609 (neon_vld1_dup<mode>): Likewise.
6610 (neon_vld1_dupv2di): Likewise.
6611 (neon_vst1<mode>): Likewise.
6612 (neon_vst1_lane<mode>): Likewise.
6613 (neon_vld2<mode>): Likewise.
6614 (neon_vld2_lane<mode>): Likewise.
6615 (neon_vld2_dup<mode>): Likewise.
6616 (neon_vst2<mode>): Likewise.
6617 (neon_vst2_lane<mode>): Likewise.
6618 (neon_vld3<mode>): Likewise.
6619 (neon_vld3qa<mode>): Likewise.
6620 (neon_vld3qb<mode>): Likewise.
6621 (neon_vld3_lane<mode>): Likewise.
6622 (neon_vld3_dup<mode>): Likewise.
6623 (neon_vst3<mode>): Likewise.
6624 (neon_vst3qa<mode>): Likewise.
6625 (neon_vst3qb<mode>): Likewise.
6626 (neon_vst3_lane<mode>): Likewise.
6627 (neon_vld4<mode>): Likewise.
6628 (neon_vld4qa<mode>): Likewise.
6629 (neon_vld4qb<mode>): Likewise.
6630 (neon_vld4_lane<mode>): Likewise.
6631 (neon_vld4_dup<mode>): Likewise.
6632 (neon_vst4<mode>): Likewise.
6633 (neon_vst4qa<mode>): Likewise.
6634 (neon_vst4qb<mode>): Likewise.
6635 (neon_vst4_lane<mode>): Likewise.
6636 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6637 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6638 (neon_vec_<US>mult_lo_<mode>): Likewise.
6639 (neon_vec_<US>mult_hi_<mode>): Likewise.
6640 (neon_vec_<US>shiftl_<mode>): Likewise.
6641 (neon_unpack<US>_<mode>): Likewise.
6642 (neon_vec_<US>mult_<mode>): Likewise.
6643 (vec_pack_trunc_<mode>): Likewise.
6644 (neon_vec_pack_trunk_<mode>): Likewise.
6645 (neon_vabd<mode>_2): Likewise.
6646 (neon_vabd<mode>_3): Likewise.
6647 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6648 (thumb2_movsi_vfp): Likewise.
6649 (movdi_vfp): Likewise.
6650 (movdi_vfp_cortexa8): Likewise.
6651 (movhf_vfp_neon): Likewise.
6652 (movhf_vfp): Likewiwse.
6653 (movsf_vfp): Likewiwse.
6654 (thumb2_movsf_vfp): Likewiwse.
6655 (movdf_vfp): Likewise.
6656 (thumb2_movdf_vfp): Likewise.
6657 (movsfcc_vfp): Likewise.
6658 (thumb2_movsfcc_vfp): Likewise.
6659 (movdfcc_vfp): Likewise.
6660 (thumb2_movdfcc_vfp): Likewise.
6661 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6662 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6663 (v10_v2c): Likewise.
6664 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6666 (cortex_a15_neon_int_2): Likewise.
6667 (cortex_a15_neon_int_3): Likewise.
6668 (cortex_a15_neon_int_4): Likewise.
6669 (cortex_a15_neon_int_5): Likewise.
6670 (cortex_a15_neon_vqneg_vqabs): Likewise.
6671 (cortex_a15_neon_vmov): Likewise.
6672 (cortex_a15_neon_vaba): Likewise.
6673 (cortex_a15_neon_vaba_qqq): Likewise.
6674 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6675 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6676 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6677 scalar_64_32_long_scalar): Likewise.
6678 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6679 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6680 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6681 lotype_qdd_64_32_long): Likewise.
6682 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6683 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6684 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6685 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6686 (cortex_a15_neon_shift_1): Likewise.
6687 (cortex_a15_neon_shift_2): Likewise.
6688 (cortex_a15_neon_shift_3): Likewise.
6689 (cortex_a15_neon_vshl_ddd): Likewise.
6690 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6691 (cortex_a15_neon_vsra_vrsra): Likewise.
6692 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6693 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6694 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6695 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6696 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6697 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6698 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6699 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6700 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6701 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6702 (cortex_a15_neon_bp_simple): Likewise.
6703 (cortex_a15_neon_bp_2cycle): Likewise.
6704 (cortex_a15_neon_bp_3cycle): Likewise.
6705 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6706 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6707 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6708 (cortex_a15_neon_vld2_4_regs): Likewise.
6709 (cortex_a15_neon_vld3_vld4): Likewise.
6710 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6711 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6712 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6713 (cortex_a15_neon_vst3_vst4): Likewise.
6714 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6715 (cortex_a15_neon_vld3_vld4_lane" 10
6716 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6717 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6718 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6719 (cortex_a15_neon_ldm_2): Likewise.0
6720 (cortex_a15_neon_stm_2): Likewise.
6721 (cortex_a15_neon_mcr): Likewise.
6722 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6723 (cortex_a15_neon_mrc): Likewise.
6724 (cortex_a15_neon_mrrc): Likewise.
6725 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6727 (cortex_a15_alu_shift): Likewise.
6728 (cortex_a15_alu_shift_reg): Likewise.
6729 (cortex_a15_mult32): Likewise.
6730 (cortex_a15_mult64): Likewise.
6731 (cortex_a15_block): Likewise.
6732 (cortex_a15_branch): Likewise.
6733 (cortex_a15_load1): Likewise.
6734 (cortex_a15_load3): Likewise.
6735 (cortex_a15_store1): Likewise.
6736 (cortex_a15_store3): Likewise.
6737 (cortex_a15_call): Likewise.
6738 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6739 (cortex_a5_f2r): Likewise.
6740 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6742 (cortex_a53_f2r): Likewise.
6743 * config/arm/cortex-a7.md
6744 (cortex_a7_branch): Update for attribute change.
6745 (cortex_a7_call): Likewise.
6746 (cortex_a7_alu_imm): Likewise.
6747 (cortex_a7_alu_reg): Likewise.
6748 (cortex_a7_alu_shift): Likewise.
6749 (cortex_a7_mul): Likewise.
6750 (cortex_a7_load1): Likewise.
6751 (cortex_a7_store1): Likewise.
6752 (cortex_a7_load2): Likewise.
6753 (cortex_a7_store2): Likewise.
6754 (cortex_a7_load3): Likewise.
6755 (cortex_a7_store3): Likewise.
6756 (cortex_a7_load4): Likewise.
6757 (cortex_a7_store4): Likewise.
6758 (cortex_a7_fpalu): Likewise.
6759 (cortex_a7_fconst): Likewise.
6760 (cortex_a7_fpmuls): Likewise.
6761 (cortex_a7_neon_mul): Likewise.
6762 (cortex_a7_fpmacs): Likewise.
6763 (cortex_a7_neon_mla: Likewise.
6764 (cortex_a7_fpmuld: Likewise.
6765 (cortex_a7_fpmacd: Likewise.
6766 (cortex_a7_fpfmad: Likewise.
6767 (cortex_a7_fdivs: Likewise.
6768 (cortex_a7_fdivd: Likewise.
6769 (cortex_a7_r2f: Likewise.
6770 (cortex_a7_f2r: Likewise.
6771 (cortex_a7_f_flags: Likewise.
6772 (cortex_a7_f_loads: Likewise.
6773 (cortex_a7_f_loadd: Likewise.
6774 (cortex_a7_f_stores: Likewise.
6775 (cortex_a7_f_stored: Likewise.
6776 (cortex_a7_neon): Likewise.
6777 * config/arm/cortex-a8-neon.md
6778 (cortex_a8_neon_mrc): Update for attribute change.
6779 (cortex_a8_neon_mrrc): Likewise.
6780 (cortex_a8_neon_int_1): Likewise.
6781 (cortex_a8_neon_int_2): Likewise.
6782 (cortex_a8_neon_int_3): Likewise.
6783 (cortex_a8_neon_int_4): Likewise.
6784 (cortex_a8_neon_int_5): Likewise.
6785 (cortex_a8_neon_vqneg_vqabs): Likewise.
6786 (cortex_a8_neon_vmov): Likewise.
6787 (cortex_a8_neon_vaba): Likewise.
6788 (cortex_a8_neon_vaba_qqq): Likewise.
6789 (cortex_a8_neon_vsma): Likewise.
6790 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6791 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6792 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6793 long_scalar): Likewise.
6794 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6795 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6796 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6797 long_scalar_qdd_64_32_long): Likewise.
6798 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6799 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6800 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6801 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6802 (cortex_a8_neon_shift_1): Likewise.
6803 (cortex_a8_neon_shift_2): Likewise.
6804 (cortex_a8_neon_shift_3): Likewise.
6805 (cortex_a8_neon_vshl_ddd): Likewise.
6806 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6807 (cortex_a8_neon_vsra_vrsra): Likewise.
6808 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6809 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6810 (cortex_a8_neon_fp_vsum): Likewise.
6811 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6812 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6813 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6814 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6815 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6816 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6817 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6818 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6819 (cortex_a8_neon_bp_simple): Likewise.
6820 (cortex_a8_neon_bp_2cycle): Likewise.
6821 (cortex_a8_neon_bp_3cycle): Likewise.
6822 (cortex_a8_neon_ldr): Likewise.
6823 (cortex_a8_neon_str): Likewise.
6824 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6825 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6826 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6827 (cortex_a8_neon_vld2_4_regs): Likewise.
6828 (cortex_a8_neon_vld3_vld4): Likewise.
6829 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6830 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6831 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6832 (cortex_a8_neon_vst3_vst4): Likewise.
6833 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6834 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6835 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6836 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6837 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6838 (cortex_a8_neon_mcr): Likewise.
6839 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6840 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6841 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6843 (ca9_neon_mrrc): Likewise.
6844 (cortex_a9_neon_int_1): Likewise.
6845 (cortex_a9_neon_int_2): Likewise.
6846 (cortex_a9_neon_int_3): Likewise.
6847 (cortex_a9_neon_int_4): Likewise.
6848 (cortex_a9_neon_int_5): Likewise.
6849 (cortex_a9_neon_vqneg_vqabs): Likewise.
6850 (cortex_a9_neon_vmov): Likewise.
6851 (cortex_a9_neon_vaba): Likewise.
6852 (cortex_a9_neon_vaba_qqq): Likewise.
6853 (cortex_a9_neon_vsma): Likewise.
6854 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6855 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6856 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6857 long_scalar): Likewise.
6858 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6859 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6860 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6861 long_scalar_qdd_64_32_long): Likewise.
6862 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6863 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6864 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6865 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6866 (cortex_a9_neon_shift_1): Likewise.
6867 (cortex_a9_neon_shift_2): Likewise.
6868 (cortex_a9_neon_shift_3): Likewise.
6869 (cortex_a9_neon_vshl_ddd): Likewise.
6870 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6871 (cortex_a9_neon_vsra_vrsra): Likewise.
6872 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6873 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6874 (cortex_a9_neon_fp_vsum): Likewise.
6875 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6876 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6877 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6878 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6879 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6880 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6881 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6882 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6883 (cortex_a9_neon_bp_simple): Likewise.
6884 (cortex_a9_neon_bp_2cycle): Likewise.
6885 (cortex_a9_neon_bp_3cycle): Likewise.
6886 (cortex_a9_neon_ldr): Likewise.
6887 (cortex_a9_neon_str): Likewise.
6888 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6889 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6890 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6891 (cortex_a9_neon_vld2_4_regs): Likewise.
6892 (cortex_a9_neon_vld3_vld4): Likewise.
6893 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6894 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6895 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6896 (cortex_a9_neon_vst3_vst4): Likewise.
6897 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6898 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6899 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6900 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6901 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6902 (cortex_a9_neon_mcr): Likewise.
6903 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6904 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6905 (cortex_a9_fps): Likewise.
6906 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6908 (cortex_m4_fmuls): Likewise.
6909 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6911 (cortex_r4_mrc): Likewise.
6912 * config/arm/iterators.md: Update comment referring to neon_type.
6913 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
6914 (iwmmxt_movsi_insn): Likewise.
6915 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6917 (pj4_core_to_vfp): Likewise.
6918 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6920 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6921 (vfp_fstore): Likewise.
6922 * doc/md.texi: Change references to neon_type to refer to type.
6924 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6926 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6928 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
6930 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6933 2013-09-04 Easwaran Raman <eraman@google.com>
6936 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6937 (build_and_add_sum): Use it.
6938 (appears_later_in_bb): Simplify code.
6940 2013-09-04 Teresa Johnson <tejohnson@google.com>
6942 * dumpfile.c (dump_finish): Don't close stderr/stdout.
6944 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
6946 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
6948 2013-09-04 Jan Hubicka <jh@suse.cz>
6950 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
6951 * ipa-devirt.c: Include diganostic.h
6952 (odr_type_d): Add types and types_set.
6953 (hash_type_name): Work for types with vtables during LTO.
6954 (odr_hasher::remove): Fix comment; destroy types_set.
6955 (add_type_duplicate): New function,
6956 (get_odr_type): Use it.
6957 (dump_type_inheritance_graph): Dump type duplicates.
6958 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
6960 * tree.c (types_same_for_odr): Give exact answers on types with
6963 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6965 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
6966 explaining their differences.
6968 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
6970 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
6972 2013-09-03 Jeff Law <law@redhat.com>
6974 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
6975 when not threading through a joiner block. Pass joiner/no joiner
6976 state to register_jump_thread.
6977 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
6978 state from argument rather than implying on path length.
6979 Dump the entire jump thread path into debugging dump.
6980 * tree-flow.h (register_jump_thread): Update prototype.
6982 2013-08-29 Xinliang David Li <davidxl@google.com>
6984 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6985 Remove a trivial gcc_assert.
6987 2013-08-29 Xinliang David Li <davidxl@google.com>
6989 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
6990 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
6991 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6992 Delay base decl alignment adjustment.
6993 * tree-vectorizer.c (vect_destroy_datarefs): New function.
6994 * tree-vectorizer.h: New data structure.
6995 (set_dr_misalignment): New function.
6996 (dr_misalignment): Ditto.
6997 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
6998 (vectorizable_load): Ditto.
6999 (ensure_base_align): New function.
7000 (vectorize_loops): Add dbg_cnt support.
7001 (execute_vect_slp): Ditto.
7002 * dbgcnt.def: New debug counter.
7003 * Makefile: New dependency.
7005 2013-09-03 Meador Inge <meadori@codesourcery.com>
7009 2013-08-30 Meador Inge <meadori@codesourcery.com>
7011 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7013 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7015 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7016 function descriptor.
7018 2013-09-03 Richard Biener <rguenther@suse.de>
7020 * tree-affine.c (add_elt_to_tree): Fix association issue,
7021 avoid useless converts and make sure to always return a
7022 properly typed result.
7024 2013-09-03 Richard Biener <rguenther@suse.de>
7027 * fold-const.c (negate_expr_p): Fix division case.
7028 (negate_expr): Likewise.
7030 2013-09-03 Richard Biener <rguenther@suse.de>
7033 * tree-streamer-out.c: Include tm.h.
7034 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7036 2013-09-03 Jan Hubicka <jh@suse.cz>
7038 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7040 2013-09-03 Alan Modra <amodra@gmail.com>
7042 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7043 multilibs to big-endian target and vice versa.
7044 * config/rs6000/t-linux64: Use := assignment on all vars.
7045 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7046 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7047 * config/rs6000/t-linux64le: New file.
7048 * config/rs6000/t-linux64bele: New file.
7049 * config/rs6000/t-linux64lebe: New file.
7051 2013-09-02 Jan Hubicka <jh@suse.cz>
7053 * ipa-inline-transform.c (inline_transform): Do not
7054 optimize_inline_calls when not optimizing.
7056 2013-09-02 Jan Hubicka <jh@suse.cz>
7058 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7059 duplicated nodes for assembler names.
7060 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7061 hard registers from assembler name hash.
7063 2013-09-02 Jan Hubicka <jh@suse.cz>
7065 * ipa-split.c (execute_split_functions): Split externally visible
7066 functions called once.
7068 2013-09-02 Martin Jambor <mjambor@suse.cz>
7071 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7072 linked list. When finding the correct duplicate, also consider also
7073 the caller in additon to its inlined_to node.
7075 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7077 * config/aarch64/aarch64-simd-builtins.def
7078 (dup_lane_scalar): Remove.
7079 * config/aarch64/aarch64-simd.md
7080 (aarch64_simd_dup): Add 'w->w' alternative.
7081 (aarch64_dup_lane<mode>): Allow for VALL.
7082 (aarch64_dup_lane_scalar<mode>): Remove.
7083 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7084 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7085 (aarch64_get_lane_unsigned<mode>): Likewise.
7086 (aarch64_get_lane<mode>): Likewise.
7087 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7088 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7089 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7090 (VCON): Change container of V2SF.
7091 (vswap_width_name): Likewise.
7092 * config/aarch64/arm_neon.h
7093 (__aarch64_vdup_lane_any): New.
7094 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7095 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7096 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7098 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7101 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7102 the source or the destination is a single hard register.
7104 2013-09-02 Richard Biener <rguenther@suse.de>
7107 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7110 2013-09-02 Richard Biener <rguenther@suse.de>
7112 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7113 arithmetic to sizetype.
7115 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7117 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7118 Find auto-increment use both before and after candidate.
7120 2013-09-02 Marek Polacek <polacek@redhat.com>
7122 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7124 2013-09-01 Jan Hubicka <jh@suse.cz>
7126 * Makefile.in: Add ipa-profile.o
7127 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7128 * cgraph.c (struct cgraph_propagate_frequency_data,
7129 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7130 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7131 * cgraph.h (cgraph_propagate_frequency): Remove.
7132 * ipa-inline-analysis.c: Include ipa-utils.h;
7133 drop duplicated cfgloop.h.
7134 (inline_update_callee_summaries): Update.
7135 * ipa-profile.c: New file.
7136 * ipa-utils.h (ipa_propagate_frequency): Declare.
7137 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7138 data-streamer.h, value-prof.h.
7139 (symtab_remove_unreachable_nodes): Update profile.
7140 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7141 account_time_size, cmp_counts, dump_histogram,
7142 ipa_profile_generate_summary, ipa_profile_write_summary,
7143 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7144 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7145 Move to ipa-profile.c.
7147 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7149 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7151 2013-09-01 Jan Hubicka <jh@suse.cz>
7153 * common.opt (fdevirtualize-speculatively): New function.
7154 * invoke.texi (fdevirtualize-speculatively): Document.
7155 * ipa-devirt.c: Include ipa-inline.h
7156 (likely_target_p): New function.
7157 (ipa_devirt): New function.
7158 (gate_ipa_devirt): New function.
7159 (pass_data_ipa_devirt): New static var.
7160 (pass_ipa_devirt): Likewise.
7161 (make_pass_ipa_devirt): New function.
7162 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7163 (common_handle_option): Disable devirtualization when
7164 value range profiling is available.
7165 * passes.def (pass_ipa_devirt): Add.
7166 * timever.def (TV_IPA_DEVIRT): New timevar.
7167 * tree-pass.h (make_pass_ipa_devirt):
7169 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7171 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7172 include sanitize(undefined).
7174 2013-08-31 Diego Novillo <dnovillo@google.com>
7176 * Makefile.in (TREE_CORE_H): Define.
7178 (GTFILES): Add tree-core.h.
7179 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7181 * tree-core.h: New file.
7182 Move all data structures, enum, typedefs, global
7183 declarations and constants from ...
7186 2013-08-31 Jan Hubicka <jh@suse.cz>
7188 * bulitins.c (expand_builtin): Do not early exit for gcov
7189 instrumented functions.
7191 2013-08-31 Marek Polacek <polacek@redhat.com>
7193 * ubsan.c: Include tm_p.h.
7195 2013-08-31 Jan Hubicka <jh@suse.cz>
7197 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7200 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7201 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7202 * tree-cfg.c (verify_gimple_label): ... here.
7203 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7204 (ipa_merge_profiles): New function.
7205 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7206 (lto_input_function_body): Likewise.
7207 * ipa-utils.h (ipa_merge_profiles): Declare.
7208 * lto-streamer.h (lto_input_function_body): Update prototype.
7209 (emit_label_in_global_context_p): Remove.
7210 * lto-symtab.c: Include ipa-utils.h
7211 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7213 2013-08-31 Jan Hubicka <jh@suse.cz>
7215 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7217 2013-08-31 Jan Hubicka <jh@suse.cz>
7219 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7221 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7223 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7224 "cmp" RTX before signed_comparison_operator check to account
7227 2013-08-30 Jan Hubicka <jh@suse.cz>
7229 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7230 (detect_type_change_1): Rename to ...
7231 (detect_type_change): ... this one; early return on non-polymorphic
7233 (detect_type_change_ssa): Add comp_type parameter; update
7234 use of detect_type_change.
7235 (compute_complex_assign_jump_func): Add param_type parameter;
7236 update use of detect_type_change_ssa.
7237 (compute_complex_ancestor_jump_func): Likewise.
7238 (ipa_get_callee_param_type): New function.
7239 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7240 update calls to the jump function computation functions.
7242 2013-08-30 Teresa Johnson <tejohnson@google.com>
7243 Steven Bosscher <steven@gcc.gnu.org>
7245 * cfgrtl.c (fixup_new_cold_bb): New routine.
7246 (commit_edge_insertions): Invoke fixup_partitions.
7247 (find_partition_fixes): New routine.
7248 (fixup_partitions): Ditto.
7249 (verify_hot_cold_block_grouping): Update comments.
7250 (rtl_verify_edges): Invoke find_partition_fixes.
7251 (rtl_verify_bb_pointers): Update comments.
7252 (rtl_verify_bb_layout): Ditto.
7253 * basic-block.h (probably_never_executed_edge_p): Declare.
7254 (fixup_partitions): Ditto.
7255 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7256 * bb-reorder.c (sanitize_hot_paths): New function.
7257 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7259 * predict.c (probably_never_executed_edge_p): New routine.
7260 * cfg.c (check_bb_profile): Add partition insanity warnings.
7262 2013-08-30 Meador Inge <meadori@codesourcery.com>
7264 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7266 2013-08-30 Marek Polacek <polacek@redhat.com>
7268 * Makefile.in (ubsan.o): Add.
7269 (c-family/c-ubsan.o): Add.
7270 (builtins.o): Add ubsan.h dependency.
7271 * ubsan.h: New file.
7272 * ubsan.c: New file.
7273 * common.opt: Add -fsanitize=undefined option.
7274 (flag_sanitize): Add variable.
7275 (fsanitize=): Add option. Add Driver.
7276 (fsanitize=thread): Remove option.
7277 (fsanitize=address): Likewise.
7278 (static-libubsan): New option.
7279 * doc/invoke.texi: Document the new flag and -static-libubsan.
7280 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7281 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7282 * builtin-attrs.def (ATTR_COLD): Define.
7283 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7284 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7285 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7286 * flag-types.h (sanitize_code): New enum.
7287 * opts.c (common_handle_option): Parse command line arguments
7288 of -fsanitize=. Add -fsanitize=unreachable option.
7289 * varasm.c (get_variable_section): Adjust.
7290 (assemble_noswitch_variable): Likewise.
7291 (assemble_variable): Likewise.
7292 (output_constant_def_contents): Likewise.
7293 (categorize_decl_for_section): Likewise.
7294 (place_block_symbol): Likewise.
7295 (output_object_block): Likewise.
7296 * builtins.def: Likewise.
7297 * toplev.c (compile_file): Likewise.
7298 (process_options): Likewise.
7299 * cppbuiltin.c: Likewise.
7300 * tsan.c (tsan_pass): Likewise.
7301 (tsan_gate): Likewise.
7302 (tsan_gate_O0): Likewise.
7303 * cfgexpand.c (partition_stack_vars): Likewise.
7304 (expand_stack_vars): Likewise.
7305 (defer_stack_allocation): Likewise.
7306 (expand_used_vars): Likewise.
7307 * cfgcleanup.c (old_insns_match_p): Likewise.
7308 * asan.c (asan_finish_file): Likewise.
7309 (asan_instrument): Likewise.
7310 (gate_asan): Likewise.
7311 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7312 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7313 (asan_global_struct): Use pointer_sized_int_node instead
7314 calling build_nonstandard_integer_type.
7315 (initialize_sanitizer_builtins): Likewise.
7316 (asan_finish_file): Likewise.
7317 * gcc.c: Document %{%:function(args):X}.
7318 (static_spec_functions): Add sanitize.
7319 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7320 store funcval != NULL there.
7321 (do_spec_1): Adjust handle_spec_function caller.
7322 (handle_braces): Allow %:function(args) as condition.
7323 (sanitize_spec_function): New function.
7324 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7325 (LIBUBSAN_SPEC): Likewise.
7326 (LIBUBSAN_EARLY_SPEC): Likewise.
7327 (SANITIZER_SPEC): Handle libubsan.
7328 (SANITIZER_EARLY_SPEC): Likewise.
7329 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7330 instead of fsanitize=address.
7331 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7332 instead of fsanitize=address*.
7333 * builtins.c: Include ubsan.h.
7334 (fold_builtin_0): Instrument __builtin_unreachable.
7335 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7336 instead of flag_asan.
7337 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7338 (pointer_sized_int_node): Define.
7339 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7341 2013-08-30 Mike Stump <mikestump@comcast.net>
7343 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7346 2013-08-29 Jan Hubicka <jh@suse.cz>
7348 * cgraph.c (cgraph_function_body_availability): Handle weakref
7350 * passes.def: Remove pass_fixup_cfg.
7351 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7352 track when we need to remove functions.
7353 (gate_ipa_inline): Execute inlining always; add comment why.
7354 (pass_data_ipa_inline): Remove TODO_remove_functions.
7355 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7356 do not produce summaries.
7357 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7358 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7359 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7360 constant pool and vtable.
7362 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7364 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7365 New arm_neon.h's internal macros to specify 64-bit constants.
7366 Avoid using stdint.h's macros.
7368 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7370 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7371 before applying REG_P.
7373 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7375 PR tree-optimization/58277
7376 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7377 after seeing too many stmts with vdef in between dombb and current
7378 bb, invalidate everything.
7380 2013-08-30 Richard Biener <rguenther@suse.de>
7382 * fold-const.c (fold_single_bit_test): Fix overflow test.
7384 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7386 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7387 and which can live in a register, always retrieve the value on entry.
7388 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7389 passed by invisible reference specially.
7390 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7391 (vt_add_function_parameter): Correctly deal with a parameter passed by
7392 invisible reference.
7394 2013-08-30 Jan Hubicka <jh@suse.cz>
7396 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7398 2013-08-30 Richard Biener <rguenther@suse.de>
7400 PR tree-optimization/58228
7401 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7402 allow invariant loads in nested loop vectorization.
7404 2013-08-30 Richard Biener <rguenther@suse.de>
7406 PR tree-optimization/58223
7407 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7408 (has_anti_or_output_dependence): ... this and adjust to also
7409 look for output dependences.
7410 (mark_nodes_having_upstream_mem_writes): Adjust.
7411 (rdg_flag_uses): Likewise.
7413 2013-08-30 Richard Biener <rguenther@suse.de>
7415 PR tree-optimization/58010
7416 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7417 assert that we have a loop-closed PHI.
7419 2013-08-29 Jan Hubicka <jh@suse.cz>
7421 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7422 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7423 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7424 (lto_free_function_in_decl_state_for_node): New function.
7426 2013-08-29 Xinliang David Li <davidxl@google.com>
7428 * loop-unroll.c (report_unroll_peel): Minor message change.
7429 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7430 Emit alignment peeling message with default -fopt-info.
7431 (vect_loop_versioning): Emit loop version info message.
7432 * tree-vectorizer.c (vectorize_loops): Minor message change.
7433 (execute_vect_slp): Ditto.
7435 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7437 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7438 of the clone from the DECL_NAME of the original function.
7440 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7442 * passes.c (register_pass): Add overload.
7443 * tree-pass.h (register_pass): Forward declare it. Add comment.
7445 2013-08-29 Jan Hubicka <jh@suse.cz>
7447 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7448 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7449 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7450 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7451 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7452 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7453 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7454 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7456 2013-08-29 Teresa Johnson <tejohnson@google.com>
7458 * dumpfile.c (dump_loc): Output column number.
7459 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7460 * doc/invoke.texi: Document optall -fopt-info flag.
7461 * profile.c (read_profile_edge_counts): Use new dump framework.
7462 (compute_branch_probabilities): Ditto.
7463 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7464 when pass not in any opt group.
7465 * pass_manager.h (pass_manager::get_pass_profile): New method.
7466 * value-prof.c (check_counter): Use new dump framework.
7467 (check_ic_target): Ditto.
7468 * coverage.c (get_coverage_counts): Ditto.
7469 (coverage_init): Setup new dump framework.
7471 2013-08-29 Richard Biener <rguenther@suse.de>
7473 PR tree-optimization/58246
7474 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7475 handle the dominance check inside a basic-block.
7477 2013-08-29 Richard Biener <rguenther@suse.de>
7480 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7481 of default defs that appear in abnormal PHI nodes.
7483 2013-08-29 Richard Biener <rguenther@suse.de>
7485 PR tree-optimization/57685
7486 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7487 single-use operands to avoid exponential complexity.
7489 2013-08-28 Dehao Chen <dehao@google.com>
7491 * ipa-inline.c (edge_badness): Fix integer underflow.
7493 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7495 * gtm-builtins.def (_ITM_free): Declare leaf.
7497 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7500 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7501 (*tls_local_dynamic_base_64_largepic): Likewise.
7502 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7503 Remove predicate from call operand.
7504 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7505 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7507 2013-08-28 Jeff Law <law@redhat.com>
7509 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7510 checks for the number of predecessors and successors allowed.
7511 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7512 which require copying a joiner block if there is a request which
7513 is a subpath that requires no joiner block copying.
7515 2013-08-28 Jan Hubicka <jh@suse.cz>
7517 * lto-streamer-out.c (DFS_write_tree_body): Drop
7518 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7519 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7520 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7521 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7522 Do not read DECL_ERROR_ISSUED.
7523 (unpack_ts_decl_with_vis_value_fields): Do not read
7525 (lto_input_ts_binfo_tree_pointers): Do not read
7526 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7527 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7528 write DECL_ERROR_ISSUED..
7529 (pack_ts_decl_with_vis_value_fields): Do not write
7531 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7532 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7533 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7534 * tree.h (tree_decl_common): Update comment.
7535 (DECL_ERROR_ISSUED): Remove.
7537 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7540 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7542 2013-08-28 Jan Hubicka <jh@suse.cz>
7544 * builtins.def (free): Declare leaf.
7546 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7549 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7550 * configure: Regenerate.
7552 2013-08-27 Martin Jambor <mjambor@suse.cz>
7554 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7555 (ipa_ancestor_jf_data): Likewise.
7556 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7557 (ipa_get_jf_pass_through_type_preserved): New function.
7558 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7559 (ipa_get_jf_ancestor_type_preserved): New function.
7560 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7561 (ipa_get_jf_ancestor_result): Likewise.
7562 (propagate_vals_accross_pass_through): Use
7563 ipa_get_jf_pass_through_result to do all the value mappings.
7564 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7565 type_preserved flag.
7566 (ipa_set_jf_cst_copy): New function.
7567 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7568 (ipa_set_jf_arith_pass_through): Likewise.
7569 (ipa_set_ancestor_jf): Likewise.
7570 (compute_complex_assign_jump_func): Set type_preserved instead of
7572 (ipa_compute_jump_functions_for_edge): Likewise.
7573 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7574 (update_jump_functions_after_inlining): Update type_preserved.
7575 Explicitely create jump functions when combining one with pass_through.
7576 (ipa_write_jump_function): Stream the type_preserved flags.
7577 (ipa_read_jump_function): Likewise.
7579 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7580 Aldy Hernandez <aldyh@redhat.com>
7582 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7583 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7584 * function.h (struct function): Add has_force_vect_loops and
7586 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7587 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7589 * gimple.def: Update CLAUSES comments.
7590 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7591 (gimple_build_omp_for): Add argument to prototype.
7592 (gimple_omp_for_kind): New.
7593 (gimple_omp_for_set_kind): New.
7594 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7595 GOVD_DATA_SHARE_CLASS.
7596 (enum omp_region_type): Add ORT_SIMD.
7597 (gimple_add_tmp_var): Handle ORT_SIMD.
7598 (gimplify_var_or_parm_decl): Same.
7599 (is_gimple_stmt): Same.
7600 (omp_firstprivatize_variable): Same.
7601 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7602 (omp_notice_variable): Handle ORT_SIMD.
7603 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7604 (omp_check_private): Handle ORT_SIMD.
7605 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7607 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7608 Handle OMP_CLAUSE_LASTPRIVATE.
7609 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7611 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7612 (gimplify_expr): Handle OMP_SIMD.
7613 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7614 (expand_GOMP_SIMD_VF): New.
7615 (expand_GOMP_SIMD_LAST_LANE): New.
7616 * internal-fn.def (GOMP_SIMD_LANE): New.
7617 (GOMP_SIMD_VF): New.
7618 (GOMP_SIMD_LAST_LANE): New.
7619 * omp-low.c: Include target.h.
7620 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7622 (check_omp_nesting_restrictions): Same.
7624 (lower_rec_simd_input_clauses): New.
7625 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7627 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7628 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7629 (expand_omp_build_assign): New.
7630 (expand_omp_for_init_counts): New.
7631 (expand_omp_for_init_vars): New.
7632 (extract_omp_for_update_vars): New.
7633 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7634 and rewrite accordingly.
7635 (expand_omp_simd): New.
7636 (expand_omp_for): Use expand_omp_simd.
7637 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7638 (lower_omp_for): Do not lower the body.
7639 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7641 * tree-flow.h (find_omp_clause): Remove prototype.
7642 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7643 forcing vectorization of the loop, or if flag_tree_vectorize.
7644 (gate_tree_if_conversion): Similarly.
7645 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7646 gimple_build_omp_for.
7647 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7648 * tree-parloops (create_parallel_loop): Pass kind argument to
7649 gimple_build_omp_for.
7650 * tree-pretty-print.c (dump_omp_clause): Add cases for
7651 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7652 OMP_CLAUSE__SIMDUID_.
7653 (dump_generic_node): Handle OMP_SIMD.
7654 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7655 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7656 unroll OMP_SIMD loops here.
7657 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7658 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7660 (vect_analyze_data_refs): Handle simd loops.
7661 * tree-vect-loop.c (vectorizable_live_operation): Handle
7663 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7664 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7665 (vectorizable_load): Same.
7666 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7667 (struct simduid_to_vf): New.
7668 (simduid_to_vf::hash): New.
7669 (simduid_to-vf::equal): New.
7670 (struct simd_array_to_simduid): New.
7671 (simd_array_to_simduid::hash): New.
7672 (simd_array_to_simduid::equal): New.
7673 (adjust_simduid_builtins): New.
7674 (struct note_simd_array_uses_struct): New.
7675 (note_simd_array_uses_cb): New.
7676 (note_simd_array_uses): New.
7677 (vectorize_loops): Handle simd hints and adjust simd builtins
7679 * tree-vectorizer.h (struct _stmt_vec_info): Add
7680 simd_lane_access_p field.
7681 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7682 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7683 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7684 (omp_clause_code_name): Same.
7685 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7686 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7687 * tree.def (OMP_SIMD): New entry.
7688 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7689 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7690 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7691 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7692 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7693 (OMP_CLAUSE_LINEAR_STEP): New.
7694 (OMP_CLAUSE_SAFELEN_EXPR): New.
7695 (OMP_CLAUSE__SIMDUID__DECL): New.
7696 (find_omp_clause): New prototype.
7698 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7700 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7701 Haswell processor detection.
7703 2013-08-27 Christian Widmer <shadow@umbrox.de>
7706 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7707 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7708 AVX2 capable processors.
7710 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7712 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7713 of vget_low_* with implementations in terms of other intrinsics.
7715 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7718 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7719 values to -1, 0 and 1.
7721 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7723 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7724 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7725 (<optab><mode>3_insn): Remove.
7726 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7727 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7728 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7729 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7730 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7731 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7732 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7733 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7734 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7735 (ror<mode>3_insn): Likewise.
7736 * config/aarch64/predicates.md (aarch64_simd_register): New.
7738 2013-08-27 Richard Biener <rguenther@suse.de>
7740 PR tree-optimization/57521
7741 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7742 one edge is non-critical.
7743 (find_phi_replacement_condition): Make sure to use a non-critical
7744 edge. Cleanup and remove old bug workarounds.
7745 (bb_postdominates_preds): Remove.
7746 (if_convertible_loop_p_1): Do not compute post-dominators.
7747 (combine_blocks): Do not free post-dominators.
7748 (main_tree_if_conversion): Likewise.
7749 (pass_data_if_conversion): Add TODO_verify_ssa.
7751 2013-08-27 DJ Delorie <dj@redhat.com>
7753 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7755 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7757 * function.c (assign_parm_find_data_types): Set passed_mode and
7758 nominal_mode to the TYPE_MODE of nominal_type for the built
7759 pointer type in case of the struct-pass-by-reference.
7761 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7763 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7764 (UINT16_TYPE): Change default to "unsigned int".
7766 * config/avr/avr.opt (mfract-convert-truncate): New option.
7767 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7768 is set, round negative fractional integers according to n1169
7769 when converting to integer types.
7771 2013-08-26 Jan Hubicka <jh@suse.cz>
7773 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7774 methods can not be called indirectly when their address is not taken.
7776 2013-08-26 Jan Hubicka <jh@suse.cz>
7778 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7781 2013-08-26 Jan Hubicka <jh@suse.cz>
7783 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7786 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7788 * reload.c (find_valid_class): Allow classes that do not include
7789 FIRST_PSEUDO_REGISTER - 1.
7791 2013-08-26 Jan Hubicka <jh@suse.cz>
7793 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7794 fix edge count/frequency when speculation failed; fix type check
7795 for the direct call.
7797 2013-08-26 Jan Hubicka <jh@suse.cz>
7799 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7801 2013-08-26 Jan Hubicka <jh@suse.cz>
7803 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7804 to be changed by edge redirection.
7806 2013-08-26 Jan Hubicka <jh@suse.cz>
7808 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7809 formating; add sanity check.
7810 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7811 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7812 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7813 (resolve_noninline_speculation): Update callee keys, too.
7815 2013-08-26 Jan Hubicka <jh@suse.cz>
7817 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7818 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7820 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7822 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7825 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
7827 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7828 (x86_64_elf_select_section): Ditto.
7829 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7831 (ix86_pass_by_reference): Ditto.
7832 (output_set_got): Ditto.
7833 (ix86_unary_operator_ok): Ditto.
7834 (ix86_expand_builtin): Ditto.
7836 2013-08-23 Jan Hubicka <jh@suse.cz>
7838 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
7840 2013-08-23 Jan Hubicka <jh@suse.cz>
7842 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7843 (tree_decl_with_vis): Add FINAL field.
7845 2013-08-23 Jeff Law <law@redhat.com>
7847 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7848 the debugging dump when the expression is fully redundant.
7850 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7852 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7853 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7854 * pretty-print.c (pp_formatted_text_data): Likewise.
7855 (pp_write_text_to_stream): Likewise.
7856 (pp_write_text_as_dot_label_to_stream): Likewise.
7857 (pp_append_r): Likewise.
7858 (pp_format): Likewise.
7859 (pp_flush): Likewise.
7860 (pp_clear_output_area): Likewise.
7861 (pp_append_text): Likewise.
7862 (pp_formatted_text): Likewise.
7863 (pp_remaining_character_count_for_line): Likewise.
7864 (pp_newline): Likewise.
7865 (pp_character): Likewise.
7866 (output_buffer::~output_buffer): Define.
7867 (pretty_printer::~pretty_printer): Destruct output buffer.
7868 * pretty-print.h (output_buffer::~output_buffer): Declare.
7869 (pretty_printer::~pretty_printer): Declare virtual.
7871 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
7874 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7875 HOST_WIDE_INT_M1U): New macros.
7876 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7877 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7878 unsigned -1 for lshift.
7879 * cse.c (cse_insn): Likewise.
7880 * double-int.c (rshift_double, lshift_double): Likewise.
7881 * builtins.c (fold_builtin_bitop): Likewise.
7882 * combine.c (force_to_mode): Likewise.
7883 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7884 * simplify-rtx.c (simplify_const_unary_operation,
7885 simplify_const_binary_operation): Likewise.
7886 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7887 check_va_list_escapes): Likewise.
7888 * rtlanal.c (nonzero_bits1): Likewise.
7889 * expmed.c (expand_smod_pow2): Likewise.
7890 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7892 2013-08-23 Jan Hubicka <jh@suse.cz>
7894 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7895 as having address taken.
7897 2013-08-23 Jan Hubicka <jh@suse.cz>
7899 * ipa-utils.h (method_class_type): Declare.
7900 * ipa-devirt.c (method_class_type): Export.
7902 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7903 do not walk base classes of anonymous types.
7905 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7907 PR rtl-optimization/58220
7909 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7910 handle SEQUENCE insns properly.
7912 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7914 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7916 (pp_newline_and_indent): Likewise.
7917 (pp_separate_with): Likewise.
7918 * pretty-print.c (pp_newline_and_flush): Define.
7919 (pp_newline_and_indent): Likewise.
7920 (pp_separate_with): Likewise.
7922 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7925 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7926 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7928 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7930 * config/i386/predicates.md (ext_sse_reg_operand): New.
7931 * config/i386/i386.md (*movti_internal): Use
7932 predicate to determine if EVEX is needed.
7933 (*movsi_internal): Ditto.
7934 (*movdf_internal): Ditto.
7935 (*movsf_internal): Ditto.
7936 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
7938 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7940 PR tree-optimization/58209
7941 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
7942 (find_tail_calls): Give up for pointer result types if m is non-NULL.
7943 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
7944 emit POINTER_PLUS_EXPR.
7945 (create_tailcall_accumulator): For pointer result type accumulate in
7948 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
7950 * configure.ac: Add backslashes missing from the last change.
7951 * configure: Regenerate.
7953 2013-08-22 Jan Hubicka <jh@susue.cz>
7955 * ipa.c (function_and_variable_visibility): First remember function
7956 was global and then make it local.
7958 2013-08-22 Julian Brown <julian@codesourcery.com>
7960 * configure.ac: Add aarch64 to list of arches which use "nop" in
7962 * configure: Regenerate.
7964 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7966 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
7967 gnu_libc_has_function.
7968 * config/s390/tpf.h: Likewise.
7970 2013-08-22 Jan Hubicka <jh@susue.cz>
7972 * timevar.c (validate_phases): Add cast.
7974 2013-08-22 Jan Hubicka <jh@susue.cz>
7976 * timevar.c (validate_phases): Use size_t for memory.
7977 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
7979 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
7981 * pretty-print.h (output_buffer::output_buffer): Declare.
7982 (pretty_printer::pretty_printer): Likewise.
7983 (pp_construct): Remove.
7984 * pretty-print.c (output_buffer::output_buffer): Define.
7985 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
7986 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
7987 (print_gimple_expr): Likewise.
7988 (print_gimple_seq): Likewise.
7989 (gimple_dump_bb): Likewise.
7990 * sched-vis.c (dump_value_slim): Likewise.
7991 (dump_insn_slim): Likewise.
7992 (dump_rtl_slim): Likewise.
7993 (str_pattern_slim): Likewise.
7994 * tree-mudflap.c (mf_varname_tree): Likewise.
7995 * graph.c (print_graph_cfg): Likewise.
7996 (start_graph_dump): Likewise.
7997 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
7999 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8000 pretty printer initialization.
8001 * coretypes.h (diagnostic_context): Remove superflous type alias
8003 (pretty_printer): Likewise. Declare directly as a class.
8004 (pretty_print_info): Remove declaration as class.
8005 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8006 and pp_clear_output_area.
8007 (asan_add_global): Likewise.
8009 2013-08-22 Jan Hubicka <jh@suse.cz>
8011 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8012 * ipa-utils.h (update_type_inheritance_graph): Declare.
8013 (possible_polymorphic_call_target_p): Declare.
8014 (possible_polymorphic_call_target_p): New.
8015 * ipa-devirt.c: Update toplevel comments.
8016 (cached_polymorphic_call_targets): Move up.
8017 (odr_type_d): Move ID down.
8018 (polymorphic_type_binfo_p): Update comment.
8019 (odr_hasher::remove): Likewise;
8020 (get_odr_type): Set anonymous_namespace.
8021 (dump_odr_type): Dump it.
8022 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8023 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8024 (record_binfo): Add comment.
8025 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8027 (devirt_node_removal_hook): Do not iCE when cache is freed.
8028 (possible_polymorphic_call_target_p): New predicate.
8029 (update_type_inheritance_graph): New function.
8031 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8032 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8033 Sergey Lega <sergey.s.lega@intel.com>
8034 Anna Tikhonova <anna.tikhonova@intel.com>
8035 Ilya Tocar <ilya.tocar@intel.com>
8036 Andrey Turetskiy <andrey.turetskiy@intel.com>
8037 Ilya Verbin <ilya.verbin@intel.com>
8038 Kirill Yukhin <kirill.yukhin@intel.com>
8039 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8041 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8042 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8043 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8044 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8045 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8046 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8047 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8048 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8049 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8050 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8051 OPT_mavx512pf, OPT_mavx512er cases.
8052 * config/i386/constraints.md (v): New constraint.
8053 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8054 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8055 (bit_AVX512CD): New.
8056 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8057 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8058 * config/i386/i386-c.c (ix86_target_macros_internal):
8059 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8061 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8062 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8063 * config/i386/i386.c (regclass_map, dbx_register_map)
8064 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8065 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8066 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8068 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8069 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8070 -mavx512pf options. Fix formatting.
8071 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8072 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8073 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8074 -mavx512cd, -mavx512pf options.
8075 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8076 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8077 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8079 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8080 (ix86_set_reg_reg_cost): Ditto.
8081 (x86_order_regs_for_local_alloc): Ditto.
8082 (MAX_VECT_LEN): Extend to 64-byte.
8083 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8084 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8085 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8086 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8087 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8088 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8089 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8090 (SSE_REG_MODE_P): Support new modes.
8091 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8092 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8093 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8094 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8095 (REG_CLASS_CONTENTS): Add new registers.
8096 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8097 (EXT_REX_SSE_REGNO_P): New.
8098 (HI_REGISTER_NAMES): Add new registers.
8099 * config/i386/i386.md: Define constants for new registers.
8100 (mode): Add new 512-bit modes.
8101 (prefix): Support evex prefix.
8102 (isa): Support avx512f, noavx512f, fma_avx512f.
8103 (ssemodesuffix): Add new 512-bit modes.
8105 (*movxi_internal_avx512f): Ditto.
8106 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8108 (*movsi_internal): Likewise.
8109 (*movdf_internal): Likewise.
8110 (*movsf_internal): Likewise.
8111 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8112 (<code><mode>3): Likewise.
8113 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8115 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8116 with the new constraint "v".
8117 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8119 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8120 with the new constraint "v".
8121 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8122 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8123 (*<plusminus_insn><mode>3): Likewise.
8124 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8125 (*mul<mode>3): Likewise.
8126 (<sse>_vmmul<mode>3): Likewise.
8127 (<sse>_div<mode>3): Likewise.
8128 (<sse>_vmdiv<mode>3): Likewise.
8129 (<sse>_sqrt<mode>2): Likewise.
8130 (<sse>_vmsqrt<mode>2): Likewise.
8131 (*<code><mode>3_finite): Likewise.
8132 (*<code><mode>3) <smaxmin>: Likewise.
8133 (<sse>_vm<code><mode>3): Likewise.
8134 (*<code><mode>3) <any_logic>: Likewise.
8135 (*fma_fmadd_<mode>): Likewise.
8136 (*fma_fmsub_<mode>): Likewise.
8137 (*fma_fnmadd_<mode>): Likewise.
8138 (*fma_fnmsub_<mode>): Likewise.
8139 (*fma_fmaddsub_<mode>): Likewise.
8140 (*fma_fmsubadd_<mode>): Likewise.
8141 (*fmai_fmadd_<mode>): Likewise.
8142 (*fmai_fmsub_<mode>): Likewise.
8143 (*fmai_fnmadd_<mode>): Likewise.
8144 (*fmai_fnmsub_<mode>): Likewise.
8145 (sse_cvtsi2ss): Likewise.
8146 (sse_cvtsi2ssq): Likewise.
8147 (sse_cvtss2si): Likewise.
8148 (sse_cvtss2si_2): Likewise.
8149 (sse_cvtss2siq): Likewise.
8150 (sse_cvtss2siq_2): Likewise.
8151 (sse_cvttss2si): Likewise.
8152 (sse_cvtss2siq_2): Likewise.
8153 (float<sseintvecmodelower><mode>2): Likewise.
8154 (sse2_cvtsd2si_2): Likewise.
8155 (sse2_cvtsd2siq_2): Likewise.
8156 (*<plusminus_insn><mode>3): Likewise.
8157 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8158 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8159 (ashr<mode>3): Likewise.
8160 (<shift_insn><mode>3): Likewise.
8161 (avx2_<code><mode>3): Likewise.
8162 (*avx2_<code><mode>3): Likewise.
8163 (*andnot<mode>3): Likewise.
8164 (*<code><mode>3) <any_logic>: Likewise.
8165 (abs<mode>2): Likewise.
8166 (avx2_permvar<mode>): Likewise.
8167 (avx2_perm<mode>_1): Likewise.
8168 (*avx_vpermilp<mode>): Likewise.
8169 (avx_vpermilvar<mode>3): Likewise.
8170 (avx2_ashrv<mode>): Likewise.
8171 (avx2_<shift_insn>v<mode>): Likewise.
8172 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8174 * doc/rtl.texi: Document XImode.
8176 2013-08-21 Jeff Law <law@redhat.com>
8178 * tree-flow.h (register_jump_thread): Pass vector of edges
8179 instead of each important edge.
8180 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8181 thread path into a vector and pass that to register_jump_thread.
8182 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8183 passed in edge vector to the current 3-edge form.
8186 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8188 * dce.c (fini_dce): Call df_analyze again just in case
8189 delete_unmarked_insns removed anything.
8191 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8193 * reload.h (struct reg_equivs): Rename to ..
8194 (struct reg_equivs_s): .. this.
8196 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8198 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8200 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8202 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8204 2013-08-21 Jeff Law <law@redhat.com>
8206 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8207 simplify assignments too. If the RHS collapses to a singleton
8208 range, then return the value for the range.
8210 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8212 * config/i386/sse.md (V16): Rename to...
8214 (mov<mode>): Update iterator name.
8215 (*mov<mode>_internal): Ditto.
8216 (push<mode>1): Ditto.
8217 (movmisalign<mode>): Ditto.
8219 2013-08-20 Jan Hubicka <jh@suse.cz>
8222 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8223 entry for direct edges.
8224 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8226 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8228 Revert my last two changes, r201865 and r201864:
8231 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8233 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8234 instances can own GC refs.
8236 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8237 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8238 (gcc::context::gt_pch_nx): Likewise.
8239 (gcc::context::gt_pch_nx): Likewise.
8240 * ggc.h (gt_ggc_mx <T>): New.
8241 (gt_pch_nx_with_op <T>): New.
8242 (gt_pch_nx <T>): New.
8243 * passes.c (opt_pass::gt_ggc_mx): New.
8244 (opt_pass::gt_pch_nx): New.
8245 (opt_pass::gt_pch_nx_with_op): New.
8246 (pass_manager::gt_ggc_mx): New.
8247 (pass_manager::gt_pch_nx): New.
8248 (pass_manager::gt_pch_nx_with_op): New.
8249 (pass_manager::operator new): Use
8250 ggc_internal_cleared_alloc_stat rather than xcalloc.
8251 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8252 (pass_manager::gt_ggc_mx): New.
8253 (pass_manager::gt_pch_nx): New.
8254 (pass_manager::gt_pch_nx_with_op): New.
8255 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8256 (opt_pass::operator new): New.
8257 (opt_pass::gt_ggc_mx): New.
8258 (opt_pass::gt_pch_nx): New.
8259 (opt_pass::gt_pch_nx_with_op): New.
8262 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8264 * Makefile.in (GTFILES): Add context.h.
8265 * context.c (gcc::context::operator new): New.
8266 (gcc::context::gt_ggc_mx): New.
8267 (gcc::context::gt_pch_nx): New.
8268 (gcc::context::gt_pch_nx): New.
8269 * context.h (gcc::context): Add GTY((user)) marking.
8270 (gcc::context::operator new): New.
8271 (gcc::context::gt_ggc_mx): New.
8272 (gcc::context::gt_pch_nx): New.
8273 (gcc::context::gt_pch_nx): New.
8274 (g): Add GTY marking.
8275 (gt_ggc_mx (gcc::context *)): New.
8276 (gt_pch_nx (gcc::context *)): New.
8277 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8278 void *cookie)): New.
8279 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8281 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8283 * dce.c (fini_dce): Call df_analyze again just in case
8284 delete_unmarked_insns removed anything.
8286 2013-08-20 Teresa Johnson <tejohnson@google.com>
8288 PR rtl-optimizations/57451
8289 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8290 from crossing split section boundaries.
8292 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8294 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8295 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8296 with MULTLIB_DEFAULTS.
8298 2013-08-20 Nick Clifton <nickc@redhat.com>
8300 * target.def (narrow_volatile_bitfield): Note that the default
8301 value is false, not !TARGET_STRICT_ALIGN.
8302 * doc/tm.texi: Regenerate.
8304 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8306 Fix LIB_SPEC for systems without libpthread.
8308 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8309 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8311 * config/i386/linux-common.h: Likewise.
8312 * config/mips/linux-common.h: Likewise.
8314 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8316 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8319 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8321 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8322 instances can own GC refs.
8324 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8325 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8326 (gcc::context::gt_pch_nx): Likewise.
8327 (gcc::context::gt_pch_nx): Likewise.
8328 * ggc.h (gt_ggc_mx <T>): New.
8329 (gt_pch_nx_with_op <T>): New.
8330 (gt_pch_nx <T>): New.
8331 * passes.c (opt_pass::gt_ggc_mx): New.
8332 (opt_pass::gt_pch_nx): New.
8333 (opt_pass::gt_pch_nx_with_op): New.
8334 (pass_manager::gt_ggc_mx): New.
8335 (pass_manager::gt_pch_nx): New.
8336 (pass_manager::gt_pch_nx_with_op): New.
8337 (pass_manager::operator new): Use
8338 ggc_internal_cleared_alloc_stat rather than xcalloc.
8339 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8340 (pass_manager::gt_ggc_mx): New.
8341 (pass_manager::gt_pch_nx): New.
8342 (pass_manager::gt_pch_nx_with_op): New.
8343 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8344 (opt_pass::operator new): New.
8345 (opt_pass::gt_ggc_mx): New.
8346 (opt_pass::gt_pch_nx): New.
8347 (opt_pass::gt_pch_nx_with_op): New.
8349 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8351 * Makefile.in (GTFILES): Add context.h.
8352 * context.c (gcc::context::operator new): New.
8353 (gcc::context::gt_ggc_mx): New.
8354 (gcc::context::gt_pch_nx): New.
8355 (gcc::context::gt_pch_nx): New.
8356 * context.h (gcc::context): Add GTY((user)) marking.
8357 (gcc::context::operator new): New.
8358 (gcc::context::gt_ggc_mx): New.
8359 (gcc::context::gt_pch_nx): New.
8360 (gcc::context::gt_pch_nx): New.
8361 (g): Add GTY marking.
8362 (gt_ggc_mx (gcc::context *)): New.
8363 (gt_pch_nx (gcc::context *)): New.
8364 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8365 void *cookie)): New.
8366 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8368 2013-08-20 Alan Modra <amodra@gmail.com>
8371 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8372 (rs6000_emit_epilogue): Likewise.
8374 2013-08-19 Dehao Chen <dehao@google.com>
8376 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8378 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8379 Jakub Jelinek <jakub@redhat.com>
8381 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8382 (BUILT_IN_FABSD64): Likewise.
8383 (BUILT_IN_FABSD128): Likewise.
8384 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8385 (fold_builtin_1): Likewise.
8386 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8387 destination and source operands.
8388 (*abstd2_fpr): Likewise.
8389 (*nabstd2_fpr): Likewise.
8391 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8393 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8396 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8398 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8401 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8403 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8404 * builtins.c (default_libc_has_function): New.
8405 (gnu_libc_has_function): Ditto.
8406 (no_c99_libc_has_function): Ditto.
8407 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8408 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8409 (fold_builtin_sincos): Likewise.
8410 (fold_builtin_cexp): Likewise.
8411 * builtins.def (DEF_C94_BUILTIN): Likewise.
8412 (DEF_C99_BUILTIN): Likewise.
8413 (DEF_C99_C90RES_BUILTIN): Likewise.
8414 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8415 definitions to using this define.
8416 * config/darwin-protos.h (darwin_libc_has_function): New.
8417 * config/darwin.c (darwin_libc_has_function): Ditto.
8418 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8419 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8420 * config/darwin.h: Ditto.
8421 * config/elfos.h: Ditto.
8422 * config/freebsd.h: Ditto.
8423 * config/i386/cygming.h: Ditto.
8424 * config/i386/djgpp.h: Ditto.
8425 * config/i386/i386-interix.h: Ditto.
8426 * config/microblaze/microblaze.h: Ditto.
8427 * config/mmix/mmix.h: Ditto.
8428 * config/gnu-user.h: Ditto.
8429 * config/ia64/hpux.h: Ditto.
8430 * config/pa/pa-hpux.h: Ditto.
8431 * config/pdp11/pdp11.h: Ditto.
8432 * config/picochip/picochip.h: Ditto.
8433 * config/linux.h: Ditto.
8434 * config/netbsd.h: Ditto.
8435 * config/openbsd.h: Ditto.
8436 * config/rs6000/aix43.h: Ditto.
8437 * config/rs6000/aix51.h: Ditto.
8438 * config/rs6000/aix52.h: Ditto.
8439 * config/rs6000/aix53.h: Ditto.
8440 * config/rs6000/aix61.h: Ditto.
8441 * config/rs6000/darwin.h: Ditto.
8442 * config/rs6000/linux.h: Ditto.
8443 * config/rs6000/linux64.h: Ditto.
8444 * config/s390/tpf.h: Ditto.
8445 * config/sol2-10.h: Ditto.
8446 * config/sol2.h: Ditto.
8447 * config/vms/vms.h: Ditto.
8448 * config/vxworks.h: Ditto.
8449 * config/linux-android.c (linux_android_libc_has_function):
8450 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8451 * config/linux-protos.h (linux_android_libc_has_function):
8453 * config/i386/i386.c (ix86_libc_has_function): New.
8454 * config/i386/i386-protos.h
8455 (ix86_libc_has_function): New declaration.
8456 * config/i386/i386.md
8457 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8458 ("isinf<mode>2): Likewise.
8459 * convert.c (convert_to_integer): Using new target hook
8460 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8461 TARGET_C99_FUNCTIONS.
8462 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8463 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8464 * coretypes.h (function_class): New enum for different
8465 classes of functions.
8466 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8467 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8468 (TARGET_HAS_SINCOS): Likewise.
8469 (TARGET_LIBC_HAS_FUNCTION): New.
8470 * doc/tm.texi: Regenerated.
8471 * targhooks.h (default_libc_has_function): New declaration.
8472 (no_c99_libc_has_function): Ditto.
8473 (gnu_libc_has_function): Ditto.
8474 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8475 and TARGET_HAS_SINCOS.
8477 2013-08-18 Jan Hubicka <jh@suse.cz>
8479 * Makeifle-in (ipa-devirt.o): New.
8480 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8481 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8482 (analyze_functions): Look into possible targets of polymorphic call.
8483 * dumpfile.c (dump_files): Add type-inheritance dump.
8484 * dumpfile.h (TDI_inheritance): New.
8485 * ipa-devirt.c: New file.
8486 * ipa-utils.h (odr_type_d): Forward declare.
8487 (odr_type): New type.
8488 (build_type_inheritance_graph): Declare.
8489 (possible_polymorphic_call_targets): Declare and introduce inline
8490 variant when only edge is pased.
8491 (dump_possible_polymorphic_call_targets): Likewise.
8492 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8493 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8494 (types_same_for_odr): ... here.
8495 * tree.h (type_in_anonymous_namespace_p): Declare.
8497 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8499 PR tree-optimization/58006
8500 * tree-parloops.c (take_address_of): Don't ICE if get_name
8502 (eliminate_local_variables_stmt): Remove clobber stmts.
8504 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8506 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8507 error message is issued for an alias to undefined symbol.
8509 2013-08-18 Jan Hubicka <jh@suse.cz>
8511 * cgraph.c (cgraph_create_indirect_edge): Discover
8512 polymorphic calls and record basic info into indirect_info.
8513 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8514 devirtualization, ignore objc function calls.
8515 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8516 call with no parm index info.
8517 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8518 * tree.c (virtual_method_call_p): New function.
8519 * tree.h (virtual_method_call_p): Declare.
8521 2013-08-16 Jan Hubicka <jh@suse.cz>
8524 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8526 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8528 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8531 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8533 * gengtype.c (type_for_name): Add special-case support for
8534 locating types within the "gcc::" namespace.
8535 (open_base_files): Emit a "using namespace gcc" directive.
8537 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8540 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8541 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8543 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8544 array instead of each individual operand as a separate argument.
8545 (emit_fusion_gpr_load): Likewise.
8546 (expand_fusion_gpr_load): Add new function declaration.
8548 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8549 signature to have the operands passed as an array, instead of as
8550 separate arguments. Allow ZERO_EXTEND to be in the memory
8551 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8552 depend on the register live/dead flags when peepholes are run.
8553 (expand_fusion_gpr_load): New function to be called from the
8554 peephole2 pass, to change the register that addis sets to be the
8556 (emit_fusion_gpr_load): Change the calling signature to have the
8557 operands passed as an array, instead of as separate arguments.
8558 Allow ZERO_EXTEND to be in the memory address, and also
8559 SIGN_EXTEND if -mpower8-fusion-sign.
8561 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8563 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8564 adjust the register addis loads up in the peephole2 pass. Do not
8565 depend on the register live/dead state when the peephole pass is done.
8567 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8569 * gengtype.c (create_user_defined_type): Ensure that the kind
8570 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8571 declaration is seen before the GTY((user)) marking.
8573 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8576 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8578 2013-08-16 Jan Hubicka <jh@suse.cz>
8580 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8581 arugment expected_type.
8582 (gimple_fold_call): Use it.
8583 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8584 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8585 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8586 (try_make_edge_direct_virtual_call): Likewise.
8587 * tree.c (obj_type_ref_class): New.
8588 * tree.h (obj_type_ref_class): Use it.
8590 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8592 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8593 (rtl_slim_pp): Likewise.
8594 (init_rtl_slim_pretty_print): Likewise.
8595 (dump_value_slim): Don't call it. Use local pretty printer.
8596 (dump_insn_slim): Likewise.
8597 (dump_rtl_slim): Likewise.
8598 (str_pattern_slim): Likewise.
8599 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8602 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8604 PR tree-optimization/58164
8605 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8606 walk gimple_goto_dest of GIMPLE_GOTO.
8608 PR tree-optimization/58165
8609 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8610 bi_call must be the last stmt in a bb, don't split_block, instead
8611 use fallthru edge from it and give up if there is none.
8612 Release conds vector when returning early.
8614 2013-08-14 Xinliang David Li <davidxl@google.com>
8616 * config/i386/i386.c (ix86_option_override_internal):
8617 Remove unused variable and field.
8619 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8622 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8623 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8624 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8625 and Linux, correct BLKmode alignment when 128-bit alignment is
8626 required and compatibility flag is not set.
8627 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8628 for zero-size arguments when compatibility flag is not set.
8630 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8632 PR tree-optimization/58145
8633 * tree-sra.c (build_ref_for_offset): If prev_base has
8634 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8636 2013-08-14 Xinliang David Li <davidxl@google.com>
8638 * config/i386/i386.c (ix86_option_override_internal):
8639 Fix uninitialized variable error.
8641 2013-08-14 Xinliang David Li <davidxl@google.com>
8643 * config/i386/i386.opt: Define two new options.
8644 * config/i386/x86-tune.def: Add arch selector field in macros.
8645 * config/i386/i386.h: Adjust macro definition.
8646 * config/i386/i386.c (ix86_option_override_internal):
8648 (parse_mtune_ctrl_str): New function.
8649 (set_ix86_tune_features): New function.
8650 (ix86_function_specific_restore): Call the new helper function.
8652 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8654 PR rtl-optimization/57662
8655 * sel-sched.c (code_motion_process_successors): When the current insn
8656 is removed after the recursive traversal, break from the loop.
8657 Add comments and debug printouts.
8659 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8660 Alexandre Oliva <aoliva@redhat.com>
8663 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8664 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8665 there also UNSPEC_PLTOFF.
8667 2013-08-14 Marek Polacek <polacek@redhat.com>
8669 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8670 NUM_CONDITIONS bit positions.
8672 2013-08-13 Cary Coutant <ccoutant@google.com>
8674 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8675 (attr_checksum): Hash vector contents instead of pointer.
8676 (attr_checksum_ordered): Likewise.
8678 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8680 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8681 when Pmode != word_mode. Add length_address attribute.
8682 (sse3_monitor_<mode>): Merge from sse3_monitor and
8683 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8684 Pmode != word_mode. Update insn length attribute.
8685 * config/i386/i386.c (ix86_option_override_internal): Update
8686 ix86_gen_monitor selection for merged sse3_monitor insn.
8688 2013-08-13 Julian Brown <julian@codesourcery.com>
8690 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8691 perform invalid legitimization on greater-than-word-size modes for
8694 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8696 * ira.c (setup_class_translate_array): Use aclass instead of cl
8697 for classes not fully covered by allocno classes.
8699 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8701 PR tree-optimization/57661
8702 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8703 * tree-inline.c (tree_function_versioning): Initialize it.
8704 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8705 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8706 that is not being copied.
8709 * asan.c (instrument_strlen_call): Fix typo in comment.
8710 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8712 2013-08-13 Steve Ellcey <sellcey@mips.com>
8714 * config/mips/mips.md (prefetch): Use lw instead of ld on
8715 loongson in 32bit mode.
8717 2013-08-13 Nick Clifton <nickc@redhat.com>
8719 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8721 2013-08-13 Jan Hubicka <jh@suse.cz>
8723 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8724 introduced edge; fix typo in sanity check.
8725 (cgraph_resolve_speculation): Export; improve diagnostic.
8726 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8727 speculation at type mismatch.
8728 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8729 (cgraph_resolve_speculation): Declare.
8730 (symtab_can_be_discarded): New function.
8731 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8732 * ipa-inline-transform.c (speculation_removed): New global var.
8733 (clone_inlined_nodes): See if speculation can be removed.
8734 (inline_call): If speculations was removed, we growths may not match.
8735 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8736 (speculation_useful_p): New function.
8737 (resolve_noninline_speculation): New function.
8738 (inline_small_functions): Resolve useless speculations.
8739 * ipa-inline.h (speculation_useful_p): Declare
8740 * ipa.c (can_replace_by_local_alias): Simplify.
8741 (ipa_profile): Produce speculative calls in non-lto, too;
8742 add simple cost model; produce local aliases.
8744 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8746 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8749 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8751 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8753 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8755 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8756 iterator instead of X87MODEF.
8758 2013-08-12 Perez Read <netfirewall@gmail.com>
8761 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8762 operand 0 for intel asm alternative.
8763 (*movabs<mode>_2): Ditto for operand 1.
8765 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8767 * config/aarch64/arm_none.h
8768 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8770 2013-08-12 Nick Clifton <nickc@redhat.com>
8772 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8774 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8776 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8777 expand for QI/HImode operand to produce more effictive code for
8778 unsigned char(short) --> float(double) conversion.
8780 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8782 * doc/invoke.texi: Mention that -ftls-model does not force the final
8785 2013-08-12 Marek Polacek <polacek@redhat.com>
8786 Marc Glisse <marc.glisse@inria.fr>
8788 PR tree-optimization/57980
8789 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8790 when creating -1 constant.
8792 2013-08-10 Jan Hubicka <jh@suse.cz>
8794 Workaround binutils PR14342.
8795 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8796 (gimple_init_edge_profiler): Likewise.
8797 (gimple_gen_ic_func_profiler): Likewise.
8799 2013-08-09 Jan Hubicka <jh@suse.cz>
8801 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8803 2013-08-09 Xinliang David Li <davidxl@google.com>
8805 * config/i386/stringop.def: New file.
8806 * config/i386/stringop.opt: New file.
8807 * config/i386/i386-opts.h: Include stringopt.def.
8808 * config/i386/i386.opt: Include stringopt.opt.
8809 * config/i386/i386.c (ix86_option_override_internal):
8810 Override default size based stringop inline strategies with options.
8811 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8814 2013-08-09 Jan Hubicka <jh@suse.cz>
8816 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
8818 2013-08-09 Jan Hubicka <jh@suse.cz>
8820 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8822 (dump_cgraph_node): Dump profile-id.
8823 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8824 and common_target_probability.
8825 * lto-cgraph.c (lto_output_edge): Stream common targets.
8826 (lto_output_node): Stream profile ids.
8827 (input_node): Stream profile ids.
8828 (input_edge): Stream common targets.
8829 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8830 * ipa.c: Include value-prof.h
8831 (ipa_profile_generate_summary): Turn indirect call statement histograms
8832 into common targets.
8833 (ipa_profile): Turn common targets into speculative edges.
8835 2013-08-09 Jan Hubicka <jh@suse.cz>
8837 * cgraph.h (cgraph_node): Add profile_id.
8838 * value-prof.c (cgraph_node_map): Turn into pointer_map.
8839 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
8840 (del_node_map): Update.
8841 (find_func_by_funcdef_no): Replace by ...
8842 (find_func_by_profile_id): ... this one.
8843 (gimple_ic_transform): Do not remove useful histograms when
8844 speculation is not done; dump info when indirect call removal
8846 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8847 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8848 (__gcov_indirect_call_profiler_v2): .. this one.
8849 * profile.h (init_node_map): Update.
8850 * coverage.c (coverage_compute_profile_id): New function.
8851 * coverage.h (coverage_compute_profile_id): Declare.
8852 * tree-profile.c (init_ic_make_global_vars): Make
8853 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8854 (gimple_init_edge_profiler): Update prototype of
8855 __gcov_indirect_call_profiler.
8856 (gimple_gen_ic_func_profiler): Simplify.
8857 (tree_profiling): Use init_node_map
8859 2013-08-09 Jan Hubicka <jh@suse.cz>
8861 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8862 non-speculative refs.
8863 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8864 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8865 (cgraph_set_call_stmt): Likewise.
8866 (cgraph_create_edge_1): Fix release checking compilatoin;
8868 (cgraph_free_edge): Free indirect info.
8869 (cgraph_turn_edge_to_speculative): New function.
8870 (cgraph_speculative_call_info): New function.
8871 (cgraph_make_edge_direct): Return direct edge; handle speculation.
8872 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
8873 (dump_cgraph_node): Dump speculation.
8874 (verify_edge_count_and_frequency): Accept speculative edges.
8875 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8876 (verify_cgraph_node): Handle speculation.
8877 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8878 (cgraph_set_call_stmt): Update prototype.
8879 (cgraph_make_edge_direct): Update prototype.
8880 (cgraph_speculative_call_info): Declare.
8881 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8882 to change; update call of ipa_find_references.
8883 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8884 lto_stmt_uid and speculative flags.
8885 (ipa_dump_references): Dump speculation.
8886 (ipa_clone_references): Clone speculative flag.
8887 (ipa_clone_referring): Likewise.
8888 (ipa_clone_ref): New function.
8889 (ipa_find_reference): Look into lto_stmt_uids
8890 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8891 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8892 (ipa_find_reference): Update declaration.
8893 (ipa_clone_ref): Declare.
8894 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8895 stream speculative flag.
8896 (lto_output_ref): Stream statements uids and speculation.
8897 (input_ref): Likewise.
8898 (input_edge): Stream speuclation.
8899 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8900 (cgraph_set_call_stmt_including_clones): Handle speculation.
8901 * ipa-inline.c (heap_edge_removal_hook): New function.
8902 (inline_small_functions): Register it.
8903 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8904 also initialize refs.
8905 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8907 (try_make_edge_direct_simple_call): Likewise.
8908 (try_make_edge_direct_simple_call): Likewise.
8909 (update_indirect_edges_after_inlining): Likewise.
8910 (remove_described_reference): Look proper lto_stmt_uid.
8911 (propagate_controlled_uses): Likewise.
8912 (propagate_controlled_uses): Liekwise.
8913 * tree-inline.c (copy_bb): Copy speculative edges.
8914 (redirect_all_calls): New function.
8915 (copy_cfg_body): Do redirection after loop info is updated.
8916 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
8918 2013-08-09 Jan Hubicka <jh@suse.cz>
8920 * lto-streamer-out.c (output_function): Renumber PHIs.
8921 * lto-streamer-in.c (input_function): Likewise.
8923 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8925 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8926 (get_lane_unsigned): Likewise.
8927 (dup_lane_scalar): Likewise.
8928 (get_lane): enable for VALL.
8929 * config/aarch64/aarch64-simd.md
8930 (aarch64_dup_lane_scalar<mode>): Remove.
8931 (aarch64_get_lane_signed<mode>): Likewise.
8932 (aarch64_get_lane_unsigned<mode>): Likewise.
8933 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8934 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8935 (aarch64_get_lane<mode>): Enable for all vector modes.
8936 (aarch64_get_lanedi): Remove misleading constraints.
8937 * config/aarch64/arm_neon.h
8938 (__aarch64_vget_lane_any): Define.
8939 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
8940 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
8941 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
8942 * config/aarch64/iterators.md (VDQQH): New.
8946 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
8948 * configure.ac: Add GAS check for LEON instructions on SPARC.
8949 * configure: Regenerate.
8950 * config.in: Likewise.
8951 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
8953 * config/sparc/sparc.opt (LEON, LEON3): New masks.
8954 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
8956 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
8957 (AS_LEON_FLAG): New macro.
8958 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
8959 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
8960 Deal with LEON and LEON3 for the memory model.
8961 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
8962 (atomic_compare_and_swap<mode>_1): Likewise.
8963 (*atomic_compare_and_swap<mode>_1): Likewise.
8965 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8967 * config/arm/neon.md (vcond): Fix floating-point vector
8968 comparisons against 0.
8970 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
8972 * lra-constraints.c (emit_spill_move): Remove assert.
8973 (process_alt_operands): Add more debugging
8974 output. Increase reject for spilling into memory. Decrease
8975 reject for reloading scratch.
8976 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
8978 2013-08-08 Steve Ellcey <sellcey@mips.com>
8980 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
8981 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
8982 micromips incompatible. Add nan2008.
8983 (MULTILIB_DIRNAMES): Add nan2008.
8984 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8985 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
8986 and micromips incompatible. Add nan2008.
8987 (MULTILIB_DIRNAMES): Add nan2008.
8988 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8990 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
8992 PR rtl-optimization/58079
8993 * combine.c (combine_simplify_rtx): Avoid using SUBST if
8994 simplify_comparison has widened a comparison with an integer.
8996 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8998 * config/arm/neon.md (movmisalign<mode>): Disable when we
8999 don't allow unaligned accesses.
9000 (*movmisalign<mode>_neon_store): Likewise.
9001 (*movmisalign<mode>_neon_load): Likewise.
9002 (*movmisalign<mode>_neon_store): Likewise.
9003 (*movmisalign<mode>_neon_load): Likewise.
9005 2013-08-08 Jan Hubicka <jh@suse.cz>
9007 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9008 (make_pass_rebuild_cgraph_edges): Also clear references.
9009 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9010 * ipa-inline-transform.c (inline_transform): Remove all references
9012 * cgraphunit.c (expand_function): Remove all references after
9014 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9015 (ipa_find_reference): Rewrite to iterator.
9016 (remove_stmt_references): Likewise.
9017 (ipa_clear_stmts_in_references): New function.
9018 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9019 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9021 * ipa-split.c (split_function): Remove references in split function.
9023 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9026 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9027 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9029 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9032 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9033 suggested mode for the assignment isn't compatible with the
9036 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9039 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9041 2013-08-07 Xinliang David Li <davidxl@google.com>
9043 * config/i386/i386.opt: New option -mtune-ctrl=.
9044 * config/i386/x86-tune.def: New file.
9045 * config/i386/i386.h: include x86-tune.def.
9046 * config/i386/i386.c (ix86_option_override_internal):
9047 Parsing -mtune-ctrl= option and set tune features.
9049 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9052 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9053 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9056 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9058 * rtl.h (update_alignments): Declare.
9059 * final.c (grow_label_align): New function extracted from...
9060 (shorten_branches): ...here. Call it.
9061 (update_alignments): New function.
9062 * reorg.c (sibling_labels): New variable.
9063 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9064 the new label along with it onto the sibling_labels vector.
9065 (fill_simple_delay_slots): Adjust call to get_label_before.
9066 (fill_slots_from_thread): Likewise.
9067 (relax_delay_slots): Likewise.
9068 (make_return_insns): Likewise.
9069 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9071 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9073 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9074 document its semantics.
9075 (diagnostic_report_diagnostic): Adjust accordingly.
9077 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9079 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9080 (sparc_option_override): ...and port to new C++ pass API.
9081 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9083 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9085 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9087 2013-08-06 Caroline Tice <cmtice@google.com>
9089 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9090 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9091 * tree-pass.h: Add pass_vtable_verify.
9092 * varasm.c (assemble_variable): Add code to properly set the comdat
9093 section and name for the .vtable_map_vars section.
9094 (assemble_vtyv_preinit_initializer): New function.
9095 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9097 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9098 * vtable-verify.c: New file.
9099 * vtable-verify.h: New file.
9100 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9101 initialiation levels.
9102 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9103 * passes.def: Insert pass_vtable_verify.
9104 * aclocal.m4: Reorder includes.
9105 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9106 -fvtv-counts options.
9107 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9108 as appropriate, if -fvtable-verify=... is used.
9109 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9110 -fvtable-verify=... is used.
9111 * Makefile.in (OBJS): Add vtable-verify.o to list.
9112 (vtable-verify.o): Add new build rule.
9113 (GTFILES): Add vtable-verify.c to list.
9114 * common.opt (fvtable-verify=): New flag.
9115 (vtv_priority): Values for fvtable-verify= flag.
9116 (fvtv-counts): New flag.
9117 (fvtv-debug): New flag.
9118 * tree.h (save_vtable_map_decl): New extern function decl.
9120 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9122 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9123 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9124 (pass_data_rl78_devirt): ...new pass_data instance and...
9125 (make_pass_rl78_devirt): ...new function.
9126 (rl78_asm_file_start): Port pass registration to new C++ API.
9128 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9130 * coretypes.h (rtl_opt_pass): Add.
9131 (gcc::context): Add.
9132 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9133 (epiphany_init): Port to new C++ pass API.
9134 (epiphany_optimize_mode_switching): Likewise.
9135 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9136 (pass_manager::get_pass_mode_switching): New.
9137 (pass_manager::get_pass_peephole2): New.
9138 * mode-switching.c (pass_mode_switching): Add clone method.
9139 * recog.c (pass_peephole2): Add clone method.
9140 (pass_split_all_insns): Add clone method.
9142 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9144 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9145 (mips_option_override): ...here, porting to new C++ API for passes.
9147 2013-08-06 Jan Hubicka <jh@suse.cz>
9149 * cgraph.c (cgraph_get_body): New function based on lto.c
9151 * cgraph.h (cgraph_get_body): Declare.
9152 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9154 * cgraphunit.c (expand_function): Get body prior expanding.
9155 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9156 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9158 * passes.c (do_per_function_toporder): Get body.
9159 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9160 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9162 2013-08-06 Martin Jambor <mjambor@suse.cz>
9165 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9166 is not re-finalized. Rename second parameter to no_collect.
9168 2013-08-06 Martin Jambor <mjambor@suse.cz>
9171 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9172 MEM_REF has proper alignment information.
9174 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9177 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9179 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9180 rtx (*) (rtx, ...) with insn_gen_fn.
9181 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9182 insn_gen_fn::stored_funcptr. Add initializer braces.
9184 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9186 Rewrite how instances of passes are cloned to remove assumptions
9187 about their sizes (thus allowing pass subclasses to have
9188 additional data fields, albeit non-GC-managed ones at this point).
9190 * passes.c (make_pass_instance): Now that passes have clone
9191 methods, rewrite this function to eliminate XNEW and memcpy
9192 calls that used hardcoded sizes. Since this function no longer
9193 creates pass instances, rename it to...
9194 (add_pass_instance): ...this. Document the old way that passes were
9195 numbered and flagged, and rework this function to continue using it.
9196 (next_pass_1): Add an initial_pass argument for use by
9198 (position_pass): When adding multiple instances of a pass, use
9199 the pass's clone method, rather than relying on the XNEW/memcpy
9200 within the former make_pass_instance (now add_pass_instance).
9201 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9202 the initial instance of the current pass within the pass manager.
9204 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9206 This is the automated part of the conversion of passes from C
9207 structs to C++ classes.
9209 Patch autogenerated by refactor_passes.py from
9210 https://github.com/davidmalcolm/gcc-refactoring-scripts
9211 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9213 * asan.c (pass_asan): Convert from a global struct to a subclass of
9214 gimple_opt_pass along with...
9215 (pass_data_asan): ...new pass_data instance and...
9216 (make_pass_asan): ...new function.
9217 (pass_asan_O0): Convert from a global struct to a subclass of
9218 gimple_opt_pass along with...
9219 (pass_data_asan_O0): ...new pass_data instance and...
9220 (make_pass_asan_O0): ...new function.
9221 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9222 subclass of rtl_opt_pass along with...
9223 (pass_data_inc_dec): ...new pass_data instance and...
9224 (make_pass_inc_dec): ...new function.
9225 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9226 a subclass of rtl_opt_pass along with...
9227 (pass_data_reorder_blocks): ...new pass_data instance and...
9228 (make_pass_reorder_blocks): ...new function.
9229 (pass_duplicate_computed_gotos): Convert from a global struct to a
9230 subclass of rtl_opt_pass along with...
9231 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9232 (make_pass_duplicate_computed_gotos): ...new function.
9233 (pass_partition_blocks): Convert from a global struct to a subclass of
9234 rtl_opt_pass along with...
9235 (pass_data_partition_blocks): ...new pass_data instance and...
9236 (make_pass_partition_blocks): ...new function.
9237 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9238 struct to a subclass of rtl_opt_pass along with...
9239 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9241 (make_pass_branch_target_load_optimize1): ...new function.
9242 (pass_branch_target_load_optimize2): Convert from a global struct to a
9243 subclass of rtl_opt_pass along with...
9244 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9246 (make_pass_branch_target_load_optimize2): ...new function.
9247 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9248 of rtl_opt_pass along with...
9249 (pass_data_jump): ...new pass_data instance and...
9250 (make_pass_jump): ...new function.
9251 (pass_jump2): Convert from a global struct to a subclass of
9252 rtl_opt_pass along with...
9253 (pass_data_jump2): ...new pass_data instance and...
9254 (make_pass_jump2): ...new function.
9255 * cfgexpand.c (pass_expand): Convert from a global struct to a
9256 subclass of rtl_opt_pass along with...
9257 (pass_data_expand): ...new pass_data instance and...
9258 (make_pass_expand): ...new function.
9259 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9260 of rtl_opt_pass along with...
9261 (pass_data_free_cfg): ...new pass_data instance and...
9262 (make_pass_free_cfg): ...new function.
9263 (pass_into_cfg_layout_mode): Convert from a global struct to a
9264 subclass of rtl_opt_pass along with...
9265 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9266 (make_pass_into_cfg_layout_mode): ...new function.
9267 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9268 subclass of rtl_opt_pass along with...
9269 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9270 (make_pass_outof_cfg_layout_mode): ...new function.
9271 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9272 struct to a subclass of gimple_opt_pass along with...
9273 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9274 (make_pass_build_cgraph_edges): ...new function.
9275 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9276 subclass of gimple_opt_pass along with...
9277 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9278 (make_pass_rebuild_cgraph_edges): ...new function.
9279 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9280 subclass of gimple_opt_pass along with...
9281 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9283 (make_pass_remove_cgraph_callee_edges): ...new function.
9284 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9285 struct to a subclass of rtl_opt_pass along with...
9286 (pass_data_stack_adjustments): ...new pass_data instance and...
9287 (make_pass_stack_adjustments): ...new function.
9288 * combine.c (pass_combine): Convert from a global struct to a subclass
9289 of rtl_opt_pass along with...
9290 (pass_data_combine): ...new pass_data instance and...
9291 (make_pass_combine): ...new function.
9292 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9293 global struct to a subclass of rtl_opt_pass along with...
9294 (pass_data_compare_elim_after_reload): ...new pass_data instance
9296 (make_pass_compare_elim_after_reload): ...new function.
9297 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9298 of rtl_opt_pass along with...
9299 (pass_data_rtl_cprop): ...new pass_data instance and...
9300 (make_pass_rtl_cprop): ...new function.
9301 * cse.c (pass_cse): Convert from a global struct to a subclass of
9302 rtl_opt_pass along with...
9303 (pass_data_cse): ...new pass_data instance and...
9304 (make_pass_cse): ...new function.
9305 (pass_cse2): Convert from a global struct to a subclass of
9306 rtl_opt_pass along with...
9307 (pass_data_cse2): ...new pass_data instance and...
9308 (make_pass_cse2): ...new function.
9309 (pass_cse_after_global_opts): Convert from a global struct to a
9310 subclass of rtl_opt_pass along with...
9311 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9312 (make_pass_cse_after_global_opts): ...new function.
9313 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9314 of rtl_opt_pass along with...
9315 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9316 (make_pass_ud_rtl_dce): ...new function.
9317 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9318 rtl_opt_pass along with...
9319 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9320 (make_pass_fast_rtl_dce): ...new function.
9321 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9322 a subclass of rtl_opt_pass along with...
9323 (pass_data_df_initialize_opt): ...new pass_data instance and...
9324 (make_pass_df_initialize_opt): ...new function.
9325 (pass_df_initialize_no_opt): Convert from a global struct to a
9326 subclass of rtl_opt_pass along with...
9327 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9328 (make_pass_df_initialize_no_opt): ...new function.
9329 (pass_df_finish): Convert from a global struct to a subclass of
9330 rtl_opt_pass along with...
9331 (pass_data_df_finish): ...new pass_data instance and...
9332 (make_pass_df_finish): ...new function.
9333 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9334 rtl_opt_pass along with...
9335 (pass_data_rtl_dse1): ...new pass_data instance and...
9336 (make_pass_rtl_dse1): ...new function.
9337 (pass_rtl_dse2): Convert from a global struct to a subclass of
9338 rtl_opt_pass along with...
9339 (pass_data_rtl_dse2): ...new pass_data instance and...
9340 (make_pass_rtl_dse2): ...new function.
9341 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9342 subclass of rtl_opt_pass along with...
9343 (pass_data_dwarf2_frame): ...new pass_data instance and...
9344 (make_pass_dwarf2_frame): ...new function.
9345 * except.c (pass_set_nothrow_function_flags): Convert from a global
9346 struct to a subclass of rtl_opt_pass along with...
9347 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9349 (make_pass_set_nothrow_function_flags): ...new function.
9350 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9351 subclass of rtl_opt_pass along with...
9352 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9354 (make_pass_convert_to_eh_region_ranges): ...new function.
9355 * final.c (pass_compute_alignments): Convert from a global struct to a
9356 subclass of rtl_opt_pass along with...
9357 (pass_data_compute_alignments): ...new pass_data instance and...
9358 (make_pass_compute_alignments): ...new function.
9359 (pass_final): Convert from a global struct to a subclass of
9360 rtl_opt_pass along with...
9361 (pass_data_final): ...new pass_data instance and...
9362 (make_pass_final): ...new function.
9363 (pass_shorten_branches): Convert from a global struct to a subclass of
9364 rtl_opt_pass along with...
9365 (pass_data_shorten_branches): ...new pass_data instance and...
9366 (make_pass_shorten_branches): ...new function.
9367 (pass_clean_state): Convert from a global struct to a subclass of
9368 rtl_opt_pass along with...
9369 (pass_data_clean_state): ...new pass_data instance and...
9370 (make_pass_clean_state): ...new function.
9371 * function.c (pass_instantiate_virtual_regs): Convert from a global
9372 struct to a subclass of rtl_opt_pass along with...
9373 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9374 (make_pass_instantiate_virtual_regs): ...new function.
9375 (pass_leaf_regs): Convert from a global struct to a subclass of
9376 rtl_opt_pass along with...
9377 (pass_data_leaf_regs): ...new pass_data instance and...
9378 (make_pass_leaf_regs): ...new function.
9379 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9380 subclass of rtl_opt_pass along with...
9381 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9383 (make_pass_thread_prologue_and_epilogue): ...new function.
9384 (pass_match_asm_constraints): Convert from a global struct to a
9385 subclass of rtl_opt_pass along with...
9386 (pass_data_match_asm_constraints): ...new pass_data instance and...
9387 (make_pass_match_asm_constraints): ...new function.
9388 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9389 subclass of rtl_opt_pass along with...
9390 (pass_data_rtl_fwprop): ...new pass_data instance and...
9391 (make_pass_rtl_fwprop): ...new function.
9392 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9393 rtl_opt_pass along with...
9394 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9395 (make_pass_rtl_fwprop_addr): ...new function.
9396 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9397 rtl_opt_pass along with...
9398 (pass_data_rtl_pre): ...new pass_data instance and...
9399 (make_pass_rtl_pre): ...new function.
9400 (pass_rtl_hoist): Convert from a global struct to a subclass of
9401 rtl_opt_pass along with...
9402 (pass_data_rtl_hoist): ...new pass_data instance and...
9403 (make_pass_rtl_hoist): ...new function.
9404 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9405 subclass of gimple_opt_pass along with...
9406 (pass_data_lower_cf): ...new pass_data instance and...
9407 (make_pass_lower_cf): ...new function.
9408 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9409 from a global struct to a subclass of gimple_opt_pass along with...
9410 (pass_data_strength_reduction): ...new pass_data instance and...
9411 (make_pass_strength_reduction): ...new function.
9412 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9413 of rtl_opt_pass along with...
9414 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9415 (make_pass_rtl_ifcvt): ...new function.
9416 (pass_if_after_combine): Convert from a global struct to a subclass of
9417 rtl_opt_pass along with...
9418 (pass_data_if_after_combine): ...new pass_data instance and...
9419 (make_pass_if_after_combine): ...new function.
9420 (pass_if_after_reload): Convert from a global struct to a subclass of
9421 rtl_opt_pass along with...
9422 (pass_data_if_after_reload): ...new pass_data instance and...
9423 (make_pass_if_after_reload): ...new function.
9424 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9425 a subclass of rtl_opt_pass along with...
9426 (pass_data_initialize_regs): ...new pass_data instance and...
9427 (make_pass_initialize_regs): ...new function.
9428 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9429 of ipa_opt_pass_d along with...
9430 (pass_data_ipa_cp): ...new pass_data instance and...
9431 (make_pass_ipa_cp): ...new function.
9432 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9433 global struct to a subclass of gimple_opt_pass along with...
9434 (pass_data_inline_parameters): ...new pass_data instance and...
9435 (make_pass_inline_parameters): ...new function.
9436 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9437 subclass of gimple_opt_pass along with...
9438 (pass_data_early_inline): ...new pass_data instance and...
9439 (make_pass_early_inline): ...new function.
9440 (pass_ipa_inline): Convert from a global struct to a subclass of
9441 ipa_opt_pass_d along with...
9442 (pass_data_ipa_inline): ...new pass_data instance and...
9443 (make_pass_ipa_inline): ...new function.
9444 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9445 struct to a subclass of gimple_opt_pass along with...
9446 (pass_data_local_pure_const): ...new pass_data instance and...
9447 (make_pass_local_pure_const): ...new function.
9448 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9449 ipa_opt_pass_d along with...
9450 (pass_data_ipa_pure_const): ...new pass_data instance and...
9451 (make_pass_ipa_pure_const): ...new function.
9452 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9453 to a subclass of ipa_opt_pass_d along with...
9454 (pass_data_ipa_reference): ...new pass_data instance and...
9455 (make_pass_ipa_reference): ...new function.
9456 * ipa-split.c (pass_split_functions): Convert from a global struct to
9457 a subclass of gimple_opt_pass along with...
9458 (pass_data_split_functions): ...new pass_data instance and...
9459 (make_pass_split_functions): ...new function.
9460 (pass_feedback_split_functions): Convert from a global struct to a
9461 subclass of gimple_opt_pass along with...
9462 (pass_data_feedback_split_functions): ...new pass_data instance and...
9463 (make_pass_feedback_split_functions): ...new function.
9464 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9465 global struct to a subclass of simple_ipa_opt_pass along with...
9466 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9468 (make_pass_ipa_function_and_variable_visibility): ...new function.
9469 (pass_ipa_free_inline_summary): Convert from a global struct to a
9470 subclass of simple_ipa_opt_pass along with...
9471 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9472 (make_pass_ipa_free_inline_summary): ...new function.
9473 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9474 subclass of ipa_opt_pass_d along with...
9475 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9477 (make_pass_ipa_whole_program_visibility): ...new function.
9478 (pass_ipa_profile): Convert from a global struct to a subclass of
9479 ipa_opt_pass_d along with...
9480 (pass_data_ipa_profile): ...new pass_data instance and...
9481 (make_pass_ipa_profile): ...new function.
9482 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9483 ipa_opt_pass_d along with...
9484 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9485 (make_pass_ipa_cdtor_merge): ...new function.
9486 * ira.c (pass_ira): Convert from a global struct to a subclass of
9487 rtl_opt_pass along with...
9488 (pass_data_ira): ...new pass_data instance and...
9489 (make_pass_ira): ...new function.
9490 (pass_reload): Convert from a global struct to a subclass of
9491 rtl_opt_pass along with...
9492 (pass_data_reload): ...new pass_data instance and...
9493 (make_pass_reload): ...new function.
9494 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9495 subclass of rtl_opt_pass along with...
9496 (pass_data_cleanup_barriers): ...new pass_data instance and...
9497 (make_pass_cleanup_barriers): ...new function.
9498 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9499 of rtl_opt_pass along with...
9500 (pass_data_loop2): ...new pass_data instance and...
9501 (make_pass_loop2): ...new function.
9502 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9503 rtl_opt_pass along with...
9504 (pass_data_rtl_loop_init): ...new pass_data instance and...
9505 (make_pass_rtl_loop_init): ...new function.
9506 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9507 rtl_opt_pass along with...
9508 (pass_data_rtl_loop_done): ...new pass_data instance and...
9509 (make_pass_rtl_loop_done): ...new function.
9510 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9511 subclass of rtl_opt_pass along with...
9512 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9513 (make_pass_rtl_move_loop_invariants): ...new function.
9514 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9515 rtl_opt_pass along with...
9516 (pass_data_rtl_unswitch): ...new pass_data instance and...
9517 (make_pass_rtl_unswitch): ...new function.
9518 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9519 subclass of rtl_opt_pass along with...
9520 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9522 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9523 (pass_rtl_doloop): Convert from a global struct to a subclass of
9524 rtl_opt_pass along with...
9525 (pass_data_rtl_doloop): ...new pass_data instance and...
9526 (make_pass_rtl_doloop): ...new function.
9527 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9528 a subclass of rtl_opt_pass along with...
9529 (pass_data_lower_subreg): ...new pass_data instance and...
9530 (make_pass_lower_subreg): ...new function.
9531 (pass_lower_subreg2): Convert from a global struct to a subclass of
9532 rtl_opt_pass along with...
9533 (pass_data_lower_subreg2): ...new pass_data instance and...
9534 (make_pass_lower_subreg2): ...new function.
9535 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9536 struct to a subclass of ipa_opt_pass_d along with...
9537 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9538 (make_pass_ipa_lto_gimple_out): ...new function.
9539 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9540 of ipa_opt_pass_d along with...
9541 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9542 (make_pass_ipa_lto_finish_out): ...new function.
9543 * mode-switching.c (pass_mode_switching): Convert from a global struct
9544 to a subclass of rtl_opt_pass along with...
9545 (pass_data_mode_switching): ...new pass_data instance and...
9546 (make_pass_mode_switching): ...new function.
9547 * modulo-sched.c (pass_sms): Convert from a global struct to a
9548 subclass of rtl_opt_pass along with...
9549 (pass_data_sms): ...new pass_data instance and...
9550 (make_pass_sms): ...new function.
9551 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9552 subclass of gimple_opt_pass along with...
9553 (pass_data_expand_omp): ...new pass_data instance and...
9554 (make_pass_expand_omp): ...new function.
9555 (pass_lower_omp): Convert from a global struct to a subclass of
9556 gimple_opt_pass along with...
9557 (pass_data_lower_omp): ...new pass_data instance and...
9558 (make_pass_lower_omp): ...new function.
9559 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9560 of gimple_opt_pass along with...
9561 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9562 (make_pass_diagnose_omp_blocks): ...new function.
9563 * passes.c (pass_early_local_passes): Convert from a global struct to
9564 a subclass of simple_ipa_opt_pass along with...
9565 (pass_data_early_local_passes): ...new pass_data instance and...
9566 (make_pass_early_local_passes): ...new function.
9567 (pass_all_early_optimizations): Convert from a global struct to a
9568 subclass of gimple_opt_pass along with...
9569 (pass_data_all_early_optimizations): ...new pass_data instance and...
9570 (make_pass_all_early_optimizations): ...new function.
9571 (pass_all_optimizations): Convert from a global struct to a subclass
9572 of gimple_opt_pass along with...
9573 (pass_data_all_optimizations): ...new pass_data instance and...
9574 (make_pass_all_optimizations): ...new function.
9575 (pass_all_optimizations_g): Convert from a global struct to a subclass
9576 of gimple_opt_pass along with...
9577 (pass_data_all_optimizations_g): ...new pass_data instance and...
9578 (make_pass_all_optimizations_g): ...new function.
9579 (pass_rest_of_compilation): Convert from a global struct to a subclass
9580 of rtl_opt_pass along with...
9581 (pass_data_rest_of_compilation): ...new pass_data instance and...
9582 (make_pass_rest_of_compilation): ...new function.
9583 (pass_postreload): Convert from a global struct to a subclass of
9584 rtl_opt_pass along with...
9585 (pass_data_postreload): ...new pass_data instance and...
9586 (make_pass_postreload): ...new function.
9587 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9588 subclass of rtl_opt_pass along with...
9589 (pass_data_gcse2): ...new pass_data instance and...
9590 (make_pass_gcse2): ...new function.
9591 * postreload.c (pass_postreload_cse): Convert from a global struct to
9592 a subclass of rtl_opt_pass along with...
9593 (pass_data_postreload_cse): ...new pass_data instance and...
9594 (make_pass_postreload_cse): ...new function.
9595 * predict.c (pass_profile): Convert from a global struct to a subclass
9596 of gimple_opt_pass along with...
9597 (pass_data_profile): ...new pass_data instance and...
9598 (make_pass_profile): ...new function.
9599 (pass_strip_predict_hints): Convert from a global struct to a subclass
9600 of gimple_opt_pass along with...
9601 (pass_data_strip_predict_hints): ...new pass_data instance and...
9602 (make_pass_strip_predict_hints): ...new function.
9603 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9604 of rtl_opt_pass along with...
9605 (pass_data_peephole2): ...new pass_data instance and...
9606 (make_pass_peephole2): ...new function.
9607 (pass_split_all_insns): Convert from a global struct to a subclass of
9608 rtl_opt_pass along with...
9609 (pass_data_split_all_insns): ...new pass_data instance and...
9610 (make_pass_split_all_insns): ...new function.
9611 (pass_split_after_reload): Convert from a global struct to a subclass
9612 of rtl_opt_pass along with...
9613 (pass_data_split_after_reload): ...new pass_data instance and...
9614 (make_pass_split_after_reload): ...new function.
9615 (pass_split_before_regstack): Convert from a global struct to a
9616 subclass of rtl_opt_pass along with...
9617 (pass_data_split_before_regstack): ...new pass_data instance and...
9618 (make_pass_split_before_regstack): ...new function.
9619 (pass_split_before_sched2): Convert from a global struct to a subclass
9620 of rtl_opt_pass along with...
9621 (pass_data_split_before_sched2): ...new pass_data instance and...
9622 (make_pass_split_before_sched2): ...new function.
9623 (pass_split_for_shorten_branches): Convert from a global struct to a
9624 subclass of rtl_opt_pass along with...
9625 (pass_data_split_for_shorten_branches): ...new pass_data instance
9627 (make_pass_split_for_shorten_branches): ...new function.
9628 * ree.c (pass_ree): Convert from a global struct to a subclass of
9629 rtl_opt_pass along with...
9630 (pass_data_ree): ...new pass_data instance and...
9631 (make_pass_ree): ...new function.
9632 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9633 subclass of rtl_opt_pass along with...
9634 (pass_data_stack_regs): ...new pass_data instance and...
9635 (make_pass_stack_regs): ...new function.
9636 (pass_stack_regs_run): Convert from a global struct to a subclass of
9637 rtl_opt_pass along with...
9638 (pass_data_stack_regs_run): ...new pass_data instance and...
9639 (make_pass_stack_regs_run): ...new function.
9640 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9641 subclass of rtl_opt_pass along with...
9642 (pass_data_cprop_hardreg): ...new pass_data instance and...
9643 (make_pass_cprop_hardreg): ...new function.
9644 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9645 subclass of rtl_opt_pass along with...
9646 (pass_data_reginfo_init): ...new pass_data instance and...
9647 (make_pass_reginfo_init): ...new function.
9648 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9649 of rtl_opt_pass along with...
9650 (pass_data_regmove): ...new pass_data instance and...
9651 (make_pass_regmove): ...new function.
9652 * regrename.c (pass_regrename): Convert from a global struct to a
9653 subclass of rtl_opt_pass along with...
9654 (pass_data_regrename): ...new pass_data instance and...
9655 (make_pass_regrename): ...new function.
9656 * reorg.c (pass_delay_slots): Convert from a global struct to a
9657 subclass of rtl_opt_pass along with...
9658 (pass_data_delay_slots): ...new pass_data instance and...
9659 (make_pass_delay_slots): ...new function.
9660 (pass_machine_reorg): Convert from a global struct to a subclass of
9661 rtl_opt_pass along with...
9662 (pass_data_machine_reorg): ...new pass_data instance and...
9663 (make_pass_machine_reorg): ...new function.
9664 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9665 of rtl_opt_pass along with...
9666 (pass_data_sched): ...new pass_data instance and...
9667 (make_pass_sched): ...new function.
9668 (pass_sched2): Convert from a global struct to a subclass of
9669 rtl_opt_pass along with...
9670 (pass_data_sched2): ...new pass_data instance and...
9671 (make_pass_sched2): ...new function.
9672 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9673 to a subclass of rtl_opt_pass along with...
9674 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9675 (make_pass_stack_ptr_mod): ...new function.
9676 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9677 to a subclass of rtl_opt_pass along with...
9678 (pass_data_rtl_store_motion): ...new pass_data instance and...
9679 (make_pass_rtl_store_motion): ...new function.
9680 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9681 of gimple_opt_pass along with...
9682 (pass_data_tracer): ...new pass_data instance and...
9683 (make_pass_tracer): ...new function.
9684 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9685 to a subclass of gimple_opt_pass along with...
9686 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9687 (make_pass_diagnose_tm_blocks): ...new function.
9688 (pass_lower_tm): Convert from a global struct to a subclass of
9689 gimple_opt_pass along with...
9690 (pass_data_lower_tm): ...new pass_data instance and...
9691 (make_pass_lower_tm): ...new function.
9692 (pass_tm_init): Convert from a global struct to a subclass of
9693 gimple_opt_pass along with...
9694 (pass_data_tm_init): ...new pass_data instance and...
9695 (make_pass_tm_init): ...new function.
9696 (pass_tm_mark): Convert from a global struct to a subclass of
9697 gimple_opt_pass along with...
9698 (pass_data_tm_mark): ...new pass_data instance and...
9699 (make_pass_tm_mark): ...new function.
9700 (pass_tm_edges): Convert from a global struct to a subclass of
9701 gimple_opt_pass along with...
9702 (pass_data_tm_edges): ...new pass_data instance and...
9703 (make_pass_tm_edges): ...new function.
9704 (pass_tm_memopt): Convert from a global struct to a subclass of
9705 gimple_opt_pass along with...
9706 (pass_data_tm_memopt): ...new pass_data instance and...
9707 (make_pass_tm_memopt): ...new function.
9708 (pass_ipa_tm): Convert from a global struct to a subclass of
9709 simple_ipa_opt_pass along with...
9710 (pass_data_ipa_tm): ...new pass_data instance and...
9711 (make_pass_ipa_tm): ...new function.
9712 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9713 subclass of gimple_opt_pass along with...
9714 (pass_data_call_cdce): ...new pass_data instance and...
9715 (make_pass_call_cdce): ...new function.
9716 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9717 subclass of gimple_opt_pass along with...
9718 (pass_data_build_cfg): ...new pass_data instance and...
9719 (make_pass_build_cfg): ...new function.
9720 (pass_split_crit_edges): Convert from a global struct to a subclass of
9721 gimple_opt_pass along with...
9722 (pass_data_split_crit_edges): ...new pass_data instance and...
9723 (make_pass_split_crit_edges): ...new function.
9724 (pass_warn_function_return): Convert from a global struct to a
9725 subclass of gimple_opt_pass along with...
9726 (pass_data_warn_function_return): ...new pass_data instance and...
9727 (make_pass_warn_function_return): ...new function.
9728 (pass_warn_function_noreturn): Convert from a global struct to a
9729 subclass of gimple_opt_pass along with...
9730 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9731 (make_pass_warn_function_noreturn): ...new function.
9732 (pass_warn_unused_result): Convert from a global struct to a subclass
9733 of gimple_opt_pass along with...
9734 (pass_data_warn_unused_result): ...new pass_data instance and...
9735 (make_pass_warn_unused_result): ...new function.
9736 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9737 a subclass of gimple_opt_pass along with...
9738 (pass_data_merge_phi): ...new pass_data instance and...
9739 (make_pass_merge_phi): ...new function.
9740 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9741 a subclass of gimple_opt_pass along with...
9742 (pass_data_lower_complex): ...new pass_data instance and...
9743 (make_pass_lower_complex): ...new function.
9744 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9745 gimple_opt_pass along with...
9746 (pass_data_lower_complex_O0): ...new pass_data instance and...
9747 (make_pass_lower_complex_O0): ...new function.
9748 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9749 subclass of gimple_opt_pass along with...
9750 (pass_data_lower_eh): ...new pass_data instance and...
9751 (make_pass_lower_eh): ...new function.
9752 (pass_refactor_eh): Convert from a global struct to a subclass of
9753 gimple_opt_pass along with...
9754 (pass_data_refactor_eh): ...new pass_data instance and...
9755 (make_pass_refactor_eh): ...new function.
9756 (pass_lower_resx): Convert from a global struct to a subclass of
9757 gimple_opt_pass along with...
9758 (pass_data_lower_resx): ...new pass_data instance and...
9759 (make_pass_lower_resx): ...new function.
9760 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9761 of gimple_opt_pass along with...
9762 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9763 (make_pass_lower_eh_dispatch): ...new function.
9764 (pass_cleanup_eh): Convert from a global struct to a subclass of
9765 gimple_opt_pass along with...
9766 (pass_data_cleanup_eh): ...new pass_data instance and...
9767 (make_pass_cleanup_eh): ...new function.
9768 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9769 to a subclass of simple_ipa_opt_pass along with...
9770 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9771 (make_pass_ipa_lower_emutls): ...new function.
9772 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9773 a subclass of gimple_opt_pass along with...
9774 (pass_data_if_conversion): ...new pass_data instance and...
9775 (make_pass_if_conversion): ...new function.
9776 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9777 subclass of gimple_opt_pass along with...
9778 (pass_data_build_ssa): ...new pass_data instance and...
9779 (make_pass_build_ssa): ...new function.
9780 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9781 global struct to a subclass of gimple_opt_pass along with...
9782 (pass_data_loop_distribution): ...new pass_data instance and...
9783 (make_pass_loop_distribution): ...new function.
9784 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9785 subclass of gimple_opt_pass along with...
9786 (pass_data_mudflap_1): ...new pass_data instance and...
9787 (make_pass_mudflap_1): ...new function.
9788 (pass_mudflap_2): Convert from a global struct to a subclass of
9789 gimple_opt_pass along with...
9790 (pass_data_mudflap_2): ...new pass_data instance and...
9791 (make_pass_mudflap_2): ...new function.
9792 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9793 subclass of gimple_opt_pass along with...
9794 (pass_data_mudflap_1): ...new pass_data instance and...
9795 (make_pass_mudflap_1): ...new function.
9796 (pass_mudflap_2): Convert from a global struct to a subclass of
9797 gimple_opt_pass along with...
9798 (pass_data_mudflap_2): ...new pass_data instance and...
9799 (make_pass_mudflap_2): ...new function.
9800 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9801 gimple_opt_pass along with...
9802 (pass_data_nrv): ...new pass_data instance and...
9803 (make_pass_nrv): ...new function.
9804 (pass_return_slot): Convert from a global struct to a subclass of
9805 gimple_opt_pass along with...
9806 (pass_data_return_slot): ...new pass_data instance and...
9807 (make_pass_return_slot): ...new function.
9808 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9809 to a subclass of gimple_opt_pass along with...
9810 (pass_data_object_sizes): ...new pass_data instance and...
9811 (make_pass_object_sizes): ...new function.
9812 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9813 global struct to a subclass of gimple_opt_pass along with...
9814 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9816 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9817 (pass_fixup_cfg): Convert from a global struct to a subclass of
9818 gimple_opt_pass along with...
9819 (pass_data_fixup_cfg): ...new pass_data instance and...
9820 (make_pass_fixup_cfg): ...new function.
9821 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9822 (make_pass_mudflap_1): ...new function.
9823 (pass_mudflap_2): Replace declaration with that of...
9824 (make_pass_mudflap_2): ...new function.
9825 (pass_asan): Replace declaration with that of...
9826 (make_pass_asan): ...new function.
9827 (pass_asan_O0): Replace declaration with that of...
9828 (make_pass_asan_O0): ...new function.
9829 (pass_tsan): Replace declaration with that of...
9830 (make_pass_tsan): ...new function.
9831 (pass_tsan_O0): Replace declaration with that of...
9832 (make_pass_tsan_O0): ...new function.
9833 (pass_lower_cf): Replace declaration with that of...
9834 (make_pass_lower_cf): ...new function.
9835 (pass_refactor_eh): Replace declaration with that of...
9836 (make_pass_refactor_eh): ...new function.
9837 (pass_lower_eh): Replace declaration with that of...
9838 (make_pass_lower_eh): ...new function.
9839 (pass_lower_eh_dispatch): Replace declaration with that of...
9840 (make_pass_lower_eh_dispatch): ...new function.
9841 (pass_lower_resx): Replace declaration with that of...
9842 (make_pass_lower_resx): ...new function.
9843 (pass_build_cfg): Replace declaration with that of...
9844 (make_pass_build_cfg): ...new function.
9845 (pass_early_tree_profile): Replace declaration with that of...
9846 (make_pass_early_tree_profile): ...new function.
9847 (pass_cleanup_eh): Replace declaration with that of...
9848 (make_pass_cleanup_eh): ...new function.
9849 (pass_sra): Replace declaration with that of...
9850 (make_pass_sra): ...new function.
9851 (pass_sra_early): Replace declaration with that of...
9852 (make_pass_sra_early): ...new function.
9853 (pass_early_ipa_sra): Replace declaration with that of...
9854 (make_pass_early_ipa_sra): ...new function.
9855 (pass_tail_recursion): Replace declaration with that of...
9856 (make_pass_tail_recursion): ...new function.
9857 (pass_tail_calls): Replace declaration with that of...
9858 (make_pass_tail_calls): ...new function.
9859 (pass_tree_loop): Replace declaration with that of...
9860 (make_pass_tree_loop): ...new function.
9861 (pass_tree_loop_init): Replace declaration with that of...
9862 (make_pass_tree_loop_init): ...new function.
9863 (pass_lim): Replace declaration with that of...
9864 (make_pass_lim): ...new function.
9865 (pass_tree_unswitch): Replace declaration with that of...
9866 (make_pass_tree_unswitch): ...new function.
9867 (pass_predcom): Replace declaration with that of...
9868 (make_pass_predcom): ...new function.
9869 (pass_iv_canon): Replace declaration with that of...
9870 (make_pass_iv_canon): ...new function.
9871 (pass_scev_cprop): Replace declaration with that of...
9872 (make_pass_scev_cprop): ...new function.
9873 (pass_empty_loop): Replace declaration with that of...
9874 (make_pass_empty_loop): ...new function.
9875 (pass_record_bounds): Replace declaration with that of...
9876 (make_pass_record_bounds): ...new function.
9877 (pass_graphite): Replace declaration with that of...
9878 (make_pass_graphite): ...new function.
9879 (pass_graphite_transforms): Replace declaration with that of...
9880 (make_pass_graphite_transforms): ...new function.
9881 (pass_if_conversion): Replace declaration with that of...
9882 (make_pass_if_conversion): ...new function.
9883 (pass_loop_distribution): Replace declaration with that of...
9884 (make_pass_loop_distribution): ...new function.
9885 (pass_vectorize): Replace declaration with that of...
9886 (make_pass_vectorize): ...new function.
9887 (pass_slp_vectorize): Replace declaration with that of...
9888 (make_pass_slp_vectorize): ...new function.
9889 (pass_complete_unroll): Replace declaration with that of...
9890 (make_pass_complete_unroll): ...new function.
9891 (pass_complete_unrolli): Replace declaration with that of...
9892 (make_pass_complete_unrolli): ...new function.
9893 (pass_parallelize_loops): Replace declaration with that of...
9894 (make_pass_parallelize_loops): ...new function.
9895 (pass_loop_prefetch): Replace declaration with that of...
9896 (make_pass_loop_prefetch): ...new function.
9897 (pass_iv_optimize): Replace declaration with that of...
9898 (make_pass_iv_optimize): ...new function.
9899 (pass_tree_loop_done): Replace declaration with that of...
9900 (make_pass_tree_loop_done): ...new function.
9901 (pass_ch): Replace declaration with that of...
9902 (make_pass_ch): ...new function.
9903 (pass_ccp): Replace declaration with that of...
9904 (make_pass_ccp): ...new function.
9905 (pass_phi_only_cprop): Replace declaration with that of...
9906 (make_pass_phi_only_cprop): ...new function.
9907 (pass_build_ssa): Replace declaration with that of...
9908 (make_pass_build_ssa): ...new function.
9909 (pass_build_alias): Replace declaration with that of...
9910 (make_pass_build_alias): ...new function.
9911 (pass_build_ealias): Replace declaration with that of...
9912 (make_pass_build_ealias): ...new function.
9913 (pass_dominator): Replace declaration with that of...
9914 (make_pass_dominator): ...new function.
9915 (pass_dce): Replace declaration with that of...
9916 (make_pass_dce): ...new function.
9917 (pass_dce_loop): Replace declaration with that of...
9918 (make_pass_dce_loop): ...new function.
9919 (pass_cd_dce): Replace declaration with that of...
9920 (make_pass_cd_dce): ...new function.
9921 (pass_call_cdce): Replace declaration with that of...
9922 (make_pass_call_cdce): ...new function.
9923 (pass_merge_phi): Replace declaration with that of...
9924 (make_pass_merge_phi): ...new function.
9925 (pass_split_crit_edges): Replace declaration with that of...
9926 (make_pass_split_crit_edges): ...new function.
9927 (pass_pre): Replace declaration with that of...
9928 (make_pass_pre): ...new function.
9929 (pass_profile): Replace declaration with that of...
9930 (make_pass_profile): ...new function.
9931 (pass_strip_predict_hints): Replace declaration with that of...
9932 (make_pass_strip_predict_hints): ...new function.
9933 (pass_lower_complex_O0): Replace declaration with that of...
9934 (make_pass_lower_complex_O0): ...new function.
9935 (pass_lower_complex): Replace declaration with that of...
9936 (make_pass_lower_complex): ...new function.
9937 (pass_lower_vector): Replace declaration with that of...
9938 (make_pass_lower_vector): ...new function.
9939 (pass_lower_vector_ssa): Replace declaration with that of...
9940 (make_pass_lower_vector_ssa): ...new function.
9941 (pass_lower_omp): Replace declaration with that of...
9942 (make_pass_lower_omp): ...new function.
9943 (pass_diagnose_omp_blocks): Replace declaration with that of...
9944 (make_pass_diagnose_omp_blocks): ...new function.
9945 (pass_expand_omp): Replace declaration with that of...
9946 (make_pass_expand_omp): ...new function.
9947 (pass_expand_omp_ssa): Replace declaration with that of...
9948 (make_pass_expand_omp_ssa): ...new function.
9949 (pass_object_sizes): Replace declaration with that of...
9950 (make_pass_object_sizes): ...new function.
9951 (pass_strlen): Replace declaration with that of...
9952 (make_pass_strlen): ...new function.
9953 (pass_fold_builtins): Replace declaration with that of...
9954 (make_pass_fold_builtins): ...new function.
9955 (pass_stdarg): Replace declaration with that of...
9956 (make_pass_stdarg): ...new function.
9957 (pass_early_warn_uninitialized): Replace declaration with that of...
9958 (make_pass_early_warn_uninitialized): ...new function.
9959 (pass_late_warn_uninitialized): Replace declaration with that of...
9960 (make_pass_late_warn_uninitialized): ...new function.
9961 (pass_cse_reciprocals): Replace declaration with that of...
9962 (make_pass_cse_reciprocals): ...new function.
9963 (pass_cse_sincos): Replace declaration with that of...
9964 (make_pass_cse_sincos): ...new function.
9965 (pass_optimize_bswap): Replace declaration with that of...
9966 (make_pass_optimize_bswap): ...new function.
9967 (pass_optimize_widening_mul): Replace declaration with that of...
9968 (make_pass_optimize_widening_mul): ...new function.
9969 (pass_warn_function_return): Replace declaration with that of...
9970 (make_pass_warn_function_return): ...new function.
9971 (pass_warn_function_noreturn): Replace declaration with that of...
9972 (make_pass_warn_function_noreturn): ...new function.
9973 (pass_cselim): Replace declaration with that of...
9974 (make_pass_cselim): ...new function.
9975 (pass_phiopt): Replace declaration with that of...
9976 (make_pass_phiopt): ...new function.
9977 (pass_forwprop): Replace declaration with that of...
9978 (make_pass_forwprop): ...new function.
9979 (pass_phiprop): Replace declaration with that of...
9980 (make_pass_phiprop): ...new function.
9981 (pass_tree_ifcombine): Replace declaration with that of...
9982 (make_pass_tree_ifcombine): ...new function.
9983 (pass_dse): Replace declaration with that of...
9984 (make_pass_dse): ...new function.
9985 (pass_nrv): Replace declaration with that of...
9986 (make_pass_nrv): ...new function.
9987 (pass_rename_ssa_copies): Replace declaration with that of...
9988 (make_pass_rename_ssa_copies): ...new function.
9989 (pass_sink_code): Replace declaration with that of...
9990 (make_pass_sink_code): ...new function.
9991 (pass_fre): Replace declaration with that of...
9992 (make_pass_fre): ...new function.
9993 (pass_check_data_deps): Replace declaration with that of...
9994 (make_pass_check_data_deps): ...new function.
9995 (pass_copy_prop): Replace declaration with that of...
9996 (make_pass_copy_prop): ...new function.
9997 (pass_vrp): Replace declaration with that of...
9998 (make_pass_vrp): ...new function.
9999 (pass_uncprop): Replace declaration with that of...
10000 (make_pass_uncprop): ...new function.
10001 (pass_return_slot): Replace declaration with that of...
10002 (make_pass_return_slot): ...new function.
10003 (pass_reassoc): Replace declaration with that of...
10004 (make_pass_reassoc): ...new function.
10005 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10006 (make_pass_rebuild_cgraph_edges): ...new function.
10007 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10008 (make_pass_remove_cgraph_callee_edges): ...new function.
10009 (pass_build_cgraph_edges): Replace declaration with that of...
10010 (make_pass_build_cgraph_edges): ...new function.
10011 (pass_local_pure_const): Replace declaration with that of...
10012 (make_pass_local_pure_const): ...new function.
10013 (pass_tracer): Replace declaration with that of...
10014 (make_pass_tracer): ...new function.
10015 (pass_warn_unused_result): Replace declaration with that of...
10016 (make_pass_warn_unused_result): ...new function.
10017 (pass_diagnose_tm_blocks): Replace declaration with that of...
10018 (make_pass_diagnose_tm_blocks): ...new function.
10019 (pass_lower_tm): Replace declaration with that of...
10020 (make_pass_lower_tm): ...new function.
10021 (pass_tm_init): Replace declaration with that of...
10022 (make_pass_tm_init): ...new function.
10023 (pass_tm_mark): Replace declaration with that of...
10024 (make_pass_tm_mark): ...new function.
10025 (pass_tm_memopt): Replace declaration with that of...
10026 (make_pass_tm_memopt): ...new function.
10027 (pass_tm_edges): Replace declaration with that of...
10028 (make_pass_tm_edges): ...new function.
10029 (pass_split_functions): Replace declaration with that of...
10030 (make_pass_split_functions): ...new function.
10031 (pass_feedback_split_functions): Replace declaration with that of...
10032 (make_pass_feedback_split_functions): ...new function.
10033 (pass_strength_reduction): Replace declaration with that of...
10034 (make_pass_strength_reduction): ...new function.
10035 (pass_ipa_lower_emutls): Replace declaration with that of...
10036 (make_pass_ipa_lower_emutls): ...new function.
10037 (pass_ipa_function_and_variable_visibility): Replace declaration with
10039 (make_pass_ipa_function_and_variable_visibility): ...new function.
10040 (pass_ipa_tree_profile): Replace declaration with that of...
10041 (make_pass_ipa_tree_profile): ...new function.
10042 (pass_early_local_passes): Replace declaration with that of...
10043 (make_pass_early_local_passes): ...new function.
10044 (pass_ipa_whole_program_visibility): Replace declaration with that
10046 (make_pass_ipa_whole_program_visibility): ...new function.
10047 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10048 (make_pass_ipa_lto_gimple_out): ...new function.
10049 (pass_ipa_increase_alignment): Replace declaration with that of...
10050 (make_pass_ipa_increase_alignment): ...new function.
10051 (pass_ipa_inline): Replace declaration with that of...
10052 (make_pass_ipa_inline): ...new function.
10053 (pass_ipa_free_lang_data): Replace declaration with that of...
10054 (make_pass_ipa_free_lang_data): ...new function.
10055 (pass_ipa_free_inline_summary): Replace declaration with that of...
10056 (make_pass_ipa_free_inline_summary): ...new function.
10057 (pass_ipa_cp): Replace declaration with that of...
10058 (make_pass_ipa_cp): ...new function.
10059 (pass_ipa_reference): Replace declaration with that of...
10060 (make_pass_ipa_reference): ...new function.
10061 (pass_ipa_pure_const): Replace declaration with that of...
10062 (make_pass_ipa_pure_const): ...new function.
10063 (pass_ipa_pta): Replace declaration with that of...
10064 (make_pass_ipa_pta): ...new function.
10065 (pass_ipa_lto_finish_out): Replace declaration with that of...
10066 (make_pass_ipa_lto_finish_out): ...new function.
10067 (pass_ipa_tm): Replace declaration with that of...
10068 (make_pass_ipa_tm): ...new function.
10069 (pass_ipa_profile): Replace declaration with that of...
10070 (make_pass_ipa_profile): ...new function.
10071 (pass_ipa_cdtor_merge): Replace declaration with that of...
10072 (make_pass_ipa_cdtor_merge): ...new function.
10073 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10075 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10076 (pass_init_datastructures): Replace declaration with that of...
10077 (make_pass_init_datastructures): ...new function.
10078 (pass_fixup_cfg): Replace declaration with that of...
10079 (make_pass_fixup_cfg): ...new function.
10080 (pass_expand): Replace declaration with that of...
10081 (make_pass_expand): ...new function.
10082 (pass_instantiate_virtual_regs): Replace declaration with that of...
10083 (make_pass_instantiate_virtual_regs): ...new function.
10084 (pass_rtl_fwprop): Replace declaration with that of...
10085 (make_pass_rtl_fwprop): ...new function.
10086 (pass_rtl_fwprop_addr): Replace declaration with that of...
10087 (make_pass_rtl_fwprop_addr): ...new function.
10088 (pass_jump): Replace declaration with that of...
10089 (make_pass_jump): ...new function.
10090 (pass_jump2): Replace declaration with that of...
10091 (make_pass_jump2): ...new function.
10092 (pass_lower_subreg): Replace declaration with that of...
10093 (make_pass_lower_subreg): ...new function.
10094 (pass_cse): Replace declaration with that of...
10095 (make_pass_cse): ...new function.
10096 (pass_fast_rtl_dce): Replace declaration with that of...
10097 (make_pass_fast_rtl_dce): ...new function.
10098 (pass_ud_rtl_dce): Replace declaration with that of...
10099 (make_pass_ud_rtl_dce): ...new function.
10100 (pass_rtl_dce): Replace declaration with that of...
10101 (make_pass_rtl_dce): ...new function.
10102 (pass_rtl_dse1): Replace declaration with that of...
10103 (make_pass_rtl_dse1): ...new function.
10104 (pass_rtl_dse2): Replace declaration with that of...
10105 (make_pass_rtl_dse2): ...new function.
10106 (pass_rtl_dse3): Replace declaration with that of...
10107 (make_pass_rtl_dse3): ...new function.
10108 (pass_rtl_cprop): Replace declaration with that of...
10109 (make_pass_rtl_cprop): ...new function.
10110 (pass_rtl_pre): Replace declaration with that of...
10111 (make_pass_rtl_pre): ...new function.
10112 (pass_rtl_hoist): Replace declaration with that of...
10113 (make_pass_rtl_hoist): ...new function.
10114 (pass_rtl_store_motion): Replace declaration with that of...
10115 (make_pass_rtl_store_motion): ...new function.
10116 (pass_cse_after_global_opts): Replace declaration with that of...
10117 (make_pass_cse_after_global_opts): ...new function.
10118 (pass_rtl_ifcvt): Replace declaration with that of...
10119 (make_pass_rtl_ifcvt): ...new function.
10120 (pass_into_cfg_layout_mode): Replace declaration with that of...
10121 (make_pass_into_cfg_layout_mode): ...new function.
10122 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10123 (make_pass_outof_cfg_layout_mode): ...new function.
10124 (pass_loop2): Replace declaration with that of...
10125 (make_pass_loop2): ...new function.
10126 (pass_rtl_loop_init): Replace declaration with that of...
10127 (make_pass_rtl_loop_init): ...new function.
10128 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10129 (make_pass_rtl_move_loop_invariants): ...new function.
10130 (pass_rtl_unswitch): Replace declaration with that of...
10131 (make_pass_rtl_unswitch): ...new function.
10132 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10133 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10134 (pass_rtl_doloop): Replace declaration with that of...
10135 (make_pass_rtl_doloop): ...new function.
10136 (pass_rtl_loop_done): Replace declaration with that of...
10137 (make_pass_rtl_loop_done): ...new function.
10138 (pass_web): Replace declaration with that of...
10139 (make_pass_web): ...new function.
10140 (pass_cse2): Replace declaration with that of...
10141 (make_pass_cse2): ...new function.
10142 (pass_df_initialize_opt): Replace declaration with that of...
10143 (make_pass_df_initialize_opt): ...new function.
10144 (pass_df_initialize_no_opt): Replace declaration with that of...
10145 (make_pass_df_initialize_no_opt): ...new function.
10146 (pass_reginfo_init): Replace declaration with that of...
10147 (make_pass_reginfo_init): ...new function.
10148 (pass_inc_dec): Replace declaration with that of...
10149 (make_pass_inc_dec): ...new function.
10150 (pass_stack_ptr_mod): Replace declaration with that of...
10151 (make_pass_stack_ptr_mod): ...new function.
10152 (pass_initialize_regs): Replace declaration with that of...
10153 (make_pass_initialize_regs): ...new function.
10154 (pass_combine): Replace declaration with that of...
10155 (make_pass_combine): ...new function.
10156 (pass_if_after_combine): Replace declaration with that of...
10157 (make_pass_if_after_combine): ...new function.
10158 (pass_ree): Replace declaration with that of...
10159 (make_pass_ree): ...new function.
10160 (pass_partition_blocks): Replace declaration with that of...
10161 (make_pass_partition_blocks): ...new function.
10162 (pass_match_asm_constraints): Replace declaration with that of...
10163 (make_pass_match_asm_constraints): ...new function.
10164 (pass_regmove): Replace declaration with that of...
10165 (make_pass_regmove): ...new function.
10166 (pass_split_all_insns): Replace declaration with that of...
10167 (make_pass_split_all_insns): ...new function.
10168 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10169 (make_pass_fast_rtl_byte_dce): ...new function.
10170 (pass_lower_subreg2): Replace declaration with that of...
10171 (make_pass_lower_subreg2): ...new function.
10172 (pass_mode_switching): Replace declaration with that of...
10173 (make_pass_mode_switching): ...new function.
10174 (pass_sms): Replace declaration with that of...
10175 (make_pass_sms): ...new function.
10176 (pass_sched): Replace declaration with that of...
10177 (make_pass_sched): ...new function.
10178 (pass_ira): Replace declaration with that of...
10179 (make_pass_ira): ...new function.
10180 (pass_reload): Replace declaration with that of...
10181 (make_pass_reload): ...new function.
10182 (pass_clean_state): Replace declaration with that of...
10183 (make_pass_clean_state): ...new function.
10184 (pass_branch_prob): Replace declaration with that of...
10185 (make_pass_branch_prob): ...new function.
10186 (pass_value_profile_transformations): Replace declaration with that
10188 (make_pass_value_profile_transformations): ...new function.
10189 (pass_postreload_cse): Replace declaration with that of...
10190 (make_pass_postreload_cse): ...new function.
10191 (pass_gcse2): Replace declaration with that of...
10192 (make_pass_gcse2): ...new function.
10193 (pass_split_after_reload): Replace declaration with that of...
10194 (make_pass_split_after_reload): ...new function.
10195 (pass_branch_target_load_optimize1): Replace declaration with that
10197 (make_pass_branch_target_load_optimize1): ...new function.
10198 (pass_thread_prologue_and_epilogue): Replace declaration with that
10200 (make_pass_thread_prologue_and_epilogue): ...new function.
10201 (pass_stack_adjustments): Replace declaration with that of...
10202 (make_pass_stack_adjustments): ...new function.
10203 (pass_peephole2): Replace declaration with that of...
10204 (make_pass_peephole2): ...new function.
10205 (pass_if_after_reload): Replace declaration with that of...
10206 (make_pass_if_after_reload): ...new function.
10207 (pass_regrename): Replace declaration with that of...
10208 (make_pass_regrename): ...new function.
10209 (pass_cprop_hardreg): Replace declaration with that of...
10210 (make_pass_cprop_hardreg): ...new function.
10211 (pass_reorder_blocks): Replace declaration with that of...
10212 (make_pass_reorder_blocks): ...new function.
10213 (pass_branch_target_load_optimize2): Replace declaration with that
10215 (make_pass_branch_target_load_optimize2): ...new function.
10216 (pass_leaf_regs): Replace declaration with that of...
10217 (make_pass_leaf_regs): ...new function.
10218 (pass_split_before_sched2): Replace declaration with that of...
10219 (make_pass_split_before_sched2): ...new function.
10220 (pass_compare_elim_after_reload): Replace declaration with that of...
10221 (make_pass_compare_elim_after_reload): ...new function.
10222 (pass_sched2): Replace declaration with that of...
10223 (make_pass_sched2): ...new function.
10224 (pass_stack_regs): Replace declaration with that of...
10225 (make_pass_stack_regs): ...new function.
10226 (pass_stack_regs_run): Replace declaration with that of...
10227 (make_pass_stack_regs_run): ...new function.
10228 (pass_df_finish): Replace declaration with that of...
10229 (make_pass_df_finish): ...new function.
10230 (pass_compute_alignments): Replace declaration with that of...
10231 (make_pass_compute_alignments): ...new function.
10232 (pass_duplicate_computed_gotos): Replace declaration with that of...
10233 (make_pass_duplicate_computed_gotos): ...new function.
10234 (pass_variable_tracking): Replace declaration with that of...
10235 (make_pass_variable_tracking): ...new function.
10236 (pass_free_cfg): Replace declaration with that of...
10237 (make_pass_free_cfg): ...new function.
10238 (pass_machine_reorg): Replace declaration with that of...
10239 (make_pass_machine_reorg): ...new function.
10240 (pass_cleanup_barriers): Replace declaration with that of...
10241 (make_pass_cleanup_barriers): ...new function.
10242 (pass_delay_slots): Replace declaration with that of...
10243 (make_pass_delay_slots): ...new function.
10244 (pass_split_for_shorten_branches): Replace declaration with that of...
10245 (make_pass_split_for_shorten_branches): ...new function.
10246 (pass_split_before_regstack): Replace declaration with that of...
10247 (make_pass_split_before_regstack): ...new function.
10248 (pass_convert_to_eh_region_ranges): Replace declaration with that
10250 (make_pass_convert_to_eh_region_ranges): ...new function.
10251 (pass_shorten_branches): Replace declaration with that of...
10252 (make_pass_shorten_branches): ...new function.
10253 (pass_set_nothrow_function_flags): Replace declaration with that of...
10254 (make_pass_set_nothrow_function_flags): ...new function.
10255 (pass_dwarf2_frame): Replace declaration with that of...
10256 (make_pass_dwarf2_frame): ...new function.
10257 (pass_final): Replace declaration with that of...
10258 (make_pass_final): ...new function.
10259 (pass_rtl_seqabstr): Replace declaration with that of...
10260 (make_pass_rtl_seqabstr): ...new function.
10261 (pass_release_ssa_names): Replace declaration with that of...
10262 (make_pass_release_ssa_names): ...new function.
10263 (pass_early_inline): Replace declaration with that of...
10264 (make_pass_early_inline): ...new function.
10265 (pass_inline_parameters): Replace declaration with that of...
10266 (make_pass_inline_parameters): ...new function.
10267 (pass_update_address_taken): Replace declaration with that of...
10268 (make_pass_update_address_taken): ...new function.
10269 (pass_convert_switch): Replace declaration with that of...
10270 (make_pass_convert_switch): ...new function.
10271 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10272 to a subclass of simple_ipa_opt_pass along with...
10273 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10274 (make_pass_ipa_tree_profile): ...new function.
10275 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10276 subclass of gimple_opt_pass along with...
10277 (pass_data_sra_early): ...new pass_data instance and...
10278 (make_pass_sra_early): ...new function.
10279 (pass_sra): Convert from a global struct to a subclass of
10280 gimple_opt_pass along with...
10281 (pass_data_sra): ...new pass_data instance and...
10282 (make_pass_sra): ...new function.
10283 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10284 gimple_opt_pass along with...
10285 (pass_data_early_ipa_sra): ...new pass_data instance and...
10286 (make_pass_early_ipa_sra): ...new function.
10287 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10288 subclass of gimple_opt_pass along with...
10289 (pass_data_ccp): ...new pass_data instance and...
10290 (make_pass_ccp): ...new function.
10291 (pass_fold_builtins): Convert from a global struct to a subclass of
10292 gimple_opt_pass along with...
10293 (pass_data_fold_builtins): ...new pass_data instance and...
10294 (make_pass_fold_builtins): ...new function.
10295 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10296 subclass of gimple_opt_pass along with...
10297 (pass_data_copy_prop): ...new pass_data instance and...
10298 (make_pass_copy_prop): ...new function.
10299 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10300 global struct to a subclass of gimple_opt_pass along with...
10301 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10302 (make_pass_rename_ssa_copies): ...new function.
10303 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10304 subclass of gimple_opt_pass along with...
10305 (pass_data_dce): ...new pass_data instance and...
10306 (make_pass_dce): ...new function.
10307 (pass_dce_loop): Convert from a global struct to a subclass of
10308 gimple_opt_pass along with...
10309 (pass_data_dce_loop): ...new pass_data instance and...
10310 (make_pass_dce_loop): ...new function.
10311 (pass_cd_dce): Convert from a global struct to a subclass of
10312 gimple_opt_pass along with...
10313 (pass_data_cd_dce): ...new pass_data instance and...
10314 (make_pass_cd_dce): ...new function.
10315 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10316 subclass of gimple_opt_pass along with...
10317 (pass_data_dominator): ...new pass_data instance and...
10318 (make_pass_dominator): ...new function.
10319 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10320 gimple_opt_pass along with...
10321 (pass_data_phi_only_cprop): ...new pass_data instance and...
10322 (make_pass_phi_only_cprop): ...new function.
10323 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10324 subclass of gimple_opt_pass along with...
10325 (pass_data_dse): ...new pass_data instance and...
10326 (make_pass_dse): ...new function.
10327 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10328 a subclass of gimple_opt_pass along with...
10329 (pass_data_forwprop): ...new pass_data instance and...
10330 (make_pass_forwprop): ...new function.
10331 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10332 struct to a subclass of gimple_opt_pass along with...
10333 (pass_data_tree_ifcombine): ...new pass_data instance and...
10334 (make_pass_tree_ifcombine): ...new function.
10335 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10336 subclass of gimple_opt_pass along with...
10337 (pass_data_ch): ...new pass_data instance and...
10338 (make_pass_ch): ...new function.
10339 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10340 subclass of gimple_opt_pass along with...
10341 (pass_data_tree_loop): ...new pass_data instance and...
10342 (make_pass_tree_loop): ...new function.
10343 (pass_tree_loop_init): Convert from a global struct to a subclass of
10344 gimple_opt_pass along with...
10345 (pass_data_tree_loop_init): ...new pass_data instance and...
10346 (make_pass_tree_loop_init): ...new function.
10347 (pass_lim): Convert from a global struct to a subclass of
10348 gimple_opt_pass along with...
10349 (pass_data_lim): ...new pass_data instance and...
10350 (make_pass_lim): ...new function.
10351 (pass_tree_unswitch): Convert from a global struct to a subclass of
10352 gimple_opt_pass along with...
10353 (pass_data_tree_unswitch): ...new pass_data instance and...
10354 (make_pass_tree_unswitch): ...new function.
10355 (pass_predcom): Convert from a global struct to a subclass of
10356 gimple_opt_pass along with...
10357 (pass_data_predcom): ...new pass_data instance and...
10358 (make_pass_predcom): ...new function.
10359 (pass_vectorize): Convert from a global struct to a subclass of
10360 gimple_opt_pass along with...
10361 (pass_data_vectorize): ...new pass_data instance and...
10362 (make_pass_vectorize): ...new function.
10363 (pass_graphite): Convert from a global struct to a subclass of
10364 gimple_opt_pass along with...
10365 (pass_data_graphite): ...new pass_data instance and...
10366 (make_pass_graphite): ...new function.
10367 (pass_graphite_transforms): Convert from a global struct to a subclass
10368 of gimple_opt_pass along with...
10369 (pass_data_graphite_transforms): ...new pass_data instance and...
10370 (make_pass_graphite_transforms): ...new function.
10371 (pass_check_data_deps): Convert from a global struct to a subclass of
10372 gimple_opt_pass along with...
10373 (pass_data_check_data_deps): ...new pass_data instance and...
10374 (make_pass_check_data_deps): ...new function.
10375 (pass_iv_canon): Convert from a global struct to a subclass of
10376 gimple_opt_pass along with...
10377 (pass_data_iv_canon): ...new pass_data instance and...
10378 (make_pass_iv_canon): ...new function.
10379 (pass_scev_cprop): Convert from a global struct to a subclass of
10380 gimple_opt_pass along with...
10381 (pass_data_scev_cprop): ...new pass_data instance and...
10382 (make_pass_scev_cprop): ...new function.
10383 (pass_record_bounds): Convert from a global struct to a subclass of
10384 gimple_opt_pass along with...
10385 (pass_data_record_bounds): ...new pass_data instance and...
10386 (make_pass_record_bounds): ...new function.
10387 (pass_complete_unroll): Convert from a global struct to a subclass of
10388 gimple_opt_pass along with...
10389 (pass_data_complete_unroll): ...new pass_data instance and...
10390 (make_pass_complete_unroll): ...new function.
10391 (pass_complete_unrolli): Convert from a global struct to a subclass of
10392 gimple_opt_pass along with...
10393 (pass_data_complete_unrolli): ...new pass_data instance and...
10394 (make_pass_complete_unrolli): ...new function.
10395 (pass_parallelize_loops): Convert from a global struct to a subclass
10396 of gimple_opt_pass along with...
10397 (pass_data_parallelize_loops): ...new pass_data instance and...
10398 (make_pass_parallelize_loops): ...new function.
10399 (pass_loop_prefetch): Convert from a global struct to a subclass of
10400 gimple_opt_pass along with...
10401 (pass_data_loop_prefetch): ...new pass_data instance and...
10402 (make_pass_loop_prefetch): ...new function.
10403 (pass_iv_optimize): Convert from a global struct to a subclass of
10404 gimple_opt_pass along with...
10405 (pass_data_iv_optimize): ...new pass_data instance and...
10406 (make_pass_iv_optimize): ...new function.
10407 (pass_tree_loop_done): Convert from a global struct to a subclass of
10408 gimple_opt_pass along with...
10409 (pass_data_tree_loop_done): ...new pass_data instance and...
10410 (make_pass_tree_loop_done): ...new function.
10411 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10412 struct to a subclass of gimple_opt_pass along with...
10413 (pass_data_cse_reciprocals): ...new pass_data instance and...
10414 (make_pass_cse_reciprocals): ...new function.
10415 (pass_cse_sincos): Convert from a global struct to a subclass of
10416 gimple_opt_pass along with...
10417 (pass_data_cse_sincos): ...new pass_data instance and...
10418 (make_pass_cse_sincos): ...new function.
10419 (pass_optimize_bswap): Convert from a global struct to a subclass of
10420 gimple_opt_pass along with...
10421 (pass_data_optimize_bswap): ...new pass_data instance and...
10422 (make_pass_optimize_bswap): ...new function.
10423 (pass_optimize_widening_mul): Convert from a global struct to a
10424 subclass of gimple_opt_pass along with...
10425 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10426 (make_pass_optimize_widening_mul): ...new function.
10427 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10428 subclass of gimple_opt_pass along with...
10429 (pass_data_phiopt): ...new pass_data instance and...
10430 (make_pass_phiopt): ...new function.
10431 (pass_cselim): Convert from a global struct to a subclass of
10432 gimple_opt_pass along with...
10433 (pass_data_cselim): ...new pass_data instance and...
10434 (make_pass_cselim): ...new function.
10435 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10436 subclass of gimple_opt_pass along with...
10437 (pass_data_phiprop): ...new pass_data instance and...
10438 (make_pass_phiprop): ...new function.
10439 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10440 subclass of gimple_opt_pass along with...
10441 (pass_data_pre): ...new pass_data instance and...
10442 (make_pass_pre): ...new function.
10443 (pass_fre): Convert from a global struct to a subclass of
10444 gimple_opt_pass along with...
10445 (pass_data_fre): ...new pass_data instance and...
10446 (make_pass_fre): ...new function.
10447 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10448 subclass of gimple_opt_pass along with...
10449 (pass_data_reassoc): ...new pass_data instance and...
10450 (make_pass_reassoc): ...new function.
10451 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10452 subclass of gimple_opt_pass along with...
10453 (pass_data_sink_code): ...new pass_data instance and...
10454 (make_pass_sink_code): ...new function.
10455 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10456 subclass of gimple_opt_pass along with...
10457 (pass_data_strlen): ...new pass_data instance and...
10458 (make_pass_strlen): ...new function.
10459 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10460 struct to a subclass of gimple_opt_pass along with...
10461 (pass_data_build_alias): ...new pass_data instance and...
10462 (make_pass_build_alias): ...new function.
10463 (pass_build_ealias): Convert from a global struct to a subclass of
10464 gimple_opt_pass along with...
10465 (pass_data_build_ealias): ...new pass_data instance and...
10466 (make_pass_build_ealias): ...new function.
10467 (pass_ipa_pta): Convert from a global struct to a subclass of
10468 simple_ipa_opt_pass along with...
10469 (pass_data_ipa_pta): ...new pass_data instance and...
10470 (make_pass_ipa_pta): ...new function.
10471 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10472 subclass of gimple_opt_pass along with...
10473 (pass_data_uncprop): ...new pass_data instance and...
10474 (make_pass_uncprop): ...new function.
10475 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10476 global struct to a subclass of gimple_opt_pass along with...
10477 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10478 (make_pass_late_warn_uninitialized): ...new function.
10479 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10480 to a subclass of gimple_opt_pass along with...
10481 (pass_data_init_datastructures): ...new pass_data instance and...
10482 (make_pass_init_datastructures): ...new function.
10483 (pass_early_warn_uninitialized): Convert from a global struct to a
10484 subclass of gimple_opt_pass along with...
10485 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10486 (make_pass_early_warn_uninitialized): ...new function.
10487 (pass_update_address_taken): Convert from a global struct to a
10488 subclass of gimple_opt_pass along with...
10489 (pass_data_update_address_taken): ...new pass_data instance and...
10490 (make_pass_update_address_taken): ...new function.
10491 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10492 struct to a subclass of gimple_opt_pass along with...
10493 (pass_data_release_ssa_names): ...new pass_data instance and...
10494 (make_pass_release_ssa_names): ...new function.
10495 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10496 subclass of gimple_opt_pass along with...
10497 (pass_data_stdarg): ...new pass_data instance and...
10498 (make_pass_stdarg): ...new function.
10499 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10500 global struct to a subclass of gimple_opt_pass along with...
10501 (pass_data_convert_switch): ...new pass_data instance and...
10502 (make_pass_convert_switch): ...new function.
10503 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10504 to a subclass of gimple_opt_pass along with...
10505 (pass_data_tail_recursion): ...new pass_data instance and...
10506 (make_pass_tail_recursion): ...new function.
10507 (pass_tail_calls): Convert from a global struct to a subclass of
10508 gimple_opt_pass along with...
10509 (pass_data_tail_calls): ...new pass_data instance and...
10510 (make_pass_tail_calls): ...new function.
10511 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10512 struct to a subclass of gimple_opt_pass along with...
10513 (pass_data_lower_vector): ...new pass_data instance and...
10514 (make_pass_lower_vector): ...new function.
10515 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10516 gimple_opt_pass along with...
10517 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10518 (make_pass_lower_vector_ssa): ...new function.
10519 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10520 to a subclass of gimple_opt_pass along with...
10521 (pass_data_slp_vectorize): ...new pass_data instance and...
10522 (make_pass_slp_vectorize): ...new function.
10523 (pass_ipa_increase_alignment): Convert from a global struct to a
10524 subclass of simple_ipa_opt_pass along with...
10525 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10526 (make_pass_ipa_increase_alignment): ...new function.
10527 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10528 gimple_opt_pass along with...
10529 (pass_data_vrp): ...new pass_data instance and...
10530 (make_pass_vrp): ...new function.
10531 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10532 subclass of simple_ipa_opt_pass along with...
10533 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10534 (make_pass_ipa_free_lang_data): ...new function.
10535 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10536 gimple_opt_pass along with...
10537 (pass_data_tsan): ...new pass_data instance and...
10538 (make_pass_tsan): ...new function.
10539 (pass_tsan_O0): Convert from a global struct to a subclass of
10540 gimple_opt_pass along with...
10541 (pass_data_tsan_O0): ...new pass_data instance and...
10542 (make_pass_tsan_O0): ...new function.
10543 * var-tracking.c (pass_variable_tracking): Convert from a global
10544 struct to a subclass of rtl_opt_pass along with...
10545 (pass_data_variable_tracking): ...new pass_data instance and...
10546 (make_pass_variable_tracking): ...new function.
10547 * web.c (pass_web): Convert from a global struct to a subclass of
10548 rtl_opt_pass along with...
10549 (pass_data_web): ...new pass_data instance and...
10550 (make_pass_web): ...new function.
10551 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10552 declaration with that of...
10553 (make_pass_mode_switch_use): ...new function.
10554 (pass_resolve_sw_modes): Replace declaration with that of...
10555 (make_pass_resolve_sw_modes): ...new function.
10556 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10557 from a global struct to a subclass of rtl_opt_pass along with...
10558 (pass_data_mode_switch_use): ...new pass_data instance and...
10559 (make_pass_mode_switch_use): ...new function.
10560 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10561 from a global struct to a subclass of rtl_opt_pass along with...
10562 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10563 (make_pass_resolve_sw_modes): ...new function.
10564 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10565 struct to a subclass of rtl_opt_pass along with...
10566 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10567 (make_pass_insert_vzeroupper): ...new function.
10568 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10569 global struct to a subclass of rtl_opt_pass along with...
10570 (pass_data_work_around_errata): ...new pass_data instance and...
10571 (make_pass_work_around_errata): ...new function.
10572 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10573 struct to a subclass of rtl_opt_pass along with...
10574 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10575 (make_pass_mips_machine_reorg2): ...new function.
10577 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10579 * passes.c (pass_manager::operator new): New.
10581 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10583 Handwritten part of conversion of passes to C++ classes.
10585 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10586 (toplev.o): Add dep on PASS_MANAGER_H.
10587 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10588 of early local pases to reflect this moving from a global to a
10589 member of gcc::pass_manager.
10590 (cgraph_add_new_function): Likewise.
10591 * lto-cgraph.c (lto_output_node): Update for conversion of
10592 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10593 * passes.c (opt_pass::clone): New.
10594 (opt_pass::gate): New.
10595 (opt_pass::execute): New.
10596 (opt_pass::opt_pass): New.
10597 (pass_manager::execute_early_local_passes): New.
10598 (pass_manager::execute_pass_mode_switching): new.
10599 (finish_optimization_passes): Convert to...
10600 (pass_manager::finish_optimization_passes): ...this.
10601 (finish_optimization_passes): Update for conversion of passes to
10603 (register_dump_files_1): Use has_gate since we cannot portably
10604 check a vtable entry against NULL.
10605 (dump_one_pass): Likewise.
10606 (ipa_write_summaries_2): Likewise.
10607 (ipa_write_optimization_summaries_1): Likewise.
10608 (ipa_read_summaries_1): Likewise.
10609 (ipa_read_optimization_summaries_1): Likewise.
10610 (execute_ipa_stmt_fixups): Likewise.
10611 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10612 pass-creation functions rather than wiring up globals, and
10613 storing the results in fields of pass_manager generated using
10614 pass-instances.def.
10615 (pass_manager::dump_profile_report): Update for conversion of
10616 passes to C++ classes.
10617 (pass_manager::execute_ipa_summary_passes): Likewise.
10618 (execute_one_ipa_transform_pass): Likewise.
10619 (execute_one_pass): Use has_gate and has_execute since we cannot
10620 portably check a vtable entry against NULL.
10621 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10622 (pass_manager): Use pass-instances.def to add fields for the
10623 various pass instances.
10624 * toplev.c (finalize): Update for move of
10625 finish_optimization_passes to a method of gcc::pass_manager.
10626 * toplev.h (finish_optimization_passes): Move to method of class
10628 * tree-pass.h (struct pass_data): New.
10629 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10630 (opt_pass::gate): Convert to virtual function.
10631 (opt_pass::~opt_pass): New.
10632 (opt_pass::clone): New.
10633 (opt_pass::execute): Convert to virtual function.
10634 (opt_pass::opt_pass): New.
10635 (opt_pass::ctxt_): new.
10636 (gimple_opt_pass): Convert to subclass of opt_pass.
10637 (gimple_opt_pass::gimple_opt_pass): New.
10638 (rtl_opt_pass): Convert to subclass of opt_pass.
10639 (rtl_opt_pass::rtl_opt_pass): New.
10640 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10641 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10642 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10643 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10644 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10645 invocation of pass_mode_switching to reflect this moving from a
10646 global to a member of gcc::pass_manager.
10647 (ix86_option_override): Rework how pass_insert_vzeroupper is
10648 added to the pass_manager to reflect autogenerated changes.
10649 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10651 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10653 PR rtl-optimization/57708
10654 * recog.c (peep2_find_free_register): Validate all regs in a
10657 2013-08-05 Jan Hubicka <jh@suse.cz>
10660 * cgraph.c (verify_cgraph_node): Accept local flags from other
10662 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10663 (function_and_variable_visibility): Likewise.
10664 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10666 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10668 * graph.c (init_graph_slim_pretty_print): Remove.
10669 (print_graph_cfg): Do not call it. Use local pretty printer.
10670 (start_graph_dump): Likewise.
10672 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10674 * gimple-pretty-print.c (buffer): Remove.
10675 (initialized): Likewise.
10676 (maybe_init_pretty_print): Likewise.
10677 (print_gimple_stmt): Do not call it. Use non-static local
10678 pretty_printer variable.
10679 (print_gimple_expr): Likewise.
10680 (print_gimple_seq): Likewise.
10681 (gimple_dump_bb): Likewise.
10683 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10685 * asan.c (asan_pp): Remove.
10686 (asan_pp_initialized): Likewise.
10687 (asan_pp_initialize): Likewise.
10688 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10689 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10690 (asan_add_global): Likewise.
10692 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10694 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10695 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10696 * pretty-print.c (pp_formatted_text_data): Likewise.
10697 (pp_indent): Rename from pp_base_indent.
10698 (pp_format): Rename from pp_base_format.
10699 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10700 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10701 (pp_flush): Rename from pp_base_flush.
10702 (pp_set_line_maximum_length): Rename from
10703 pp_base_set_line_maximum_length.
10704 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10705 (pp_set_prefix): Rename from pp_base_set_prefix.
10706 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10707 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10708 (pp_append_text): Rename from pp_base_append_text.
10709 (pp_formatted_text): Rename from pp_base_formatted_text.
10710 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10711 (pp_remaining_character_count_for_line): Rename from
10712 pp_base_remaining_character_count_for_line.
10713 (pp_newline): Rename from pp_base_newline.
10714 (pp_character): Rename from pp_base_character.
10715 (pp_string): Rename from pp_base_string.
10716 (pp_maybe_space): Rename from pp_base_maybe_space.
10717 * asan.c (asan_pp_string): Adjust.
10718 (asan_emit_stack_protection): Likewise.
10719 (asan_add_global): Likewise.
10720 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10721 * tree-mudflap.c (mf_varname_tree): Likewise.
10722 * tree-pretty-print.c (pp_tree_identifier): Rename from
10723 pp_base_tree_identifier.
10724 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10725 Declare as function.
10727 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10729 * pretty-print.h (pp_bar_bar): New.
10730 (pp_ampersand_ampersand): Likewise.
10731 (pp_less_equal): Likewise.
10732 (pp_greater_equal): Likewise.
10733 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10734 printer functions instead of pp_string or operators and punctuators.
10735 (dump_gimple_call): Likewise.
10736 (dump_gimple_omp_for): Likewise.
10737 (dump_gimple_transaction): Likewise.
10738 (dump_gimple_phi): Likewise.
10739 (pp_gimple_stmt_1): Likewise.
10740 * sched-vis.c (print_insn): Likewise.
10741 * tree-mudflap.c (mf_varname_tree): Likewise.
10742 * tree-pretty-print.c (dump_block_node): Likewise.
10743 (dump_generic_node): Likewise.
10745 2013-08-02 Jan Hubicka <jh@suse.cz>
10747 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10749 * lto-streamer-out.c (tree_is_indexable): Results decls and
10750 parm decls are not indexable.
10751 (DFS_write_tree_body): Do not follow args and results.
10752 (hash_tree): Likewise.
10753 (output_functions): Rearrange so struct function is needed
10754 only when real body is output; be able to also ouptut abstract
10755 functions; output DECL_ARGUMENTS and DECL_RESULT.
10756 (lto_output): When not in WPA, ale store abstract functions.
10757 (write_symbol): Do not care about RESULT_DECL.
10758 (output_symbol_p): Handle correctly sbtract decls.
10759 * lto-streamer-in.c (input_function): Rearrange so struct
10760 function can be NULL at entry; allow streaming of
10761 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10762 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10763 sanity check during LTO.
10764 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10765 RESULT_DECl and DECL_ARGUMENTS.
10766 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10769 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10771 * pretty-print.h (pp_underscore): New.
10773 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10774 printer functions instead of pp_character.
10775 (dump_binary_rhs): Likewise.
10776 (dump_ternary_rhs): Likewise.
10777 (dump_gimple_call_args): Likewise.
10778 (pp_points_to_solution): Likewise.
10779 (dump_gimple_call): Likewise.
10780 (dump_gimple_switch): Likewise.
10781 (dump_gimple_cond): Likewise.
10782 (dump_gimple_bind): Likewise.
10783 (dump_gimple_try): Likewise.
10784 (dump_gimple_omp_for): Likewise.
10785 (dump_gimple_omp_continue): Likewise.
10786 (dump_gimple_omp_single): Likewise.
10787 (dump_gimple_omp_sections): Likewise.
10788 (dump_gimple_omp_block): Likewise.
10789 (dump_gimple_omp_critical): Likewise.
10790 (dump_gimple_transaction): Likewise.
10791 (dump_gimple_asm): Likewise.
10792 (dump_gimple_phi): Likewise.
10793 (dump_gimple_omp_parallel): Likewise.
10794 (dump_gimple_omp_task): Likewise.
10795 (dump_gimple_omp_atomic_load): Likewise.
10796 (dump_gimple_omp_atomic_store): Likewise.
10797 (dump_gimple_mem_ops): Likewise.
10798 (pp_gimple_stmt_1): Likewise.
10799 (pp_cfg_jump): Likewise.
10800 (dump_implicit_edges): Likewise.
10801 (gimple_dump_bb_for_graph): Likewise.
10802 * graph.c (draw_cfg_node): Likewise.
10803 * langhooks.c (lhd_print_error_function): Likewise.
10804 * sched-vis.c (print_exp): Likewise.
10805 (print_value): Likewise.
10806 (print_pattern): Likewise.
10807 (print_insn): Likewise.
10808 (rtl_dump_bb_for_graph): Likewise.
10809 * tree-pretty-print.c (dump_function_declaration): Likewise.
10810 (dump_array_domain): Likewise.
10811 (dump_omp_clause): Likewise.
10812 (dump_location): Likewise.
10813 (dump_generic_node): Likewise.
10814 (print_struct_decl): Likewise.
10815 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10817 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10819 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10820 candidate table when replacing a candidate statement.
10821 (replace_rhs_if_not_dup): Likewise.
10822 (replace_one_candidate): Likewise.
10824 2013-08-02 Jan Hubicka <jh@suse.cz>
10825 Martin Liska <marxin.liska@gmail.com>
10827 * cgraphunit.c (add_new_function): Fix logic when adding from
10829 (assemble_thunk): Rename to ...
10830 (expand_thunk); .. this one; export; get it working with
10831 general functions; make produced gimple valid.
10832 * cgraph.h (expand_thunk): Declare.
10834 2013-08-02 Jan Hubicka <jh@suse.cz>
10836 * ipa-cp.c (gather_context_independent_values): Use
10837 ipa_get_param_move_cost.
10838 (get_replacement_map): Remove PARAM; move parameter folding
10840 (create_specialized_node): Update.
10841 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10842 assert that we have gimple body; update move_cost.
10843 (count_formal_params): Assert that we have gimple body.
10844 (ipa_dump_param): New function.
10845 (ipa_alloc_node_params): Break out from ...
10846 (ipa_initialize_node_params): ... here.
10847 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10848 (ipa_write_node_info): Stream move costs.
10849 (ipa_read_node_info): Read move costs.
10850 (ipa_update_after_lto_read): Do not recompute node params.
10851 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10852 (ipa_get_param): Check we are not in WPA.
10853 (ipa_get_param_move_cost): New.
10854 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10855 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10856 parm numbers to be present.
10858 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10860 PR rtl-optimization/58048
10861 * lra-constraints.c (process_alt_operands): Don't check asm
10862 operand on register.
10864 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10866 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10867 the implied StoreLoad barrier for atomic operations if before.
10869 2013-08-02 Jan Hubicka <jh@suse.cz>
10870 Martin Liska <marxin.liska@gmail.com>
10872 * cgraph.c (cgraph_function_body_availability): Do not check
10874 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10875 symtab_node_availability): Declare.
10876 * ipa.c (can_replace_by_local_alias): New.
10877 (function_and_variable_visibility): Use it.
10878 * symtab.c (symtab_for_node_and_aliases,
10879 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
10881 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10883 PR rtl-optimization/57963
10884 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
10885 (lra_constraints): Use them.
10887 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10889 * config/arm/types.md (define_attr "type"): Add "load_acq"
10891 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10893 (cortex_a53_store1): Likewise.
10895 2013-08-01 Jan Hubicka <jh@suse.cz>
10897 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10898 partitions are not needed.
10900 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
10902 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10903 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10904 MAYBE_NON_Q_CLASS_P where appropriate.
10906 2013-08-01 Jan Hubicka <jh@suse.cz>
10908 * cgraph.h (release_function_body): Declare.
10909 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10910 of unused delcarations.
10912 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10914 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10915 RTL form when subtracting a constant.
10917 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10919 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10920 Generate canonical plus rtx with negated immediate instead of minus
10922 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10924 2013-08-01 Jan Hubicka <jh@suse.cz>
10926 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10927 (cgraph_release_function_body): Likewise.
10928 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10929 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10930 to used_as_abstract_origin.
10931 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10932 symbols used as abstract origins.
10933 * cgraphunit.c (analyze_functions): Update.
10934 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10935 used_as_abstract_origin.
10936 * tree-inline.c (tree_function_versioning): Update
10937 used_as_abstract_origin; be ready for DECL_RESULT and
10938 DECL_ARGUMENTS to be NULL.
10940 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
10941 for abstract functions.
10942 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
10945 2013-08-01 Jan Hubicka <jh@suse.cz>
10947 * profile.c (compute_value_histograms): Fix thinko.
10949 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10951 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
10952 aarch-common-protos.h to extra_headers.
10953 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10954 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
10955 * config/aarch64/t-aarch64 (aarch-common.o): Define.
10957 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10959 * config/aarch64/aarch64.md (define_attr "type"): Delete.
10960 Include "../arm/types.md". Define "type" attribute for all patterns.
10961 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
10964 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
10966 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
10967 to support power8 load fusion.
10968 (fusion_gpr_mem_load): Likewise.
10970 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
10972 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
10973 declarations for power8 load fusion.
10974 (emit_fusion_gpr_load): Likewise.
10976 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
10977 tuning for power8, turn on fusion mode by default. Turn on sign
10978 extending fusion mode if normal fusion mode is on, and we are at
10980 (fusion_gpr_load_p): New function, return true if we can fuse an
10981 addis instruction with a dependent load to a GPR.
10982 (emit_fusion_gpr_load): Emit the instructions for power8 load
10985 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
10986 (VSX load fusion peepholes): New peepholes to fuse together an
10987 addi instruction with a VSX load instruction.
10989 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
10990 peepholes to fuse an addis instruction with a load to a GPR base
10991 register. If we are supporting sign extending fusions, convert
10992 sign extending loads to zero extending loads and add an explicit
10995 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10997 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
10998 aarch-common-protos.h to extra_headers.
10999 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11000 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11001 (arm_early_store_addr_dep): Likewise.
11002 (arm_no_early_alu_shift_dep): Likewise.
11003 (arm_no_early_alu_shift_value_dep): Likewise.
11004 (arm_no_early_mul_dep): Likewise.
11005 (arm_no_early_store_addr_dep): Likewise.
11006 (arm_mac_accumulator_is_mul_result): Likewise.
11007 (arm_mac_accumulator_is_result): Likewise.
11008 * config/arm/aarch-common.c: ... here. New file.
11009 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11011 (arm_early_store_addr_dep): Likewise.
11012 (arm_no_early_alu_shift_dep): Likewise.
11013 (arm_no_early_alu_shift_value_dep): Likewise.
11014 (arm_no_early_mul_dep): Likewise.
11015 (arm_no_early_store_addr_dep): Likewise.
11016 (arm_mac_accumulator_is_mul_result): Likewise.
11017 (arm_mac_accumulator_is_result): Likewise.
11018 * config/arm/aarch-common-protos.h: ... here. New file.
11019 * config/arm/t-arm (aarch-common.o): Define.
11021 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11023 * config/arm/arm.md: Include new file "types.md".
11024 (define_attr "type"): Move from here to ...
11025 (define_attr "mul32"): Likewise.
11026 (define_attr "mul64"): Likewise.
11027 * config/arm/types.md: ... here. New file.
11029 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11031 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11032 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11034 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11036 * gen-pass-instances.awk: Fix offset of substr().
11038 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11040 * Makefile.in (pass-instances.def): New.
11041 (passes.o): Replace dependency on passes.def with one on
11044 * gen-pass-instances.awk: New.
11046 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11047 rather than passes.def, updating local definition of NEXT_PASS
11048 macro to add an extra NUM parameter (currently unused).
11050 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11052 * Makefile.in (PASS_MANAGER_H): New.
11053 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11054 (passes.o): Likewise.
11055 (statistics.o): Likewise.
11056 (cgraphunit.o): Likewise.
11057 (context.o): Depend on PASS_MANAGER_H.
11059 * pass_manager.h: New.
11061 * cgraphunit.c (cgraph_add_new_function): Update for moves
11062 of globals to fields of pass_manager.
11063 (analyze_function): Likewise.
11064 (expand_function): Likewise.
11065 (ipa_passes): Likewise.
11066 (compile): Likewise.
11068 * context.c (context::context): New.
11069 * context.h (context::context): New.
11070 (context::get_passes): New.
11071 (context::passes_): New.
11073 * lto-cgraph.c (input_node): Update for moves of globals to
11074 fields of pass_manager.
11076 * passes.c (all_passes): Remove, in favor of a field of the
11077 same name within the new class pass_manager.
11078 (all_small_ipa_passes): Likewise.
11079 (all_lowering_passes): Likewise.
11080 (all_regular_ipa_passes): Likewise.
11081 (all_late_ipa_passes): Likewise.
11082 (all_lto_gen_passes): Likewise.
11083 (passes_by_id): Likewise.
11084 (passes_by_id_size): Likewise.
11085 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11086 the new class pass_manager.
11087 (set_pass_for_id): Convert to...
11088 (pass_manager::set_pass_for_id): ...method.
11089 (get_pass_for_id): Convert to...
11090 (pass_manager::get_pass_for_id): ...method.
11091 (register_one_dump_file): Move body of implementation into...
11092 (pass_manager::register_one_dump_file): ...here.
11093 (register_dump_files_1): Convert to...
11094 (pass_manager::register_dump_files_1): ...method.
11095 (register_dump_files): Convert to...
11096 (pass_manager::register_dump_files): ...method.
11097 (create_pass_tab): Update for moves of globals to fields of
11099 (dump_passes): Move body of implementation into...
11100 (pass_manager::dump_passes): ...here.
11101 (register_pass): Move body of implementation into...
11102 (pass_manager::register_pass): ...here.
11103 (init_optimization_passes): Convert into...
11104 (pass_manager::pass_manager): ...constructor for new
11105 pass_manager class, and initialize the pass_lists array.
11106 (check_profile_consistency): Update for moves of globals to
11107 fields of pass_manager.
11108 (dump_profile_report): Move body of implementation into...
11109 (pass_manager::dump_profile_report): ...here.
11110 (ipa_write_summaries_1): Update for moves of pass lists from
11111 being globals to fields of pass_manager.
11112 (ipa_write_optimization_summaries): Likewise.
11113 (ipa_read_summaries): Likewise.
11114 (ipa_read_optimization_summaries): Likewise.
11115 (execute_all_ipa_stmt_fixups): Likewise.
11117 * statistics.c (statistics_fini): Update for moves of globals to
11118 fields of pass_manager.
11120 * toplev.c (general_init): Replace call to
11121 init_optimization_passes with construction of the pass_manager
11124 * tree-pass.h (all_passes): Remove, in favor of a field of the
11125 same name within the new class pass_manager.
11126 (all_small_ipa_passes): Likewise.
11127 (all_lowering_passes): Likewise.
11128 (all_regular_ipa_passes): Likewise.
11129 (all_lto_gen_passes): Likewise.
11130 (all_late_ipa_passes): Likewise.
11131 (passes_by_id): Likewise.
11132 (passes_by_id_size): Likewise.
11133 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11134 the new class pass_manager.
11135 (get_pass_for_id): Remove.
11137 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11139 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11142 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11144 * arm.md (mulhi3): New expand pattern.
11146 2013-07-30 Jan Hubicka <jh@suse.cz>
11147 Martin Liska <marxin.liska@gmail.com>
11149 * profile.c (compute_value_histograms): Do not ICE when
11150 there is mismatch only on some counters.
11152 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11154 PR rtl-optimization/57637
11155 * function.c (move_insn_for_shrink_wrap): Also check the
11156 GEN set of the LIVE problem for the liveness analysis
11157 if it exists, otherwise give up.
11159 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11161 PR tree-optimization/57993
11162 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11163 replaced statement in the candidate table.
11164 (phi_add_costs): Return infinite cost when the hidden basis does
11165 not dominate all phis on which the candidate is dependent.
11166 (replace_one_candidate): Record replaced statement in the
11169 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11171 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11172 (ashlv2si3): New expander.
11173 (*ashlv2si3_i): New define_insn_and_split.
11174 * predicates.md (float_operation): Allow patterns with three
11175 basic sub-patterns.
11177 PR rtl-optimization/58021
11178 * mode-switching.c (create_pre_exit): Always split off preceding
11179 insns if we are not at the basic block head.
11181 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11183 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11184 (UCLIBC_DYNAMIC_LINKER): New macro.
11185 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11187 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11188 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11190 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11191 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11192 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11193 for SF and DF modes. Use ieee_quad_format for TF mode.
11194 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11195 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11196 (mips_option_override): Handle `-mnan=legacy'.
11197 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11198 `-mabs=2008' and `-mnan=2008'.
11199 (OPTION_DEFAULT_SPECS): Add "nan" default.
11200 (ASM_SPEC): Handle `-mnan='.
11201 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11202 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11203 comment accordingly.
11204 (neg<mode>2): Likewise.
11205 * config/mips/mips.opt (mabs, mnan): New options.
11206 * doc/install.texi (Configuration): Document `--with-nan=' option.
11207 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11209 (MIPS Options): Document them.
11210 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11211 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11212 * configure: Regenerate.
11213 * config.in: Regenerate.
11215 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11217 * config/i386/i386.md (float post-reload splitters): Do not check
11218 for subregs of SSE registers.
11220 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11221 H.J. Lu <hongjiu.lu@intel.com>
11225 * config/i386/i386.md (post-reload splitter
11226 to avoid partial SSE reg dependency stalls): New pattern.
11228 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11230 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11232 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11234 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11235 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11237 2013-07-26 Cary Coutant <ccoutant@google.com>
11239 * dwarf2out.c (die_checksum_ordered): Don't include template
11240 instantiations in signature.
11241 (is_template_parameter): New function.
11242 (is_template_instantiation): New function.
11243 (generate_skeleton_bottom_up): Don't include template instantiations
11245 (generate_skeleton): Likewise.
11246 (break_out_comdat_types): Move recursive call to break out nested
11248 (prune_unused_types_mark_generic_parms_dies): Call
11249 is_template_parameter.
11251 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11253 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11254 uses vector registers.
11255 * config/aarch64/iterators.md: Add attributes rtn and vas.
11257 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11258 Richard Earnshaw <richard.earnshaw@arm.com>
11260 * combine.c (simplify_comparison): Re-canonicalize operands
11262 * config/arm/arm.md (movcond_addsi): New splitter.
11264 2013-07-25 Sterling Augustine <saugustine@google.com>
11266 * dwarf2out.c (size_of_pubnames): Move code to...
11267 (include_pubname_in_output): ...here. New.
11268 (want_pubnames): Rearrange.
11269 (output_pubnames): Call include_pubname_in_output. Move assertion.
11271 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11273 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11275 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11278 * doc/extend.texi: Remove obsolete builtins. Fix
11279 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11281 2013-07-25 Jan Hubicka <jh@suse.cz>
11283 * cgraph.c (release_function_body): Break out from ...
11284 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11285 and DECL_ARGUMENTS.
11286 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11287 old_tree in the map.
11288 (create_specialized_node): Update.
11289 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11291 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11292 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11293 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11294 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11297 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11299 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11300 addsi3_carryin_alt2_<optab>): Correct output template.
11302 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11304 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11305 Adjust for arm_restrict_it.
11306 Remove trailing whitespace.
11308 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11310 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11311 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11313 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11315 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11317 PR rtl-optimization/57960
11318 * lra-constraints.c (process_alt_operands): Use the right mode
11319 when checking strict_low.
11321 2013-07-25 Jan Hubicka <jh@suse.cz>
11323 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11324 * cgraph.c (cgraph_remove_node): Do not release function body
11325 when in cgraph streaming.
11326 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11327 in other partitions are not considered reachable; fix handling of
11330 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11332 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11334 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11339 * config/arm/arm.md ("*sibcall_insn): Replace use of
11340 Ss with US. Adjust output for v5 and v4t.
11341 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11343 * config/arm/constraints.md ("Ss"): Rename to US.
11345 2013-07-25 Terry Guo <terry.guo@arm.com>
11347 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11348 shift_add/shift_sub0/shift_sub1 RTXs.
11350 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11351 Anton Blanchard <anton@au1.ibm.com>
11353 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11354 (altivec_vpks<VI_char>ss): Likewise.
11355 (altivec_vpks<VI_char>us): Likewise.
11356 (altivec_vpku<VI_char>us): Likewise.
11357 (altivec_vpku<VI_char>um): Likewise.
11359 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11361 Introduce context class.
11363 * Makefile.in (CONTEXT_H): New.
11364 (OBJS): Add context.o.
11365 (toplev.o): Add CONTEXT_H to dependencies.
11368 * toplev.c (general_init): Create the singleton gcc::context instance.
11374 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11376 PR rtl-optimization/57968
11377 * mode-switching.c (create_pre_exit): Allow instructions that
11378 don't set a return register to need a non-exit mode.
11380 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11381 Anton Blanchard <anton@au1.ibm.com>
11383 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11384 operands to vperm for little endian.
11385 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11386 of lvsl to create the control mask for a vperm for little endian.
11388 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11389 Anton Blanchard <anton@au1.ibm.com>
11391 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11392 two operands for little-endian.
11394 2013-07-23 Steve Ellcey <sellcey@mips.com>
11396 * config/mips/mips.c (mips_case_values_threshold): New.
11397 (TARGET_CASE_VALUES_THRESHOLD): Define.
11399 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11400 Anton Blanchard <anton@au1.ibm.com>
11402 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11403 selection of field for vector splat in little endian mode.
11405 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11407 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11408 expanders to rs6000.md.
11409 (ior<mode>3): Likewise.
11410 (and<mode>3): Likewise.
11411 (one_cmpl<mode>2): Likewise.
11412 (nor<mode>3): Likewise.
11413 (andc<mode>3): Likewise.
11414 (eqv<mode>3): Likewise.
11415 (nand<mode>3): Likewise.
11416 (orc<mode>3): Likewise.
11418 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11421 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11422 to split multi-word logical operations.
11423 (rs6000_split_logical_di): Likewise.
11424 (rs6000_split_logical): Likewise.
11426 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11427 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11428 and allow TImode operations in 32-bit.
11429 (vsx_and<mode>3_64bit): Likewise.
11430 (vsx_ior<mode>3_32bit): Likewise.
11431 (vsx_ior<mode>3_64bit): Likewise.
11432 (vsx_xor<mode>3_32bit): Likewise.
11433 (vsx_xor<mode>3_64bit): Likewise.
11434 (vsx_one_cmpl<mode>2_32bit): Likewise.
11435 (vsx_one_cmpl<mode>2_64bit): Likewise.
11436 (vsx_nor<mode>3_32bit): Likewise.
11437 (vsx_nor<mode>3_64bit): Likewise.
11438 (vsx_andc<mode>3_32bit): Likewise.
11439 (vsx_andc<mode>3_64bit): Likewise.
11440 (vsx_eqv<mode>3_32bit): Likewise.
11441 (vsx_eqv<mode>3_64bit): Likewise.
11442 (vsx_nand<mode>3_32bit): Likewise.
11443 (vsx_nand<mode>3_64bit): Likewise.
11444 (vsx_orc<mode>3_32bit): Likewise.
11445 (vsx_orc<mode>3_64bit): Likewise.
11447 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11448 logical types in GPRs.
11450 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11451 logical insns to rs6000.md, and allow TImode operations in
11453 (altivec_ior<mode>3): Likewise.
11454 (altivec_xor<mode>3): Likewise.
11455 (altivec_one_cmpl<mode>2): Likewise.
11456 (altivec_nor<mode>3): Likewise.
11457 (altivec_andc<mode>3): Likewise.
11459 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11460 attributes for moving the 128-bit logical operations into
11462 (BOOL_REGS_OUTPUT): Likewise.
11463 (BOOL_REGS_OP1): Likewise.
11464 (BOOL_REGS_OP2): Likewise.
11465 (BOOL_REGS_UNARY): Likewise.
11466 (BOOL_REGS_AND_CR0): Likewise.
11467 (one_cmpl<mode>2): Add support for DI logical operations on
11468 32-bit, splitting the operations to 32-bit.
11469 (anddi3): Likewise.
11470 (iordi3): Likewise.
11471 (xordi3): Likewise.
11472 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11473 changes to combine the 32/64-bit code, allow logical operations on
11474 TI mode in 32-bit, and to use similar match_operator patterns like
11475 scalar mode uses. Combine the Altivec and VSX code for logical
11476 operations, and move it here.
11477 (ior<mode>3, 128-bit types): Likewise.
11478 (xor<mode>3, 128-bit types): Likewise.
11479 (one_cmpl<mode>3, 128-bit types): Likewise.
11480 (nor<mode>3, 128-bit types): Likewise.
11481 (andc<mode>3, 128-bit types): Likewise.
11482 (eqv<mode>3, 128-bit types): Likewise.
11483 (nand<mode>3, 128-bit types): Likewise.
11484 (orc<mode>3, 128-bit types): Likewise.
11485 (and<mode>3_internal): Likewise.
11486 (bool<mode>3_internal): Likewise.
11487 (boolc<mode>3_internal1): Likewise.
11488 (boolc<mode>3_internal2): Likewise.
11489 (boolcc<mode>3_internal1): Likewise.
11490 (boolcc<mode>3_internal2): Likewise.
11491 (eqv<mode>3_internal1): Likewise.
11492 (eqv<mode>3_internal2): Likewise.
11493 (one_cmpl1<mode>3_internal): Likewise.
11495 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11497 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11498 Rename flag_stack_usage to flag_stack_usage_info.
11500 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11502 * config/microblaze/sync.md: New file.
11503 * config/microblaze/microblaze.md: Include sync.md
11504 * config/microblaze/microblaze.c: Add print_operand 'y'.
11505 * config/microblaze/constraints.md: Add memory_contraint
11506 'Q' which is a single register.
11508 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11510 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11512 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11514 * reload.c (find_reloads): Exit loop once we find this operand
11515 cannot be reloaded somehow for this alternative.
11517 * reload.c (find_reloads): Exit loop once we find a hard register.
11519 * rtlanal.c (computed_jump_p): Exit loop once we find label
11522 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11524 * cfgloopmanip.c (remove_path): Exit loop after setting
11527 * gensupport.c (subst_dup): Avoid loop if code is not
11528 MATCH_DUP nor MATCH_OP_DUP.
11530 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11532 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11534 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11536 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11537 true for SP_REGNUM if mode == ptr_mode.
11538 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11539 with value R0_REGNUM + 31.
11541 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11543 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11544 pad pointer-typed argument downward.
11546 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11548 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11549 and __ILP32__ when the ILP32 model is in use.
11551 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11553 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11554 (aarch64_load_symref_appropriately): In the case of
11555 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11556 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11557 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11558 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11559 if 'mode' doesn't equal to 'ptr_mode'.
11560 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11561 'vcall_offset'; change to call aarch64_emit_move differently depending
11562 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11563 to calculate the upper bound of 'vcall_offset'.
11564 (aarch64_cannot_force_const_mem): Change to also return true if
11566 (aarch64_legitimize_reload_address): In the case of large
11567 displacements, add new local variable 'xmode' and an assertion
11568 based on it; change to use 'xmode' to generate the new rtx and
11570 (aarch64_asm_trampoline_template): Change to generate the template
11571 differently depending on TARGET_ILP32 or not; change to use
11572 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11573 (aarch64_trampoline_size): Removed.
11574 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11575 and replace immediate literals with it. Change to use 'ptr_mode'
11576 instead of 'DImode' and call convert_memory_address if the mode
11577 of 'fnaddr' doesn't equal to 'ptr_mode'.
11578 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11580 (aarch64_elf_asm_destructor): Likewise.
11581 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11582 on TARGET_ILP32 instead of aarch64_trampoline_size.
11583 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11584 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11585 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11586 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11587 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11588 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11589 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11590 depending on the value of 'mode'.
11591 (add_losym_<mode>): New.
11592 (ldr_got_small_<mode>): New, based on ldr_got_small.
11593 (ldr_got_small): Remove.
11594 (ldr_got_small_sidi): New.
11595 * config/aarch64/iterators.md (P): New.
11596 (PTR): Change to 'ptr_mode' in the condition.
11598 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11600 * config.gcc (aarch64*-*-*): Support --with-abi.
11601 (aarch64*-*-elf): Support --with-multilib-list.
11602 (aarch64*-*-linux*): Likewise.
11603 (supported_defaults): Add abi to aarch64*-*-*.
11604 * configure.ac: Mention AArch64 for --with-multilib-list.
11605 * configure: Re-generated.
11606 * config/aarch64/biarchilp32.h: New file.
11607 * config/aarch64/biarchlp64.h: New file.
11608 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11610 (MULTILIB_DEFAULTS): Ditto.
11611 (DRIVER_SELF_SPECS): Ditto.
11612 (ASM_SPEC): Update to also substitute -mabi.
11613 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11614 file whose name depends on -mabi= and -mbig-endian.
11615 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11617 (POINTER_SIZE): New define.
11618 (POINTERS_EXTEND_UNSIGNED): Ditto.
11619 (enum aarch64_abi_type): New enumeration tag.
11620 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11621 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11622 (TARGET_ILP32): New define.
11623 * config/aarch64/aarch64.opt (mabi): New.
11624 (aarch64_abi): New.
11625 (ilp32, lp64): New values for -mabi.
11626 * config/aarch64/t-aarch64 (comma): New define.
11627 (MULTILIB_OPTIONS): Ditto.
11628 (MULTILIB_DIRNAMES): Ditto.
11629 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11630 * doc/invoke.texi: Document -mabi for AArch64.
11632 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11634 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11636 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11637 Anton Blanchard <anton@au1.ibm.com>
11639 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11640 endianness when selecting field to splat.
11642 2013-07-22 Eric Christopher <echristo@gmail.com>
11644 * dwarf2out.c (die_odr_checksum): New function to use
11645 CHECKSUM_ macros and ULEB128 for DIE tag.
11646 (generate_type_signature): Use.
11648 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11650 * config.gcc (sparc*-*-*): Accept leon3 processor.
11651 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11652 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11653 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11654 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11655 (mfix-ut699): Adjust comment.
11656 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11657 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11658 (CPP_CPU_SPEC): Likewise.
11659 (ASM_CPU_SPEC): Likewise.
11660 * config/sparc/sparc.c (leon3_cost): New constant.
11661 (sparc_option_override): Add leon3 support.
11662 (mem_ref): New function.
11663 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11664 (sparc_do_work_around_errata): Look into the instruction in the delay
11665 slot and adjust accordingly. Add fix for the data cache nullify issues
11666 of the UT699. Change insertion position for the NOP.
11667 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11668 (leon3_load): New reservation.
11669 (leon_store): Bump latency to 2.
11670 (grfpu): New automaton.
11671 (grfpu_alu): New unit.
11672 (grfpu_ds): Likewise.
11673 (leon_fp_alu): Adjust.
11674 (leon_fp_mult): Delete.
11675 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11676 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11677 * config/sparc/sparc.md (cpu): Add leon3.
11678 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11679 (swapsi): Likewise.
11680 (atomic_test_and_set): Likewise.
11681 (ldstub): Likewise.
11683 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11685 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11686 default for R5900 targets.
11687 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11688 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11689 * config/mips/mips.c (mips_option_override): Report an error for
11690 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11691 for -march=r5900 -mhard-float.
11693 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11695 * df-problems.c (can_move_insns_across): Exit loop once we
11696 find a non-fixed, non-global register.
11698 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11701 * omega.c (omega_eliminate_red): Break after setting red_found.
11702 (omega_problem_has_red_equations): Similarly after setting found.
11703 (omega_query_variable): Similarly after setting coupled.
11705 2013-07-22 Marek Polacek <polacek@redhat.com>
11707 * gimplify.c: Don't include gimple.h twice.
11709 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11711 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11712 instead of TARGET_THUMB1.
11713 (Pz): New constraint.
11714 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11716 (compare_negsi_si): Likewise.
11717 (compare_addsi2_op0): Likewise.
11718 (compare_addsi2_op1): Likewise.
11719 (addsi3_carryin_<optab>): Likewise.
11720 (addsi3_carryin_alt2_<optab>): Likewise.
11721 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11722 for arm_restrict_it.
11723 (subsi3_carryin): Likewise.
11724 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11725 (minmax_arithsi): Disable for arm_restrict_it.
11726 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11727 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11728 (satsi_<SAT:code>_shift): Likewise.
11729 (arm_shiftsi3): Add alternative for 16-bit encoding.
11730 (arm32_movhf): Disable for arm_restrict_it.
11731 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11732 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11734 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11736 * config/arm/arm.md (attribute "insn"): Delete.
11737 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11738 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11739 (not_shiftsi): Update for attribute change.
11740 (not_shiftsi_compare0): Likewise.
11741 (not_shiftsi_compare0_scratch): Likewise.
11742 (arm_one_cmplsi2): Likewise.
11743 (thumb1_one_cmplsi2): Likewise.
11744 (notsi_compare0): Likewise.
11745 (notsi_compare0_scratch): Likewise.
11746 (thumb1_movdi_insn): Likewise.
11747 (arm_movsi_insn): Likewise.
11748 (movhi_insn_arch4): Likewise.
11749 (movhi_bytes): Likewise.
11750 (arm_movqi_insn): Likewise.
11751 (thumb1_movqi_insn): Likewise.
11752 (arm32_movhf): Likewise.
11753 (thumb1_movhf): Likewise.
11754 (arm_movsf_soft_insn): Likewise.
11755 (thumb1_movsf_insn): Likewise.
11756 (thumb_movdf_insn): Likewise.
11757 (movsicc_insn): Likewise.
11758 (movsfcc_soft_insn): Likewise.
11759 (and_scc): Likewise.
11760 (cond_move): Likewise.
11761 (if_move_not): Likewise.
11762 (if_not_move): Likewise.
11763 (if_shift_move): Likewise.
11764 (if_move_shift): Likewise.
11765 (if_shift_shift): Likewise.
11766 (if_not_arith): Likewise.
11767 (if_arith_not): Likewise.
11768 (cond_move_not): Likewise.
11769 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11770 (neon_mov<mode>): Likewise.
11771 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11772 (thumb2_movsi_vfp): Likewise.
11773 (movsf_vfp): Likewise.
11774 (thumb2_movsf_vfp): Likewise.
11775 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11777 (cortexa7_older_only): Likewise.
11778 (cortexa7_younger): Likewise.
11779 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11780 (1020alu_shift_op): Likewise.
11781 (1020alu_shift_reg_op): Likewise.
11782 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11783 (alu_shift_op): Likewise.
11784 (alu_shift_reg_op): Likewise.
11785 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11786 (11_alu_shift_op): Likewise.
11787 (11_alu_shift_reg_op): Likewise.
11788 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11789 (9_alu_shift_reg_op): Likewise.
11790 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11792 (cortex_a15_alu_shift): Likewise.
11793 (cortex_a15_alu_shift_reg): Likewise.
11794 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11795 (cortex_a5_alu_shift): Likewise.
11796 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11798 (cortex_a53_alu_shift): Likewise.
11799 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11801 (cortex_a7_alu_reg): Likewise.
11802 (cortex_a7_alu_shift): Likewise.
11803 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11804 (cortex_a8_alu_shift): Likewise.
11805 (cortex_a8_alu_shift_reg): Likewise.
11806 (cortex_a8_mov): Likewise.
11807 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11808 (cortex_a9_dp_shift): Likewise.
11809 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11810 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11811 (cortex_r4_mov): Likewise.
11812 (cortex_r4_alu_shift): Likewise.
11813 (cortex_r4_alu_shift_reg): Likewise.
11814 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11815 (526_alu_shift_op): Likewise.
11816 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11817 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11818 (626te_alu_shift_op): Likewise.
11819 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11820 (726te_alu_op): Likewise.
11821 (726te_alu_shift_op): Likewise.
11822 (726te_alu_shift_reg_op): Likewise.
11823 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11824 (mp626_alu_shift_op): Likewise.
11825 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11826 (pj4_alu_e1_conds): Likewise.
11827 (pj4_alu): Likewise.
11828 (pj4_alu_conds): Likewise.
11829 (pj4_shift): Likewise.
11830 (pj4_shift_conds): Likewise.
11831 (pj4_alu_shift): Likewise.
11832 (pj4_alu_shift_conds): Likewise.
11834 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11836 * config/arm/predicates.md (shiftable_operator_strict_it):
11838 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11839 Disable cond_exec version for arm_restrict_it.
11840 (thumb2_smaxsi3): Convert to generate cond_exec.
11841 (thumb2_sminsi3): Likewise.
11842 (thumb32_umaxsi3): Likewise.
11843 (thumb2_uminsi3): Likewise.
11844 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11845 (thumb2_neg_abssi2): Likewise.
11846 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11847 (thumb2_movsicc_insn): Adjust alternatives.
11848 (thumb2_mov_negscc): Disable for arm_restrict_it.
11849 (thumb2_mov_negscc_strict_it): New pattern.
11850 (thumb2_mov_notscc_strict_it): New pattern.
11851 (thumb2_mov_notscc): Disable for arm_restrict_it.
11852 (thumb2_ior_scc): Likewise.
11853 (thumb2_ior_scc_strict_it): New pattern.
11854 (thumb2_cond_move): Adjust for arm_restrict_it.
11855 (thumb2_cond_arith): Disable for arm_restrict_it.
11856 (thumb2_cond_arith_strict_it): New pattern.
11857 (thumb2_cond_sub): Adjust for arm_restrict_it.
11858 (thumb2_movcond): Likewise.
11859 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11860 (thumb2_zero_extendhisi2_v6): Likewise.
11861 (thumb2_zero_extendqisi2_v6): Likewise.
11862 (orsi_notsi_si): Likewise.
11863 (orsi_not_shiftsi_si): Likewise.
11865 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
11867 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11868 instruction sequence is 1 byte shorter.
11870 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
11872 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11873 it is not needed after split.
11875 2013-07-20 Iain Sandoe <iain@codesourcery.com>
11878 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11879 second label for nonlocal goto receivers. Don't output pic base labels
11880 unless we're producing PIC; mark that action unreachable().
11881 (ix86_save_reg): If the function contains a nonlocal label, save the
11883 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
11884 * config/darwin.c (emitted_pic_label_num): New GTY.
11885 (update_pic_label_number_if_needed): New.
11886 (machopic_output_function_base_name): Adjust for nonlocal receiver
11888 (machopic_should_output_picbase_label): New.
11889 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11890 (nonlocal_goto_receiver): New insn and split.
11892 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11894 * config/aarch64/aarch64-builtins.c
11895 (aarch64_fold_builtin): Fold abs in all modes.
11896 * config/aarch64/aarch64-simd-builtins.def
11897 (abs): Enable for all modes.
11898 * config/aarch64/arm_neon.h
11899 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11900 (vabs_f64): Add missing intrinsic.
11902 2013-07-19 Ian Bolton <ian.bolton@arm.com>
11904 * config/aarch64/arm_neon.h (vabs_s64): New function
11906 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
11909 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11910 * config/avr/avr.md (adjust_len): Add `round'.
11911 * config/avr/avr-protos.h (avr_out_round): New prototype.
11912 (avr_out_plus): Add `out_label' argument.
11913 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11914 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11915 Handle the case where `insn' is just a pattern.
11916 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11917 (avr_out_round): New function.
11918 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11920 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
11922 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11923 Add check for flag_stack_usage to handle -fstack-usage support
11925 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
11927 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11928 interaction for new Power8 flags and VSX.
11930 2013-07-18 Sriraman Tallam <tmsriram@google.com>
11932 PR middle-end/57698
11933 * tree-inline.c (expand_call_inline): Emit errors during
11934 early_inlining only if optimization is not turned on.
11936 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11940 * passes.c (init_optimization_passes): Move the construction of
11941 the pass hierarchy into a new passes.def file.
11943 * Makefile.in (passes.o): Add dependency on passes.def.
11945 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11947 * passes.c (init_optimization_passes): Introduce macros for
11948 constructing the tree of passes (INSERT_PASSES_AFTER,
11949 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
11950 TERMINATE_PASS_LIST).
11952 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
11953 Wei Mi <wmi@google.com>
11955 PR rtl-optimization/57878
11956 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
11958 (reload_pseudo_compare_func): Check nregs first for reload
11961 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11963 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
11965 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
11967 * read-rtl.c (validate_const_int): Once an invalid character is
11968 seen, quit the loop.
11970 * gengtype.c (write_roots): Similarly once we find the "deletable"
11971 or "if_marked" option.
11973 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11975 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
11976 "xtab" and "sat". Move value "clz" from here to ...
11977 (attriubte "type"): ... here.
11978 (satsi_<SAT:code>): Delete "insn" attribute.
11979 (satsi_<SAT:code>_shift): Likewise.
11980 (arm_zero_extendqisi2addsi): Likewise.
11981 (arm_extendqisi2addsi): Likewise.
11982 (clzsi2): Update for attribute changes.
11983 (rbitsi2): Likewise.
11984 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
11986 (arm_usatsihi): Likewise.
11987 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11989 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11991 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
11992 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
11993 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
11994 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
11995 in alphabetical order.
11996 (attribute "core_cycles"): Update for attribute changes.
11997 (arm_addsi3): Likewise.
11998 (addsi3_compare0): Likewise.
11999 (addsi3_compare0_scratch): Likewise.
12000 (addsi3_compare_op1): Likewise.
12001 (addsi3_compare_op2): Likewise.
12002 (compare_addsi2_op0): Likewise.
12003 (compare_addsi2_op1): Likewise.
12004 (addsi3_carryin_shift_<optab>): Likewise.
12005 (subsi3_carryin_shift): Likewise.
12006 (rsbsi3_carryin_shift): Likewise.
12007 (arm_subsi3_insn): Likewise.
12008 (subsi3_compare0): Likewise.
12009 (subsi3_compare): Likewise.
12010 (arm_andsi3_insn): Likewise.
12011 (thumb1_andsi3_insn): Likewise.
12012 (andsi3_compare0): Likewise.
12013 (andsi3_compare0_scratch): Likewise.
12014 (zeroextractsi_compare0_scratch
12015 (andsi_not_shiftsi_si): Likewise.
12016 (iorsi3_insn): Likewise.
12017 (iorsi3_compare0): Likewise.
12018 (iorsi3_compare0_scratch): Likewise.
12019 (arm_xorsi3): Likewise.
12020 (thumb1_xorsi3_insn): Likewise.
12021 (xorsi3_compare0): Likewise.
12022 (xorsi3_compare0_scratch): Likewise.
12023 (satsi_<SAT:code>_shift): Likewise.
12025 (arm_shiftsi3): Likewise.
12026 (shiftsi3_compare0): Likewise.
12027 (not_shiftsi): Likewise.
12028 (not_shiftsi_compare0): Likewise.
12029 (not_shiftsi_compare0_scratch): Likewise.
12030 (arm_one_cmplsi2): Likewise.
12031 (thumb_one_complsi2): Likewise.
12032 (notsi_compare0): Likewise.
12033 (notsi_compare0_scratch): Likewise.
12034 (thumb1_zero_extendhisi2): Likewise.
12035 (arm_zero_extendhisi2): Likewise.
12036 (arm_zero_extendhisi2_v6): Likewise.
12037 (arm_zero_extendhisi2addsi): Likewise.
12038 (thumb1_zero_extendqisi2): Likewise.
12039 (thumb1_zero_extendqisi2_v6): Likewise.
12040 (arm_zero_extendqisi2): Likewise.
12041 (arm_zero_extendqisi2_v6): Likewise.
12042 (arm_zero_extendqisi2addsi): Likewise.
12043 (thumb1_extendhisi2): Likewise.
12044 (arm_extendhisi2): Likewise.
12045 (arm_extendhisi2_v6): Likewise.
12046 (arm_extendqisi): Likewise.
12047 (arm_extendqisi_v6): Likewise.
12048 (arm_extendqisi2addsi): Likewise.
12049 (thumb1_extendqisi2): Likewise.
12050 (thumb1_movdi_insn): Likewise.
12051 (arm_movsi_insn): Likewise.
12052 (movsi_compare0): Likewise.
12053 (movhi_insn_arch4): Likewise.
12054 (movhi_bytes): Likewise.
12055 (arm_movqi_insn): Likewise.
12056 (thumb1_movqi_insn): Likewise.
12057 (arm32_movhf): Likewise.
12058 (thumb1_movhf): Likewise.
12059 (arm_movsf_soft_insn): Likewise.
12060 (thumb1_movsf_insn): Likewise.
12061 (movdf_soft_insn): Likewise.
12062 (thumb_movdf_insn): Likewise.
12063 (arm_cmpsi_insn): Likewise.
12064 (cmpsi_shiftsi): Likewise.
12065 (cmpsi_shiftsi_swp): Likewise.
12066 (arm_cmpsi_negshiftsi_si): Likewise.
12067 (movsicc_insn): Likewise.
12068 (movsfcc_soft_insn): Likewise.
12069 (arith_shiftsi): Likewise.
12070 (arith_shiftsi_compare0
12071 (arith_shiftsi_compare0_scratch
12072 (sub_shiftsi): Likewise.
12073 (sub_shiftsi_compare0
12074 (sub_shiftsi_compare0_scratch
12075 (and_scc): Likewise.
12076 (cond_move): Likewise.
12077 (if_plus_move): Likewise.
12078 (if_move_plus): Likewise.
12079 (if_move_not): Likewise.
12080 (if_not_move): Likewise.
12081 (if_shift_move): Likewise.
12082 (if_move_shift): Likewise.
12083 (if_shift_shift): Likewise.
12084 (if_not_arith): Likewise.
12085 (if_arith_not): Likewise.
12086 (cond_move_not): Likewise.
12087 (thumb1_ashlsi3): Set type attribute.
12088 (thumb1_ashrsi3): Likewise.
12089 (thumb1_lshrsi3): Likewise.
12090 (thumb1_rotrsi3): Likewise.
12091 (shiftsi3_compare0_scratch): Likewise.
12092 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12093 (neon_mov<mode>): Likewise.
12094 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12096 (thumb2_movsi_insn): Likewise.
12097 (thumb2_cmpsi_neg_shiftsi): Likewise.
12098 (thumb2_extendqisi_v6): Likewise.
12099 (thumb2_zero_extendhisi2_v6): Likewise.
12100 (thumb2_zero_extendqisi2_v6): Likewise.
12101 (thumb2_shiftsi3_short): Likewise.
12102 (thumb2_addsi3_compare0_scratch): Likewise.
12103 (orsi_not_shiftsi_si): Likewise.
12104 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12105 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12107 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12108 (1020alu_shift_op): Likewise.
12109 (1020alu_shift_reg_op): Likewise.
12110 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12111 (alu_shift_op): Likewise.
12112 (alu_shift_reg_op): Likewise.
12113 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12114 (11_alu_shift_op): Likewise.
12115 (11_alu_shift_reg_op): Likewise.
12116 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12117 (9_alu_shift_reg_op): Likewise.
12118 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12120 (cortex_a15_alu_shift): Likewise.
12121 (cortex_a15_alu_shift_reg): Likewise.
12122 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12124 (cortex_a5_alu_shift): Likewise.
12125 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12127 (cortex_a53_alu_shift): Likewise.
12128 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12130 (cortex_a7_alu_reg): Likewise.
12131 (cortex_a7_alu_shift): Likewise.
12132 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12134 (cortex_a8_alu_shift): Likewise.
12135 (cortex_a8_alu_shift_reg): Likewise.
12136 (cortex_a8_mov): Likewise.
12137 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12138 (cortex_a9_dp_shift): Likewise.
12139 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12141 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12143 (cortex_r4_mov): Likewise.
12144 (cortex_r4_alu_shift): Likewise.
12145 (cortex_r4_alu_shift_reg): Likewise.
12146 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12147 (526_alu_shift_op): Likewise.
12148 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12149 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12150 (626te_alu_shift_op): Likewise.
12151 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12152 (726te_alu_op): Likewise.
12153 (726te_alu_shift_op): Likewise.
12154 (726te_alu_shift_reg_op): Likewise.
12155 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12156 (mp626_alu_shift_op): Likewise.
12157 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12158 (pj4_alu_e1_conds): Likewise.
12159 (pj4_alu): Likewise.
12160 (pj4_alu_conds): Likewise.
12161 (pj4_shift): Likewise.
12162 (pj4_shift_conds): Likewise.
12163 (pj4_alu_shift): Likewise.
12164 (pj4_alu_shift_conds): Likewise.
12165 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12167 (cortexa7_older_only): Likewise.
12168 (cortexa7_younger): Likewise.
12170 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12172 * ipa-pure-const.c (generate_summary): Rename to...
12173 (pure_const_generate_summary): ... this.
12175 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12177 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12179 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12182 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12185 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12187 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12188 enabled without -march=zEC12.
12189 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12192 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12194 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12195 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12196 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12197 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12198 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12199 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12200 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12202 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12203 and ISA_HAS_NMADD3_NMSUB3.
12204 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12205 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12206 (nmsub4<mode>, nmsub3<mode>): Likewise.
12207 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12209 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12211 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12212 TARGET_MIPS5400 checking.
12214 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12215 Peter Bergner <bergner@vnet.ibm.com>
12217 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12218 registers in the comment.
12219 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12220 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12221 rather than FIRST_PSEUDO_REGISTERS.
12223 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12225 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12226 enable extra ISA flags with TARGET_HTM.
12228 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12230 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12233 2013-07-15 Cong Hou <congh@google.com>
12235 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12236 in compare function for sorting.
12238 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12240 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12241 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12242 * config/rs6000/rs6000.opt: Add -mhtm option.
12243 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12244 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12245 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12246 __HTM__ if the HTM instructions are available.
12247 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12248 htm_spr_reg_operand): New define_predicates.
12249 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12250 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12252 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12253 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12254 HTM builtin functions.
12255 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12256 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12257 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12258 (rs6000_builtin_mask_calculate): Likewise.
12259 (rs6000_option_override_internal): Likewise.
12260 (bdesc_htm): Add new HTM builtin support.
12261 (htm_spr_num): New function.
12262 (htm_spr_regno): Likewise.
12263 (rs6000_htm_spr_icode): Likewise.
12264 (htm_expand_builtin): Likewise.
12265 (htm_init_builtins): Likewise.
12266 (rs6000_expand_builtin): Add support for HTM builtin functions.
12267 (rs6000_init_builtins): Likewise.
12268 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12270 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12271 (TARGET_HTM, MASK_HTM): Define macros.
12272 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12273 (FIXED_REGISTERS): Likewise.
12274 (CALL_USED_REGISTERS): Likewise.
12275 (CALL_REALLY_USED_REGISTERS): Likewise.
12276 (REG_ALLOC_ORDER): Likewise.
12277 (enum reg_class): Likewise.
12278 (REG_CLASS_NAMES): Likewise.
12279 (REG_CLASS_CONTENTS): Likewise.
12280 (REGISTER_NAMES): Likewise.
12281 (ADDITIONAL_REGISTER_NAMES): Likewise.
12282 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12283 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12284 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12285 * config/rs6000/htm.md: New file.
12286 * config/rs6000/htmintrin.h: New file.
12287 * config/rs6000/htmxlintrin.h: New file.
12289 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12291 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12292 Define SYMBOL_TINY_GOT, update comment.
12293 * config/aarch64/aarch64.c
12294 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12295 (aarch64_expand_mov_immediate): Likewise.
12296 (aarch64_print_operand): Likewise.
12297 (aarch64_classify_symbol): Likewise.
12298 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12299 (ldr_got_tiny): Define.
12301 2013-07-13 Tobias Grosser <tobias@grosser.es>
12303 PR tree-optimization/54094
12304 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12305 scheduling dimension for the parallelism check from the polyhedral
12306 information in the AST.
12307 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12310 2013-07-13 Jason Merrill <jason@redhat.com>
12312 * print-tree.c (debug_vec_tree): Use debug_raw.
12313 (debug_raw (vec<tree, va_gc> &)): New.
12314 (debug_raw (vec<tree, va_gc> *)): New.
12315 * tree.h: Declare them.
12317 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12319 * ifcvt.c (ifcvt_after_combine): New static variable.
12320 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12322 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12323 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12324 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12326 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12328 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12330 2013-07-12 Michael Matz <matz@suse.de>
12332 PR middle-end/55771
12333 * convert.c (convert_to_real): Reject non-float inner types.
12335 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12337 * config/aarch64/aarch64-protos.h
12338 (aarch64_simd_immediate_valid_for_move): Remove.
12339 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12340 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12342 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12344 2013-07-11 Steve Ellcey <sellcey@mips.com>
12346 * config/mips/mips.c (mips_conditional_register_usage): Do not
12347 use t[0-7] registers in MIPS16 mode when optimizing for size.
12349 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12351 * config/i386/i386.c (dispatch_function_versions): Fix array
12352 indexing of function_version_info to match actual_versions.
12354 2013-07-11 Teresa Johnson <tejohnson@google.com>
12356 * vec.h (struct va_gc): Move release out-of-line.
12357 (va_gc::release): Call ggc_free on released vec.
12359 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12361 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12362 Require GOT register as additional operand in UNSPEC.
12363 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12364 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12365 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12366 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12367 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12368 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12369 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12371 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12374 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12375 name seen by assembler/linker rather if available.
12377 2013-07-11 Andreas Schwab <schwab@suse.de>
12379 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12381 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12383 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12385 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12387 * doc/tm.texi.in: Move hook documentation to ....
12388 * target.def: ... here.
12390 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12391 text on @hook line.
12392 * doc/tm.texi: Regenerate.
12394 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12397 * doc/invoke.texi: Document -Wconditionally-supported.
12399 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12402 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12405 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12408 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12409 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12410 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12411 Remove duplicate devices.
12412 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12413 * config/avr/t-multilib: Regenerate.
12414 * config/avr/avr-tables.opt: Regenerate.
12415 * doc/avr-mmcu.texi: Regenerate.
12417 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12420 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12422 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12424 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12425 the cost of MULT when optimizing for size.
12427 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12429 * config/cr16/cr16-protos.h: Don't include target.h.
12431 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12433 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12434 adjust register size for TDmode and TFmode for VSX registers.
12436 2013-07-08 Kai Tietz <ktietz@redhat.com>
12439 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12440 hook_bool_const_tree_true.
12442 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12444 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12445 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12446 * config/s390/s390.md: Define FPR*_REGNUM constants.
12447 Fix FPR2_REGNUM constant (18 -> 17).
12448 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12449 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12450 ("trunc<BFP:mode><DFP_ALL:mode>2")
12451 ("trunc<DFP_ALL:mode><BFP:mode>2")
12452 ("*extend<BFP:mode><DFP_ALL:mode>2")
12453 ("*extend<DFP_ALL:mode><BFP:mode>2")
12454 ("extend<BFP:mode><DFP_ALL:mode>2")
12455 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12458 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12460 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12462 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12464 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12465 and cfun_fpr_bit_p to cfun_fpr_save_p.
12466 (s390_frame_area, s390_register_info, s390_frame_info)
12467 (s390_emit_prologue, s390_emit_epilogue)
12468 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12470 * config/s390/s390.h: Define *_REGNUM macros for floating point
12473 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12475 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12477 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12479 PR rtl-optimization/57786
12480 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12481 and break out of the loop when it is set to false.
12483 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12486 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12487 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12488 (const_int 63)) 0)).
12489 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12490 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12491 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12493 PR rtl-optimization/57829
12494 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12495 mask bits outside of mode are just sign-extension from mode to HWI.
12497 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12499 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12500 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12501 adjust_address instead of change_address to keep info about alignment.
12502 (emit_strmov): Remove.
12503 (emit_memmov): New function.
12504 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12505 (expand_movmem_epilogue): Likewise and return updated rtx for
12507 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12508 destination and source.
12509 (decide_alignment): Refactor, handle vector_loop.
12510 (ix86_expand_movmem): Likewise.
12511 (ix86_expand_setmem): Likewise.
12512 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12514 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12516 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12517 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12519 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12521 * config/i386/sse.md (sse_movlhps): Change alternative 3
12522 of operand 2 to "m".
12524 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12527 * config/i386/sse.md (iptr): New mode attribute.
12528 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12529 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12530 (<sse>_vmmul<mode>3): Ditto.
12531 (<sse>_vmdiv<mode>3): Ditto.
12532 (sse_vmrcpv4sf2): Ditto.
12533 (<sse>_vmsqrt<mode>2): Ditto.
12534 (sse_vmrsqrtv4sf2): Ditto.
12535 (<sse>_vm<code><mode>3): Ditto.
12536 (avx_vmcmp<mode>3): Ditto.
12537 (<sse>_vmmaskcmp<mode>3): Ditto.
12538 (<sse>_comi): Ditto.
12539 (<sse>_ucomi): Ditto.
12540 (*xop_vmfrcz_<mode>): Ditto.
12541 (*fmai_fmadd_<mode>): Ditto.
12542 (*fmai_fmsub_<mode>): Ditto.
12543 (*fmai_fnmadd_<mode>): Ditto.
12544 (*fmai_fnmsub_<mode>): Ditto.
12545 (*fma4i_vmfmadd_<mode>): Ditto.
12546 (*fma4i_vmfmsub_<mode>): Ditto.
12547 (*fma4i_vmfnmadd_<mode>): Ditto.
12548 (*fma4i_vmfnmsub_<mode>): Ditto.
12549 (*xop_vmfrcz_<mode>): Ditto.
12550 (sse_cvtps2pi): Ditto.
12551 (sse_cvttps2pi): Ditto.
12552 (sse_cvtss2si): Ditto.
12553 (sse_cvtss2si_2): Ditto.
12554 (sse_cvtss2siq_2): Ditto.
12555 (sse_cvttss2si): Ditto.
12556 (sse_cvttss2siq): Ditto.
12557 (sse_cvtsd2si): Ditto.
12558 (sse_cvtsd2si_2): Ditto.
12559 (sse_cvtsd2siq_2): Ditto.
12560 (sse_cvttsd2si): Ditto.
12561 (sse_cvttsd2siq): Ditto.
12562 (sse_cvtsd2ss): Ditto.
12563 (sse_cvtss2sd): Ditto.
12564 (avx2_pbroadcast<mode>): Ditto.
12565 (avx2_pbroadcast<mode>_1): Ditto.
12566 (*avx_vperm_broadcast_v4sf): Ditto.
12568 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12569 (sse_movlhps): Ditto.
12570 (sse_storehps): Ditto.
12571 (sse_loadhps): Ditto.
12572 (sse_storelps): Ditto.
12573 (sse_loadlps): Ditto.
12574 (*vec_concatv4sf): Ditto.
12575 (*vec_interleave_highv2df): Ditto.
12576 (*vec_interleave_lowv2df): Ditto.
12577 (*vec_extractv2df_1_sse): Ditto.
12578 (*vec_extractv2df_0_sse): Ditto.
12579 (sse2_storelpd): Ditto.
12580 (sse2_loadlpd): Ditto.
12581 (sse2_movsd): Ditto.
12582 (*vec_concatv4si): Ditto.
12583 (vec_concatv2di): Ditto.
12585 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12586 for Intel asm dialect.
12587 (mmx_punpcklwd): Ditto.
12588 (mmx_punpckldq): Ditto.
12590 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12591 for intel assembler dialect.
12593 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12596 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12597 for BUILT_IN_C{LZ,LRSB}*.
12598 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12599 * tree-vrp.c (extract_range_basic): Handle
12600 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12601 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12602 fall thru to code calling set_value*.
12603 * builtins.c (expand_builtin): Remove *IMAX cases.
12604 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12605 if width is bigger than 2*HWI.
12607 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12609 PR rtl-optimization/55342
12610 * lra-int.h (lra_subreg_reload_pseudos): New.
12611 * lra.c: Add undoing optional reloads to the block diagram.
12612 (lra_subreg_reload_pseudos): New.
12613 (lra_optional_reload_pseudos): Change comments.
12614 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12615 lra_optional_reload_pseudos after undo transformations.
12616 * lra-assigns.c (pseudo_prefix_title): New.
12617 (lra_setup_reg_renumber): Use it.
12618 (spill_for): Ditto. Check subreg reload pseudos too.
12619 (assign_by_spills): Consider subreg reload pseudos too.
12620 * lra-constraints.c (simplify_operand_subreg): Use
12621 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12622 (curr_insn_transform): Recognize and do optional reloads.
12623 (undo_optional_reloads): New.
12624 (lra_undo_inheritance): Call undo_optional_reloads.
12626 2013-07-05 Thomas Quinot <quinot@adacore.com>
12628 * tree-complex.c (expand_complex_operations_1): Fix typo.
12630 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12632 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12633 (tune_params): New member 'const vec_costs'.
12634 * config/aarch64/aarch64.c (generic_vector_cost): New.
12635 (generic_tunings): New member 'generic_vector_cost'.
12636 (aarch64_builtin_vectorization_cost): New.
12637 (aarch64_add_stmt_cost): New.
12638 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12639 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12641 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12644 * config/i386/predicates.md (vsib_address_operand): Disallow
12645 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12647 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12649 PR middle-end/55030
12650 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12651 expand_builtin_setjmp_receiver.
12652 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12653 with NULL for the label parameter.
12654 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12655 the frame-pointer. Adjust comments.
12656 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12657 only if LABEL is non-NULL.
12659 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12661 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12662 (ARM_ABI_AAPCS64): Ditto.
12664 (ARM_DEFAULT_ABI): Ditto.
12666 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12668 * config/aarch64/aarch64-builtins.c
12669 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12670 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12672 * config/aarch64/aarch64-simd.md
12673 (aarch64_ld1<VALL:mode>): New.
12674 (aarch64_st1<VALL:mode>): Likewise.
12675 * config/aarch64/arm_neon.h
12676 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12678 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12680 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12682 (ix86_option_override_internal):Turn on all -mavx target flags by
12683 default as they are dependent on AVX anyway.
12685 2013-07-02 Cary Coutant <ccoutant@google.com>
12687 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12688 deterministic hash.
12689 (loc_checksum_ordered): Likewise.
12690 (hash_loc_operands): Remove inline keyword.
12692 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12694 PR tree-optimization/57741
12695 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12696 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12697 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12698 Allow REAL_CST step_exprs if flag_associative_math.
12699 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12701 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12703 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12705 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12707 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12709 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12711 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12713 (iorsi3_insn): Likewise.
12714 (arm_xorsi3): Likewise.
12716 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12718 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12720 (attribute "type"): ... here, and prefix with "wmmx_".
12721 (attribute "core_cycles"): Update for attribute changes.
12722 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12723 (tbcstv4hi): Likewise.
12724 (tbcstv2si): Likewise.
12725 (iwmmxt_iordi3): Likewise.
12726 (iwmmxt_xordi3): Likewise.
12727 (iwmmxt_anddi3): Likewise.
12728 (iwmmxt_nanddi3): Likewise.
12729 (iwmmxt_arm_movdi): Likewise.
12730 (iwmmxt_movsi_insn): Likewise.
12731 (mov<mode>_internal): Likewise.
12732 (and<mode>3_iwmmxt): Likewise.
12733 (ior<mode>3_iwmmxt): Likewise.
12734 (xor<mode>3_iwmmxt): Likewise.
12735 (add<mode>3_iwmmxt): Likewise.
12736 (ssaddv8qi3): Likewise.
12737 (ssaddv4hi3): Likewise.
12738 (ssaddv2si3): Likewise.
12739 (usaddv8qi3): Likewise.
12740 (usaddv4hi3): Likewise.
12741 (usaddv2si3): Likewise.
12742 (sub<mode>3_iwmmxt): Likewise.
12743 (sssubv8qi3): Likewise.
12744 (sssubv4hi3): Likewise.
12745 (sssubv2si3): Likewise.
12746 (ussubv8qi3): Likewise.
12747 (ussubv4hi3): Likewise.
12748 (ussubv2si3): Likewise.
12749 (mulv4hi3_iwmmxt): Likewise.
12750 (smulv4hi3_highpart): Likewise.
12751 (umulv4hi3_highpart): Likewise.
12752 (iwmmxt_wmacs): Likewise.
12753 (iwmmxt_wmacsz): Likewise.
12754 (iwmmxt_wmacu): Likewise.
12755 (iwmmxt_wmacuz): Likewise.
12756 (iwmmxt_clrdi): Likewise.
12757 (iwmmxt_clrv8qi): Likewise.
12758 (iwmmxt_clr4hi): Likewise.
12759 (iwmmxt_clr2si): Likewise.
12760 (iwmmxt_uavgrndv8qi3): Likewise.
12761 (iwmmxt_uavgrndv4hi3): Likewise.
12762 (iwmmxt_uavgv8qi3): Likewise.
12763 (iwmmxt_uavgv4hi3): Likewise.
12764 (iwmmxt_tinsrb): Likewise.
12765 (iwmmxt_tinsrh): Likewise.
12766 (iwmmxt_tinsrw): Likewise.
12767 (iwmmxt_textrmub): Likewise.
12768 (iwmmxt_textrmsb): Likewise.
12769 (iwmmxt_textrmuh): Likewise.
12770 (iwmmxt_textrmsh): Likewise.
12771 (iwmmxt_textrmw): Likewise.
12772 (iwmxxt_wshufh): Likewise.
12773 (eqv8qi3): Likewise.
12774 (eqv4hi3): Likewise.
12775 (eqv2si3): Likewise.
12776 (gtuv8qi3): Likewise.
12777 (gtuv4hi3): Likewise.
12778 (gtuv2si3): Likewise.
12779 (gtv8qi3): Likewise.
12780 (gtv4hi3): Likewise.
12781 (gtv2si3): Likewise.
12782 (smax<mode>3_iwmmxt): Likewise.
12783 (umax<mode>3_iwmmxt): Likewise.
12784 (smin<mode>3_iwmmxt): Likewise.
12785 (umin<mode>3_iwmmxt): Likewise.
12786 (iwmmxt_wpackhss): Likewise.
12787 (iwmmxt_wpackwss): Likewise.
12788 (iwmmxt_wpackdss): Likewise.
12789 (iwmmxt_wpackhus): Likewise.
12790 (iwmmxt_wpackwus): Likewise.
12791 (iwmmxt_wpackdus): Likewise.
12792 (iwmmxt_wunpckihb): Likewise.
12793 (iwmmxt_wunpckihh): Likewise.
12794 (iwmmxt_wunpckihw): Likewise.
12795 (iwmmxt_wunpckilb): Likewise.
12796 (iwmmxt_wunpckilh): Likewise.
12797 (iwmmxt_wunpckilw): Likewise.
12798 (iwmmxt_wunpckehub): Likewise.
12799 (iwmmxt_wunpckehuh): Likewise.
12800 (iwmmxt_wunpckehuw): Likewise.
12801 (iwmmxt_wunpckehsb): Likewise.
12802 (iwmmxt_wunpckehsh): Likewise.
12803 (iwmmxt_wunpckehsw): Likewise.
12804 (iwmmxt_wunpckelub): Likewise.
12805 (iwmmxt_wunpckeluh): Likewise.
12806 (iwmmxt_wunpckeluw): Likewise.
12807 (iwmmxt_wunpckelsb): Likewise.
12808 (iwmmxt_wunpckelsh): Likewise.
12809 (iwmmxt_wunpckelsw): Likewise.
12810 (ror<mode>3): Likewise.
12811 (ashr<mode>3_iwmmxt): Likewise.
12812 (lshr<mode>3_iwmmxt): Likewise.
12813 (ashl<mode>3_iwmmxt): Likewise.
12814 (ror<mode>3_di): Likewise.
12815 (ashr<mode>3_di): Likewise.
12816 (lshr<mode>3_di): Likewise.
12817 (ashl<mode>3_di): Likewise.
12818 (iwmmxt_wmadds): Likewise.
12819 (iwmmxt_wmaddu): Likewise.
12820 (iwmmxt_tmia): Likewise.
12821 (iwmmxt_tmiaph): Likewise.
12822 (iwmmxt_tmiabb): Likewise.
12823 (iwmmxt_tmiatb): Likewise.
12824 (iwmmxt_tmiabt): Likewise.
12825 (iwmmxt_tmiatt): Likewise.
12826 (iwmmxt_tmovmskb): Likewise.
12827 (iwmmxt_tmovmskh): Likewise.
12828 (iwmmxt_tmovmskw): Likewise.
12829 (iwmmxt_waccb): Likewise.
12830 (iwmmxt_wacch): Likewise.
12831 (iwmmxt_waccw): Likewise.
12832 (iwmmxt_waligni): Likewise.
12833 (iwmmxt_walignr): Likewise.
12834 (iwmmxt_walignr0): Likewise.
12835 (iwmmxt_walignr1): Likewise.
12836 (iwmmxt_walignr2): Likewise.
12837 (iwmmxt_walignr3): Likewise.
12838 (iwmmxt_wsadb): Likewise.
12839 (iwmmxt_wsadh): Likewise.
12840 (iwmmxt_wsadbz): Likewise.
12841 (iwmmxt_wsadhz): Likewise.
12842 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12843 (iwmmxt_wabsdiffb): Likewise.
12844 (iwmmxt_wabsdiffh): Likewise.
12845 (iwmmxt_wabsdiffw): Likewise.
12846 (iwmmxt_waddsubhx): Likewise
12847 (iwmmxt_wsubaddhx): Likewise.
12848 (addc<mode>3): Likewise.
12849 (iwmmxt_avg4): Likewise.
12850 (iwmmxt_avg4r): Likewise.
12851 (iwmmxt_wmaddsx): Likewise.
12852 (iwmmxt_wmaddux): Likewise.
12853 (iwmmxt_wmaddsn): Likewise.
12854 (iwmmxt_wmaddun): Likewise.
12855 (iwmmxt_wmulwsm): Likewise.
12856 (iwmmxt_wmulwum): Likewise.
12857 (iwmmxt_wmulsmr): Likewise.
12858 (iwmmxt_wmulumr): Likewise.
12859 (iwmmxt_wmulwsmr): Likewise.
12860 (iwmmxt_wmulwumr): Likewise.
12861 (iwmmxt_wmulwl): Likewise.
12862 (iwmmxt_wqmulm): Likewise.
12863 (iwmmxt_wqmulwm): Likewise.
12864 (iwmmxt_wqmulmr): Likewise.
12865 (iwmmxt_wqmulwmr): Likewise.
12866 (iwmmxt_waddbhusm): Likewise.
12867 (iwmmxt_waddbhusl): Likewise.
12868 (iwmmxt_wqmiabb): Likewise.
12869 (iwmmxt_wqmiabt): Likewise.
12870 (iwmmxt_wqmiatb): Likewise.
12871 (iwmmxt_wqmiatt): Likewise.
12872 (iwmmxt_wqmiabbn): Likewise.
12873 (iwmmxt_wqmiabtn): Likewise.
12874 (iwmmxt_wqmiatbn): Likewise.
12875 (iwmmxt_wqmiattn): Likewise.
12876 (iwmmxt_wmiabb): Likewise.
12877 (iwmmxt_wmiabt): Likewise.
12878 (iwmmxt_wmiatb): Likewise.
12879 (iwmmxt_wmiatt): Likewise.
12880 (iwmmxt_wmiabbn): Likewise.
12881 (iwmmxt_wmiabtn): Likewise.
12882 (iwmmxt_wmiatbn): Likewise.
12883 (iwmmxt_wmiattn): Likewise.
12884 (iwmmxt_wmiawbb): Likewise.
12885 (iwmmxt_wmiawbt): Likewise.
12886 (iwmmxt_wmiawtb): Likewise.
12887 (iwmmxt_wmiawtt): Likewise.
12888 (iwmmxt_wmiawbbn): Likewise.
12889 (iwmmxt_wmiawbtn): Likewise.
12890 (iwmmxt_wmiawtbn): Likewise.
12891 (iwmmxt_wmiawttn): Likewise.
12892 (iwmmxt_wmerge): Likewise.
12893 (iwmmxt_tandc<mode>3): Likewise.
12894 (iwmmxt_torc<mode>3): Likewise.
12895 (iwmmxt_torvsc<mode>3): Likewise.
12896 (iwmmxt_textrc<mode>3): Likewise.
12897 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12898 (wmmxt_pack): Likewise.
12899 (wmmxt_mult_c1): Likewise.
12900 (wmmxt_mult_c2): Likewise.
12901 (wmmxt_alu_c1): Likewise.
12902 (wmmxt_alu_c2): Likewise.
12903 (wmmxt_alu_c3): Likewise.
12904 (wmmxt_transfer_c1): Likewise.
12905 (wmmxt_transfer_c2): Likewise.
12906 (wmmxt_transfer_c3): Likewise.
12907 (marvell_f_iwmmxt_wstr): Likewise.
12908 (marvell_f_iwmmxt_wldr): Likewise.
12910 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12912 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
12914 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12917 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12918 * lra-constraints.c (need_for_split_p): Check call used hard regs
12919 living through calls.
12921 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12922 call used regs for call insn.
12924 2013-06-28 Jakub Jelinek <jakub@redhat.com>
12927 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12928 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
12930 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12932 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
12934 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12936 * lra-constraints.c (need_for_split_p): Check call used hard regs
12937 living through calls.
12939 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
12942 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
12943 to tie with any other modes. Eliminate Altivec vector mode tests,
12944 since these are a subset of ALTIVEC or VSX vector modes. Simplify
12945 code, to return 0 if testing MODE2 for a condition, if we've
12946 already tested MODE1 for the same condition.
12948 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12950 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
12953 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12955 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12956 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
12958 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12960 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
12962 (aarch64_symbolic_constant_p): Remove.
12963 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
12964 static. Fix line length and white space.
12965 (aarch64_symbolic_constant_p): Remove.
12966 * config/aarch64/predicates.md (aarch64_valid_symref):
12967 Use aarch64_classify_symbol_expression.
12969 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12971 * config/arm/constraints.md (Ts): New constraint.
12972 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
12974 (compare_scc): Use "Ts" constraint for operand 0.
12975 (ior_scc_scc): Likewise.
12976 (and_scc_scc): Likewise.
12977 (and_scc_scc_nodom): Likewise.
12978 (ior_scc_scc_cmp): Likewise for operand 7.
12979 (and_scc_scc_cmp): Likewise.
12980 * config/arm/thumb2.md (thumb2_movsi_insn):
12981 Add alternatives for 16-bit encodings.
12982 (thumb2_movhi_insn): Likewise.
12983 (thumb2_movsicc_insn): Likewise.
12984 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
12985 (thumb2_negscc): Use "Ts" constraint.
12986 Move mvn instruction outside cond_exec block.
12987 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
12988 for 16-bit encodings.
12990 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12992 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
12994 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
12995 (mulsi3subsi): Likewise.
12996 (mulsidi3adddi): Likewise.
12997 (mulsidi3_v6): Likewise.
12998 (umulsidi3_v6): Likewise.
12999 (umulsidi3adddi_v6): Likewise.
13000 (smulsi3_highpart_v6): Likewise.
13001 (umulsi3_highpart_v6): Likewise.
13002 (mulhisi3tb): Likewise.
13003 (mulhisi3bt): Likewise.
13004 (mulhisi3tt): Likewise.
13005 (maddhisi4): Likewise.
13006 (maddhisi4tb): Likewise.
13007 (maddhisi4tt): Likewise.
13008 (maddhidi4): Likewise.
13009 (maddhidi4tb): Likewise.
13010 (maddhidi4tt): Likewise.
13011 (zeroextractsi_compare0_scratch): Likewise.
13012 (insv_zero): Likewise.
13013 (insv_t2): Likewise.
13014 (anddi_notzesidi_di): Likewise.
13015 (anddi_notsesidi_di): Likewise.
13016 (andsi_notsi_si): Likewise.
13017 (iordi_zesidi_di): Likewise.
13018 (xordi_zesidi_di): Likewise.
13019 (andsi_iorsi3_notsi): Likewise.
13020 (smax_0): Likewise.
13021 (smax_m1): Likewise.
13022 (smin_0): Likewise.
13023 (not_shiftsi): Likewise.
13024 (unaligned_loadsi): Likewise.
13025 (unaligned_loadhis): Likewise.
13026 (unaligned_loadhiu): Likewise.
13027 (unaligned_storesi): Likewise.
13028 (unaligned_storehi): Likewise.
13029 (extv_reg): Likewise.
13030 (extzv_t2): Likewise.
13031 (divsi3): Likewise.
13032 (udivsi3): Likewise.
13033 (arm_zero_extendhisi2addsi): Likewise.
13034 (arm_zero_extendqisi2addsi): Likewise.
13035 (compareqi_eq0): Likewise.
13036 (arm_extendhisi2_v6): Likewise.
13037 (arm_extendqisi2addsi): Likewise.
13038 (arm_movt): Likewise.
13039 (thumb2_ldrd): Likewise.
13040 (thumb2_ldrd_base): Likewise.
13041 (thumb2_ldrd_base_neg): Likewise.
13042 (thumb2_strd): Likewise.
13043 (thumb2_strd_base): Likewise.
13044 (thumb2_strd_base_neg): Likewise.
13045 (arm_negsi2): Add alternative for 16-bit encoding.
13046 (arm_one_cmplsi2): Likewise.
13048 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13050 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13051 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13052 (movdfcc): Likewise.
13053 * config/arm/vfp.md (*thumb2_movsf_vfp):
13054 Disable predication for arm_restrict_it.
13055 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13056 (*thumb2_movdfcc_vfp): Likewise.
13057 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13058 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13059 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13060 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13061 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13062 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13063 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13064 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13065 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13066 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13067 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13068 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13069 Disable predication for arm_restrict_it.
13071 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13073 * config/i386/bmiintrin.h (_bextr_u32): New.
13074 (_bextr_u64): Ditto.
13076 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13078 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13079 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13080 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13081 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13082 * config/mips/n32-elf.h: ...this new file.
13084 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13087 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13088 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13090 2013-06-27 Catherine Moore <clm@codesourcery.com>
13092 * config/mips/mips-tables.opt: Regenerate.
13093 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13094 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13095 * doc/invoke.texi: Add -m14kc.
13097 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13100 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13101 constraints of operand 1 and 2.
13104 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13105 to match RTL canonicalization. Swap predicates and
13106 constraints of operand 1 and 2.
13108 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13110 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13111 Process OP_INOUT regs for splitting too.
13113 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13115 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13116 decl before the loop, initialize to NULL.
13117 (vectorizable_load): Initialize ptr_incr to NULL.
13119 2013-06-27 Martin Jambor <mjambor@suse.cz>
13122 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13123 * ipa-ref.c (ipa_maybe_record_reference): New function.
13124 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13125 * ipa-cp.c (create_specialized_node): Record potential references from
13127 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13129 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13131 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13132 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13133 'mode' to force_reg.
13134 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13135 (aarch64_expand_mov_immediate): Likewise.
13137 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13139 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13140 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13142 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13144 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13145 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13146 (struct machine_function): Add tbegin_p.
13147 (s390_canonicalize_comparison): Fold CC mode compares to
13148 conditional jump if possible.
13149 (s390_emit_jump): Return the emitted jump.
13150 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13151 Handle CCRAWmode compares.
13152 (s390_option_override): Default to -mhtm if available.
13153 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13154 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13155 FPRs instead of df_regs_ever_live_p.
13156 (s390_optimize_nonescaping_tx): New function.
13157 (s390_init_frame_layout): Extend clobbered_regs array to cover
13159 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13160 (s390_expand_tbegin): New function.
13161 (enum s390_builtin): New enum definition.
13162 (code_for_builtin): New array definition.
13163 (s390_init_builtins): New function.
13164 (s390_expand_builtin): New function.
13165 (TARGET_INIT_BUILTINS): Define.
13166 (TARGET_EXPAND_BUILTIN): Define.
13167 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13168 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13169 (s390_alc_comparison): Likewise.
13170 * config/s390/s390-modes.def: Add CCRAWmode.
13171 * config/s390/s390.h (processor_flags): Add PF_TX.
13172 (TARGET_CPU_HTM): Define macro.
13173 (TARGET_HTM): Define macro.
13174 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13175 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13176 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13177 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13178 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13179 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13180 ("*cc_to_int"): Move up.
13181 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13182 constants other than 0.
13183 ("*ccraw_to_int"): New insn and splitter definition.
13184 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13185 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13186 ("tx_assist"): New expander.
13187 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13188 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13189 * config/s390/s390.opt: Add -mhtm option.
13190 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13191 * config/s390/htmxlintrin.h: New file.
13192 * config/s390/htmintrin.h: New file.
13193 * config/s390/s390intrin.h: New file.
13194 * doc/extend.texi: Document htm builtins.
13195 * config.gcc: Add the new header files to extra_headers.
13197 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13199 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13200 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13202 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13203 Pat Haugen <pthaugen@us.ibm.com>
13204 Peter Bergner <bergner@vnet.ibm.com>
13206 * config/rs6000/power8.md: New.
13207 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13208 setting for power8 entry.
13209 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13210 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13211 test for Power4/Power5 only.
13212 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13214 (force_new_group): Adjust comment.
13215 * config/rs6000/rs6000.md: Include power8.md.
13217 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13219 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13220 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13222 (tune_params): Update comment.
13223 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13224 (arm_max_conditional_execute): New function.
13225 (thumb2_final_prescan_insn): Use max_insn_skipped and
13226 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13228 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13230 PR tree-optimization/57705
13231 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13232 SSA_NAME step, provided that it is not defined inside the loop.
13233 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13234 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13236 2013-06-25 Martin Jambor <mjambor@suse.cz>
13238 PR middle-end/57670
13239 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13240 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13242 (ipa_note_param_call): Initialize member_ptr flag.
13243 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13244 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13245 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13246 (ipa_read_indirect_edge_info): Likewise.
13248 2013-06-25 Richard Biener <rguenther@suse.de>
13250 PR middle-end/56977
13251 * passes.c (init_optimization_passes): Move pass_fold_builtins
13252 and pass_dce earlier with -Og.
13254 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13256 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13257 <BIT_FIELD_REF>: Remove trailing TAB.
13258 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13261 2013-06-24 Martin Jambor <mjambor@suse.cz>
13263 PR tree-optimization/57358
13264 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13265 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13266 (ipa_analyze_params_uses): Generate pessimistic info when true.
13268 2013-06-24 Martin Jambor <mjambor@suse.cz>
13270 PR tree-optimization/57539
13271 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13272 global.inlined_to of the new node to it. All callers changed.
13273 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13274 inlining_into, pass it to cgraph_clone_node.
13275 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13276 ipa_free_edge_args_substructures.
13277 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13278 rdesc linked list. Do not assert rdesc edges have inlined caller.
13279 Assert we have found an rdesc in the rdesc list.
13281 2013-06-24 Richard Biener <rguenther@suse.de>
13283 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13284 (pointer_set_lookup): Declare.
13285 (class pointer_map): New template class implementing a
13286 generic pointer to T map.
13287 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13288 pointer_map<T>::contains, pointer_map<T>::insert,
13289 pointer_map<T>::traverse): New functions.
13290 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13291 (pointer_set_lookup): New function.
13292 (pointer_set_contains): Use pointer_set_lookup.
13293 (pointer_set_insert): Likewise.
13294 (insert_aux): Remove.
13295 (struct pointer_map_t): Embed a pointer_set_t.
13296 (pointer_map_create): Adjust.
13297 (pointer_map_destroy): Likewise.
13298 (pointer_map_contains): Likewise.
13299 (pointer_map_insert): Likewise.
13300 (pointer_map_traverse): Likewise.
13301 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13302 pointer_map<unsigned> instead of a pointer_map_t.
13303 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13304 (streamer_tree_cache_lookup): Likewise.
13305 (streamer_tree_cache_create): Likewise.
13306 (streamer_tree_cache_delete): Likewise.
13307 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13308 pointer_map<unsigned> instead of a pointer_map_t.
13309 (lto_init_tree_ref_encoder): Adjust.
13310 (lto_destroy_tree_ref_encoder): Likewise.
13311 * lto-section-out.c (lto_output_decl_index): Likewise.
13312 (lto_record_function_out_decl_state): Likewise.
13313 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13315 2013-06-24 Richard Biener <rguenther@suse.de>
13317 PR tree-optimization/57488
13318 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13320 2013-06-24 Alan Modra <amodra@gmail.com>
13322 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13323 (gen_easy_altivec_constant): Likewise.
13324 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13325 easy_vector_constant_msb): Likewise.
13327 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13330 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13331 add missing return true.
13333 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13336 * config/sh/predicates.md (general_extend_operand): Invoke
13337 general_movsrc_operand for memory operands.
13338 (general_movsrc_operand): Allow reg+reg addressing, do not use
13339 general_operand for memory operands.
13341 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13343 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13344 when current target options does not apply.
13345 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13346 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13347 * config/i386/bmiintrin.h: Pass appropriate target
13348 attributes to header.
13349 * config/i386/mmintrin.h: Ditto.
13350 * config/i386/nmmintrin.h: Ditto.
13351 * config/i386/avx2intrin.h: Ditto.
13352 * config/i386/fxsrintrin.h: Ditto.
13353 * config/i386/tbmintrin.h: Ditto.
13354 * config/i386/xsaveintrin.h: Ditto.
13355 * config/i386/f16cintrin.h: Ditto.
13356 * config/i386/xtestintrin.h: Ditto.
13357 * config/i386/xsaveoptintrin.h: Ditto.
13358 * config/i386/bmi2intrin.h: Ditto.
13359 * config/i386/lzcntintrin.h: Ditto.
13360 * config/i386/smmintrin.h: Ditto.
13361 * config/i386/wmmintrin.h: Ditto.
13362 * config/i386/x86intrin.h: Remove all header include guards.
13363 * config/i386/prfchwintrin.h: Ditto.
13364 * config/i386/pmmintrin.h: Ditto.
13365 * config/i386/tmmintrin.h: Ditto.
13366 * config/i386/xmmintrin.h: Ditto.
13367 * config/i386/popcntintrin.h: Ditto.
13368 * config/i386/rdseedintrin.h: Ditto.
13369 * config/i386/ammintrin.h: Ditto.
13370 * config/i386/emmintrin.h: Ditto.
13371 * config/i386/immintrin.h: Remove all header include guards.
13372 * config/i386/fma4intrin.h: Ditto.
13373 * config/i386/lwpintrin.h: Ditto.
13374 * config/i386/xopintrin.h: Ditto.
13375 * config/i386/ia32intrin.h: Ditto.
13376 * config/i386/avxintrin.h: Ditto.
13377 * config/i386/rtmintrin.h: Ditto.
13378 * config/i386/fmaintrin.h: Ditto.
13379 * config/i386/mm3dnow.h: Ditto.
13381 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13383 * common/config/i386/i386-common.c: Handle LZCNT.
13385 2013-06-22 Andi Kleen <ak@linux.intel.com>
13387 * doc/extend.texi: Use __atomic_store_n instead of
13388 __atomic_store in HLE example.
13390 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13392 * config/sh/sh.c: Remove <cstdlib> workaround.
13394 2013-06-21 Andi Kleen <ak@linux.intel.com>
13396 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13398 2013-06-21 Andi Kleen <ak@linux.intel.com>
13400 * doc/extend.texi: Document that __atomic_clear and
13401 __atomic_test_and_set should only be used with bool.
13403 2013-06-20 Jan Hubicka <jh@suse.cz>
13405 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13406 types_same_for_odr.
13407 * tree.c (decls_same_for_odr): New function.
13408 (same_for_edr): New function.
13409 (types_same_for_odr): New function.
13410 (get_binfo_at_offset): Use it.
13411 * tree.h (types_same_for_odr): Declare.
13413 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13414 Jason Merrill <jason@redhat.com>
13416 * system.h: Include <cstdlib> as well as <stdlib.h>.
13418 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13421 * config/i386/i386.c (construct_container): Report error if
13422 long double is used with disabled x87 float returns.
13424 2013-06-20 Jan Hubicka <jh@suse.cz>
13426 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13428 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13430 PR rtl-optimization/57425
13431 PR rtl-optimization/57569
13432 * alias.c (write_dependence_p): Remove parameters mem_mode and
13433 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13434 Changed all callers.
13435 (canon_anti_dependence): Get comments and semantics in sync.
13436 Add parameter mem_canonicalized. Changed all callers.
13437 * rtl.h (canon_anti_dependence): Update prototype.
13439 2013-06-20 Richard Biener <rguenther@suse.de>
13441 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13442 byte case, inline streamer_read_uchar and defer section
13445 2013-06-20 Richard Biener <rguenther@suse.de>
13447 PR tree-optimization/57584
13448 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13449 SSA names into the expanded expression that take part in
13450 abnormal coalescing.
13452 2013-06-19 Sharad Singhai <singhai@google.com>
13454 * gcov.c (print_usage): Handle new option.
13455 (process_args): Ditto.
13456 (get_gcov_intermediate_filename): New function.
13457 (output_intermediate_file): New function.
13458 (output_gcov_file): New function
13459 (generate_results): Handle new option.
13460 (release_function): Relase demangled name.
13461 (read_graph_file): Handle demangled name.
13462 (output_lines): Ditto.
13463 * doc/gcov.texi: Document gcov intermediate format.
13465 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13468 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13469 (lra_emit_add): Use the functions. Add comment about Y as an
13472 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13475 * collect2.c (collect_atexit): New.
13476 (collect_exit): Delete.
13477 (main): Register collect_atexit with atexit.
13478 (collect_wait): Change collect_exit to exit.
13480 * collect2.h (collect_exit): Delete.
13481 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13483 2013-06-19 Wei Mi <wmi@google.com>
13485 PR rtl-optimization/57518
13486 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13487 if regno is used in paradoxical subreg.
13488 (update_equiv_regs): Check pdx_subregs[regno] before
13489 set a reg to be equivalent with a mem.
13491 2013-06-19 Matthias Klose <doko@ubuntu.com>
13494 * file-find.h (find_a_file): Add a mode parameter.
13495 * file-find.c (find_a_file): Likewise.
13496 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13497 with X_OK for the executables.
13498 * collect2.c (main): Call find_a_file with X_OK.
13500 2013-06-19 Steve Ellcey <sellcey@mips.com>
13503 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13504 Use NEXT_INSN instead of next_real_insn.
13506 2013-06-19 Jan Hubicka <jh@suse.cz>
13508 * cgraph.h (const_value_known_p): Replace by ...
13509 (ctor_for_folding): .. this one.
13510 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13511 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13512 * expr.c (expand_expr_real_1): Likewise.
13513 (string_constant): Likewise.
13514 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13515 * ipa.c (process_references): Likewise.
13516 (symtab_remove_unreachable_nodes): Likewise.
13517 * ipa-inline-analysis.c (param_change_prob): Likewise.
13518 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13519 (get_base_constructor): Likwise.
13520 * varpool.c (varpool_remove_node): Likewise.
13521 (varpool_remove_initializer): LIkewise.
13522 (dump_varpool_node): LIkwise.
13523 (const_value_known_p): Rewrite to ...
13524 (ctor_for_folding): ... this one.
13526 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13529 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13530 PERSONALITY in $PATH derived prefixes.
13532 2013-06-19 Jeff Law <law@redhat.com>
13534 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13537 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13538 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13541 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13543 * config/arm/vfp.md: Move VFP instruction classification documentation
13545 * config/arm/arm.md: ... here. Update instruction classification
13548 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13550 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13551 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13552 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13554 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13557 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13558 with NEXT_INSN. Use tablejump_p to check for jump table data
13561 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13564 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13565 that now in C++ the value is correct per the C++ standards.
13567 2013-06-19 Richard Biener <rguenther@suse.de>
13569 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13570 for global context.
13572 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13575 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13578 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13579 with next_active_insn.
13581 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13583 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13584 functions are inlined during failures to flag an error.
13585 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13586 in early inline pass.
13588 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13590 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13593 2013-06-18 Julian Brown <julian@codesourcery.com>
13595 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13596 Permit virtual register pre-reload if !strict.
13597 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13599 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13601 * config/arm/neon.md (movmisalign<mode>): Use
13602 neon_perm_struct_or_reg_operand instead of
13603 neon_struct_or_register_operand.
13604 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13605 neon_permissive_struct_operand instead of neon_struct_operand.
13606 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13607 neon_vector_mem_operand.
13608 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13609 neon_vector_mem_operand.
13610 (neon_permissive_struct_operand): New.
13611 (neon_struct_or_register_operand): Rename to...
13612 (neon_perm_struct_or_reg_operand): This. Adjust call to
13613 neon_vector_mem_operand.
13615 2013-06-18 Richard Biener <rguenther@suse.de>
13617 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13618 * lto-streamer.h: Include pointer-set.h.
13619 (struct lto_decl_slot): Remove.
13620 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13621 Remove next_index entry.
13622 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13623 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13624 (lto_init_tree_ref_encoder): Adjust.
13625 (lto_destroy_tree_ref_encoder): Likewise.
13626 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13627 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13628 (lto_output_decl_index): Adjust.
13629 (lto_new_out_decl_state): Likewise.
13630 (lto_record_function_out_decl_state): Likewise.
13631 * lto-streamer-out.c (copy_function): Likewise.
13633 2013-06-18 Richard Biener <rguenther@suse.de>
13635 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13636 * cgraphunit.c: Include cfgloop.h.
13637 (init_lowered_empty_function): Initialize the loop tree.
13638 (assemble_thunk): Insert new BBs into loops.
13640 2013-06-18 Richard Biener <rguenther@suse.de>
13642 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13643 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13644 the map from cache entry to cache index optional.
13645 (streamer_tree_cache_replace_tree): Adjust accordingly.
13646 (streamer_tree_cache_append): Likewise.
13647 (streamer_tree_cache_delete): Likewise.
13648 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13649 streamer cache map from cache entry to cache index.
13650 * lto-streamer-out.c (create_output_block): Adjust.
13652 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13654 * config/arm/arm.md (attribute "insn"): Move multiplication and
13655 division attributes to...
13656 (attribute "type"): ... here. Remove mult.
13657 (attribute "mul32"): New attribute.
13658 (attribute "mul64"): Add umaal.
13659 (*arm_mulsi3): Update attributes.
13660 (*arm_mulsi3_v6): Likewise.
13661 (*thumb_mulsi3): Likewise.
13662 (*thumb_mulsi3_v6): Likewise.
13663 (*mulsi3_compare0): Likewise.
13664 (*mulsi3_compare0_v6): Likewise.
13665 (*mulsi_compare0_scratch): Likewise.
13666 (*mulsi_compare0_scratch_v6): Likewise.
13667 (*mulsi3addsi): Likewise.
13668 (*mulsi3addsi_v6): Likewise.
13669 (*mulsi3addsi_compare0): Likewise.
13670 (*mulsi3addsi_compare0_v6): Likewise.
13671 (*mulsi3addsi_compare0_scratch): Likewise.
13672 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13673 (*mulsi3subsi): Likewise.
13674 (*mulsidi3adddi): Likewise.
13675 (*mulsi3addsi_v6): Likewise.
13676 (*mulsidi3adddi_v6): Likewise.
13677 (*mulsidi3_nov6): Likewise.
13678 (*mulsidi3_v6): Likewise.
13679 (*umulsidi3_nov6): Likewise.
13680 (*umulsidi3_v6): Likewise.
13681 (*umulsidi3adddi): Likewise.
13682 (*umulsidi3adddi_v6): Likewise.
13683 (*smulsi3_highpart_nov6): Likewise.
13684 (*smulsi3_highpart_v6): Likewise.
13685 (*umulsi3_highpart_nov6): Likewise.
13686 (*umulsi3_highpart_v6): Likewise.
13687 (mulhisi3): Likewise.
13688 (*mulhisi3tb): Likewise.
13689 (*mulhisi3bt): Likewise.
13690 (*mulhisi3tt): Likewise.
13691 (maddhisi4): Likewise.
13692 (*maddhisi4tb): Likewise.
13693 (*maddhisi4tt): Likewise.
13694 (maddhidi4): Likewise.
13695 (*maddhidi4tb): Likewise.
13696 (*maddhidi4tt): Likewise.
13697 (divsi3): Likewise.
13698 (udivsi3): Likewise.
13699 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13700 (thumb2_mulsi_short_compare0): Likewise.
13701 (thumb2_mulsi_short_compare0_scratch): Likewise.
13702 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13703 (1020mult2): Likewise.
13704 (1020mult3): Likewise.
13705 (1020mult4): Likewise.
13706 (1020mult5): Likewise.
13707 (1020mult6): Likewise.
13708 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13710 (cortex_a15_mult64): Likewise.
13711 (cortex_a15_sdiv): Likewise.
13712 (cortex_a15_udiv): Likewise.
13713 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13719 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13720 (pj4_ir_div): Likewise.
13721 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13722 (11_mult2): Likewise.
13723 (11_mult3): Likewise.
13724 (11_mult4): Likewise.
13725 (11_mult5): Likewise.
13726 (11_mult6): Likewise.
13727 (11_mult7): Likewise.
13728 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13729 (cortex_a8_mla): Likewise.
13730 (cortex_a8_mull): Likewise.
13731 (cortex_a8_smulwy): Likewise.
13732 (cortex_a8_smlald): Likewise.
13733 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13734 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13735 (cortex_r4_mul_3): Likewise.
13736 (cortex_r4_mla_4): Likewise.
13737 (cortex_r4_mla_3): Likewise.
13738 (cortex_r4_smlald): Likewise.
13739 (cortex_r4_mull): Likewise.
13740 (cortex_r4_sdiv): Likewise.
13741 (cortex_r4_udiv): Likewise.
13742 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13743 (cortex_a7_idiv): Likewise.
13744 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13745 (9_mult2): Likewise.
13746 (9_mult3): Likewise.
13747 (9_mult4): Likewise.
13748 (9_mult5): Likewise.
13749 (9_mult6): Likewise.
13750 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13751 (cortex_a53_sdiv): Likewise.
13752 (cortex_a53_udiv): Likewise.
13753 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13754 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13755 (mp626_mult2): Likewise.
13756 (mp626_mult3): Likewise.
13757 (mp626_mult4): Likewise.
13758 * config/arm/fa526.md (526_mult1): Update attribute change.
13759 (526_mult2): Likewise.
13760 * config/arm/arm-generic.md (mult): Update attribute change.
13761 (mult_ldsched_strongarm): Likewise.
13762 (mult_ldsched): Likewise.
13763 (multi_cycle): Likewise.
13764 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13765 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13766 (606te_mult2): Likewise.
13767 (606te_mult3): Likewise.
13768 (606te_mult4): Likewise.
13769 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13770 (cortex_a9_mac16): Likewise.
13771 (cortex_a9_multiply): Likewise.
13772 (cortex_a9_mac): Likewise.
13773 (cortex_a9_multiply_long): Likewise.
13774 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13775 (626te_mult2): Likewise.
13776 (626te_mult3): Likewise.
13777 (626te_mult4): Likewise.
13779 2013-06-18 Richard Biener <rguenther@suse.de>
13782 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13784 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13787 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13788 with next_active_insn.
13790 2013-06-18 Alan Modra <amodra@gmail.com>
13792 * config/rs6000/rs6000.h (enum data_align): New.
13793 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13794 (DATA_ABI_ALIGNMENT): Define.
13795 (CONSTANT_ALIGNMENT): Correct comment.
13796 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13797 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13799 2013-06-17 David Malcolm <dmalcolm@redhat.com>
13801 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13802 ATTRIBUTE_UNUSED marking.
13804 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13806 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13807 alternative and update.
13808 (aarch64_dup_lanedi): Delete.
13809 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13810 * config/aarch64/aarch64-simd-builtins.def: Update.
13812 2013-06-17 Richard Biener <rguenther@suse.de>
13814 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13815 (lto_input_scc): Declare.
13816 (lto_input_tree_1): Likewise.
13817 (struct lto_stats_d): Add num_tree_bodies_output and
13818 num_pickle_refs_output.
13819 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13820 (lto_read_tree_1): Split out from ...
13821 (lto_read_tree): ... this.
13822 (lto_input_scc): New function.
13823 (lto_input_tree_1): Split out from ...
13824 (lto_input_tree): ... this. Handle LTO_tree_scc.
13825 (lto_data_in_create): Create the streamer cache without hashes.
13826 * lto-streamer-out.c (create_output_block): Create the streamer
13827 cache with hashes when not doing WPA.
13828 (lto_write_tree_1): Split out from ...
13829 (lto_write_tree): ... this.
13830 (get_symbol_initial_value): New function.
13831 (lto_output_tree_1): Split out from ...
13832 (lto_output_tree): ... this. Write trees as series of SCCs
13833 using a DFS walk via DFS_write_tree.
13834 (struct sccs, struct scc_entry): New types.
13835 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13836 (DFS_write_tree_body): New function.
13837 (DFS_write_tree): Likewise.
13838 (hash_tree): Likewise.
13839 (scc_entry_compare): Likewise.
13840 (hash_scc): Likewise.
13841 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13842 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13843 TREE_CHAIN as regular reference.
13844 (streamer_read_integer_cst): Remove.
13845 (streamer_get_pickled_tree): Adjust.
13846 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13847 of DECL_EXTERNALs in BLOCK_VARS for now.
13848 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13850 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13851 Add hash value argument and record that if hashes are recorded
13853 (streamer_tree_cache_insert_1): Adjust.
13854 (streamer_tree_cache_insert): Likewise.
13855 (streamer_tree_cache_insert_at): Rename to ...
13856 (streamer_tree_cache_replace_tree): ... this and adjust.
13857 (streamer_tree_cache_append): Adjust.
13858 (record_common_node): Likewise.
13859 (streamer_tree_cache_create): Add argument whether to
13860 record hash values together with trees.
13861 (streamer_tree_cache_delete): Adjust.
13862 * tree-streamer.h (struct streamer_tree_cache_d): Add
13864 (streamer_read_integer_cst): Remove.
13865 (streamer_tree_cache_insert): Adjust.
13866 (streamer_tree_cache_append): Likewise.
13867 (streamer_tree_cache_insert_at): Rename to ...
13868 (streamer_tree_cache_replace_tree): ... this and adjust.
13869 (streamer_tree_cache_create): Add argument whether to record hashes.
13870 (streamer_tree_cache_get): Rename to ...
13871 (streamer_tree_cache_get_tree): ... this.
13872 (streamer_tree_cache_get_hash): New function.
13873 * tree.c (cache_integer_cst): New function.
13874 * tree.h (cache_integer_cst): Declare.
13875 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13876 * lto-symtab.c (lto_varpool_replace_node): Only release
13877 DECL_INITIAL of non-prevailing decls.
13878 * varpool.c (varpool_remove_initializer): Do not release
13879 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13881 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13883 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13884 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13885 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13886 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13887 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13888 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13889 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13890 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13891 instead of TARGET_64BIT.
13892 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13893 Require ISA_HAS_<D>DIV.
13895 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13897 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13898 (mips*-*-linux*): Move default with_llsc setting to where other
13900 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13902 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13903 Likewise. Remove default with_tune setting. Move default float
13904 setting to its own block. Handle with_llsc in the same block as above.
13906 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
13908 PR rtl-optimization/57425
13909 PR rtl-optimization/57569
13910 * alias.c (write_dependence_p): Add new parameters mem_mode,
13911 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13912 Changed all callers.
13913 (canon_anti_dependence): New function.
13914 * cse.c (check_dependence): Use canon_anti_dependence.
13915 * cselib.c (cselib_invalidate_mem): Likewise.
13916 * rtl.h (canon_anti_dependence): Declare.
13918 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13920 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13921 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13922 ".set mips3" for 64-bit targets.
13924 2013-06-15 Dehao Chen <dehao@google.com>
13926 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13927 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13928 (gimple_check_call_args): Likewise.
13929 * value-prof.c (check_ic_target): Likewise.
13930 * ipa-inline.c (early_inliner): Likewise.
13931 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13932 * cgraph.c (cgraph_create_edge_1): Likewise.
13933 (cgraph_make_edge_direct): Likewise.
13935 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
13938 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
13939 rs6000_output_move_128bit to handle emitting quad memory
13940 operations. Set attribute length to 8 bytes.
13942 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
13944 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
13946 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
13947 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
13948 (aarch64_<su>mlsl<mode>): Likewise.
13950 2013-06-14 Mike Stump <mikestump@comcast.net>
13952 * Makefile.in (TARGET_H): Add insn-codes.h.
13954 2013-06-14 Alan Modra <amodra@gmail.com>
13956 PR middle-end/57134
13957 PR middle-end/57586
13958 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
13959 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
13960 bitfield expansion when EXPAND_MEMORY.
13961 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
13963 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
13965 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
13966 test for clearing quad memory on 32-bit later.
13968 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
13970 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
13971 (fold_negate_expr): Likewise.
13972 (fold_real_zero_addition_p): Handle vectors.
13973 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
13975 2013-06-14 Alan Modra <amodra@gmail.com>
13977 * varasm.c (force_const_mem): Revert 2013-06-07 change.
13979 2013-06-13 Jan Hubicka <jh@suse.cz>
13981 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13982 Local comdats are not externally visible.
13983 * symtab.c (dump_symtab_base): Dump externally visible.
13984 (verify_symtab_base): Verify back links in the symtab hash.
13986 2013-06-13 Bin Cheng <bin.cheng@arm.com>
13988 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
13989 CONVERT_EXPR as equal nodes.
13991 2013-06-13 Bin Cheng <bin.cheng@arm.com>
13993 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
13995 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
13997 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
13998 Generalize to complex and vector.
13999 * tree.c (build_all_ones_cst): New function.
14000 * tree.h (build_all_ones_cst): Declare it.
14002 2013-06-13 Alan Modra <amodra@gmail.com>
14004 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14005 * config/rs6000/rs6000.md (signbittf2): New insn.
14006 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14007 (abstf2_internal, cmptf_internal2): Likewise.
14008 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14010 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14011 Pat Haugen <pthaugen@us.ibm.com>
14012 Peter Bergner <bergner@vnet.ibm.com>
14014 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14015 power8 byte, half-word, and quad-word atomic instructions.
14016 (emit_store_conditional): Likewise.
14017 (rs6000_expand_atomic_compare_and_swap): Likewise.
14018 (rs6000_expand_atomic_op): Likewise.
14020 * config/rs6000/sync.md (larx): Add new modes for power8.
14022 (AINT): New mode iterator to include TImode as well as normal
14023 integer modes on power8.
14024 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14025 that VSX registers are not considered. Use AINT mode iterator
14026 instead of INT1 to allow inclusion of quad word atomic operations
14028 (load_locked<mode>): Likewise.
14029 (store_conditional<mode>): Likewise.
14030 (atomic_compare_and_swap<mode>): Likewise.
14031 (atomic_exchange<mode>): Likewise.
14032 (atomic_nand<mode>): Likewise.
14033 (atomic_fetch_<fetchop_name><mode>): Likewise.
14034 (atomic_nand_fetch<mode>): Likewise.
14035 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14037 (ATOMIC): On power8, add QImode, HImode modes.
14038 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14039 modes that promote to SImode.
14040 (load_lockedti): Convert TImode arguments to PTImode, so that we
14041 get a guaranteed even/odd register pair.
14042 (load_lockedpti): Likewise.
14043 (store_conditionalti): Likewise.
14044 (store_conditionalpti): Likewise.
14046 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14047 atomic load/store instructions.
14050 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14052 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14054 (insn_count): New attribute, with most cases extracted from...
14055 (length): ...here. Redefine most cases in terms of insn_count.
14056 (single_insn): Delete.
14057 (can_delay): Use insn_count to check for single instructions.
14058 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14059 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14060 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14061 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14062 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14063 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14064 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14065 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14066 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14067 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14068 rather than "length".
14069 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14070 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14071 Use "insn_count" rather than "length".
14072 * config/mips/mips-dsp.md
14073 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14074 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14077 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14079 PR tree-optimization/57361
14080 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14082 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14084 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14086 (aarch64_simd_combine<mode>): New instruction expansion.
14087 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14088 function prototype.
14089 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14090 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14092 2013-06-12 Jan Hubicka <jh@suse.cz>
14094 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14095 decl has when in streaming stage.
14096 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14097 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14099 2013-06-12 Roland Stigge <stigge@antcom.de>
14102 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14104 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14106 PR tree-optimization/57537
14107 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14108 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14110 2013-06-12 Richard Biener <rguenther@suse.de>
14112 * data-streamer.h (streamer_write_char_stream): CSE
14113 obs->current_pointer.
14114 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14115 streamer_write_char_stream manually and optimize the resulting loop.
14116 (streamer_write_hwi_stream): Likewise.
14118 2013-06-12 Jan Hubicka <jh@suse.cz>
14120 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14121 * cgraph.h (varpool_create_empty_node): Declare.
14122 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14124 * symtab.c (symtab_unregister_node): Be lax about missin entries
14126 (symtab_get_node): Update comment.
14127 * varpool.c (varpool_create_empty_node): Break out from ...
14128 (varpool_node_for_decl): ... here.
14129 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14131 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14133 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14134 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14135 part. Use straight-line flow at the end.
14136 <COMPONENT_REF>: Remove superfluous else.
14137 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14139 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14142 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14143 target hook even for !TREE_PUBLIC decls. If no resolution info
14144 is available, return false for common and external decls.
14146 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14148 * config/rl78/constraints.md (U): New constraint.
14149 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14152 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14155 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14156 to allow returning address to AT_PLATFORM name.
14158 2013-06-11 Jan Hubicka <jh@suse.cz>
14160 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14161 * cgraph.h (symtab_node_base): Add weakref flag.
14162 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14163 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14164 (output_weakrefs): Use weakref flag.
14165 * fold-const.c (simple_operand_p): Handle WEAK.
14166 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14167 * ipa.c (varpool_externally_visible_p): Drop weakref.
14168 (function_and_variable_visibility): Update comment; fix weakref
14169 sanity checks; do not clear DECL_WEAK on them.
14170 * lto-cgraph.c (lto_output_node): update.
14171 (lto_output_varpool_node): Update.
14172 (input_overwrite_node): Update.
14173 (input_node): Update.
14174 (input_varpool_node): Update.
14175 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14176 (lto_symtab_merge_symbols): Add sanity check.
14177 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14178 * passes.c (rest_of_decl_compilation): Set static flag, too.
14179 * symtab.c (dump_symtab_base): Dump weakref.
14180 (verify_symtab_base): Sanity check weakrefs.
14181 (symtab_make_decl_local): Remove duplicated code.
14182 (symtab_alias_ultimate_target): Simplify.
14183 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14185 2013-06-11 Tom de Vries <tom@codesourcery.com>
14187 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14188 sequence_vect == NULL.
14190 2013-06-11 DJ Delorie <dj@redhat.com>
14192 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14193 (rl78_unwind_word_mode): New.
14195 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14197 * final.c (debug_prefix_maps): Make static.
14199 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14201 * function.c (initial_trampoline): Remove stray copy.
14203 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14205 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14207 2013-06-11 Martin Jambor <mjambor@suse.cz>
14209 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14210 within bounds at the beginning of the function.
14212 2013-06-11 Alan Modra <amodra@gmail.com>
14214 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14216 (get_named_section): Don't NULL !DECL_P decl.
14218 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14220 * doc/invoke.texi (core-avx2): Document.
14222 (atom): Updated with MOVBE.
14224 2013-06-11 Richard Biener <rguenther@suse.de>
14226 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14228 2013-06-11 Anton Blanchard <anton@samba.org>
14230 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14231 correct shift value in little-endian mode.
14233 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14236 * varasm.c (get_variable_align): Move #endif to the right place.
14238 2013-06-10 Cary Coutant <ccoutant@google.com>
14240 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14241 for hash so that hash table traversal order is deterministic.
14243 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14244 Pat Haugen <pthaugen@us.ibm.com>
14245 Peter Bergner <bergner@vnet.ibm.com>
14247 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14248 of vectors in GPRS if they are direct moves or quad word load or
14251 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14253 (direct_move_p): Likewise.
14254 (quad_load_store_p): Likewise.
14256 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14257 classes into bins based on the physical register type.
14258 (reg_class_to_reg_type): Likewise.
14259 (IS_STD_REG_TYPE): Likewise.
14260 (IS_FP_VECT_REG_TYPE): Likewise.
14261 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14262 use direct move instructions.
14263 (reload_gpr_vsx): Likewise.
14264 (reload_vsx_gpr): Likewise.
14265 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14266 information that is a simplification of register classes. Also
14267 precalculate direct move reload helpers.
14268 (direct_move_p): New function to return true if the operation can
14269 be done as a direct move instruciton.
14270 (quad_load_store_p): New function to return true if the operation
14271 is a quad memory operation.
14272 (rs6000_legitimize_address): If quad memory, only allow register
14273 indirect for TImode addresses.
14274 (rs6000_legitimate_address_p): Likewise.
14275 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14276 (rs6000_reload_register_type): Likewise.
14277 (register_to_reg_type): Return register type.
14278 (rs6000_secondary_reload_simple_move): New helper function for
14279 secondary reload and secondary memory needed to identify anything
14280 that is a simple move, and does not need reloading.
14281 (rs6000_secondary_reload_direct_move): New helper function for
14282 secondary reload to identify cases that can be done with several
14283 instructions via the direct move instructions.
14284 (rs6000_secondary_reload_move): New helper function for secondary
14285 reload to identify moves between register types that can be done.
14286 (rs6000_secondary_reload): Add support for quad memory operations
14287 and for direct move.
14288 (rs6000_secondary_memory_needed): Likewise.
14289 (rs6000_debug_secondary_memory_needed): Change argument names.
14290 (rs6000_output_move_128bit): New function to return the move to
14291 use for 128-bit moves, including knowing about the various
14292 limitations of quad memory operations.
14294 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14295 memory operations. call rs6000_output_move_128bit for the actual
14296 instruciton(s) to generate.
14297 (vsx_movti_64bit): Likewise.
14299 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14300 (UNSPEC_P8V_MTVSRWZ): Likewise.
14301 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14302 (UNSPEC_P8V_MTVSRD): Likewise.
14303 (UNSPEC_P8V_XXPERMDI): Likewise.
14304 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14305 (UNSPEC_FUSION_GPR): Likewise.
14306 (FMOVE128_GPR): New iterator for direct move.
14307 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14308 (f32_sv): Likewise.
14309 (f32_dm): Likewise.
14310 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14311 loads and direct move instructions.
14312 (zero_extendsidi2_lfiwzx): Likewise.
14313 (extendsidi2_lfiwax): Likewise.
14314 (extendsidi2_nocell): Likewise.
14315 (floatsi<mode>2_lfiwax): Likewise.
14316 (lfiwax): Likewise.
14317 (floatunssi<mode>2_lfiwzx): Likewise.
14318 (lfiwzx): Likewise.
14319 (fix_trunc<mode>_stfiwx): Likewise.
14320 (fixuns_trunc<mode>_stfiwx): Likewise.
14321 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14322 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14323 (parity<mode>2_cmpb): Set length/type attr.
14324 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14325 for 'mr.' to fast_compare.
14326 (bpermd_<mode>): Change type attr to popcnt.
14327 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14328 (p8_mtvsrwz_1): Likewise.
14329 (p8_mtvsrwz_2): Likewise.
14330 (reload_fpr_from_gpr<mode>): Likewise.
14331 (p8_mtvsrd_1): Likewise.
14332 (p8_mtvsrd_2): Likewise.
14333 (p8_xxpermdi_<mode>): Likewise.
14334 (reload_vsx_from_gpr<mode>): Likewise.
14335 (reload_vsx_from_gprsf): Likewise.
14336 (p8_mfvsrd_3_<mode>): LIkewise.
14337 (reload_gpr_from_vsx<mode>): Likewise.
14338 (reload_gpr_from_vsxsf): Likewise.
14339 (p8_mfvsrd_4_disf): Likewise.
14340 (multi-word GPR splits): Do not split direct moves or quad memory
14343 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14345 * tree-into-ssa.c (interesting_blocks): Make static.
14347 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14350 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14351 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14352 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14353 (get_variable_align): New function.
14354 (get_variable_section, emit_bss, emit_common,
14355 assemble_variable_contents, place_block_symbol): Use
14356 get_variable_align instead of DECL_ALIGN.
14357 (assemble_noswitch_variable): Add align argument, use it
14358 instead of DECL_ALIGN.
14359 (assemble_variable): Adjust caller. Use get_variable_align
14360 instead of DECL_ALIGN.
14361 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14363 (DATA_ABI_ALIGNMENT): Define.
14364 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14365 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14366 opt is false, only return the psABI mandated alignment increase.
14367 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14368 (DATA_ABI_ALIGNMENT): ... this.
14369 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14370 (DATA_ABI_ALIGNMENT): ... this.
14371 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14372 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14373 (DATA_ABI_ALIGNMENT): ... this.
14374 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14375 * doc/tm.texi: Regenerated.
14377 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14379 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14380 cmp_code to construct REG_EQUAL note.
14382 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14385 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14386 that operands[2] doesn't overlap with operands[0].
14388 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14389 Jan Hubicka <jh@suse.cz>
14391 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14392 hack to mark symbols as used.
14394 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14396 PR rtl-optimization/57559
14397 * lra-constraints.c (process_alt_operands): Don't discourage
14398 memory with known offset for offsetable memory constraint.
14399 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14401 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14403 * varasm.c (struct oc_local_state): Reorder fields.
14404 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14405 and adjust accordingly.
14406 (output_constructor): Reorder initialization code and adjust call to
14407 output_constructor_bitfield.
14409 2013-06-07 Jan Hubicka <jh@suse.cz>
14411 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14413 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14415 * tree-object-size.c (unknown): Make const.
14417 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14419 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14420 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14421 for last alternative in the cpu_facility attribute.
14423 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14426 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14427 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14428 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14429 * config/arm/constraints.md (Dg): New constraint.
14430 * config/arm/neon.md (xordi3_neon): Remove.
14431 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14432 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14434 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14436 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14437 Clean up alternatives.
14439 2013-06-07 Alan Modra <amodra@gmail.com>
14441 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14444 2013-06-07 Alan Modra <amodra@gmail.com>
14446 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14448 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14450 * config/arm/constraints.md (Df): New constraint.
14451 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14452 Correct length attribute for last two alternatives.
14454 2013-06-07 Alan Modra <amodra@gmail.com>
14456 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14457 override user -mfp-in-toc.
14458 (offsettable_ok_by_alignment): Consider just the current access
14459 rather than the whole object, unless BLKmode. Handle
14460 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14461 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14462 for -mcmodel=medium.
14463 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14464 override user -mfp-in-toc or -msum-in-toc. Default to
14465 -mno-fp-in-toc for -mcmodel=medium.
14467 2013-06-06 DJ Delorie <dj@redhat.com>
14469 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14470 TARGET_VALID_POINTER_MODE.
14472 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14473 Pat Haugen <pthaugen@us.ibm.com>
14474 Peter Bergner <bergner@vnet.ibm.com>
14476 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14477 Document new power8 builtins.
14479 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14480 condition code register, to allow 128-bit logical operations to be
14481 done in the VSX or GPR registers.
14482 (nor<mode>3): Use the canonical form for nor.
14483 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14484 vclz*, and vpopcnt* vector instructions.
14485 (nand<mode>3): Likewise.
14486 (orc<mode>3): Likewise.
14487 (clz<mode>2): LIkewise.
14488 (popcount<mode>2): Likewise.
14490 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14491 that only the GPRs are recognized.
14493 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14494 support for new power8 builtins.
14496 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14498 (xscvdpspn): Likewise.
14504 (vpopcnt): Likewise.
14505 (vpopcntb): Likewise.
14506 (vpopcnth): Likewise.
14507 (vpopcntw): Likewise.
14508 (vpopcntd): Likewise.
14510 (vmrgew): Likewise.
14511 (vmrgow): Likewise.
14513 (eqv_v16qi3): Likewise.
14514 (eqv_v8hi3): Likewise.
14515 (eqv_v4si3): Likewise.
14516 (eqv_v2di3): Likewise.
14517 (eqv_v4sf3): Likewise.
14518 (eqv_v2df3): Likewise.
14520 (nand_v16qi3): Likewise.
14521 (nand_v8hi3): Likewise.
14522 (nand_v4si3): Likewise.
14523 (nand_v2di3): Likewise.
14524 (nand_v4sf3): Likewise.
14525 (nand_v2df3): Likewise.
14527 (orc_v16qi3): Likewise.
14528 (orc_v8hi3): Likewise.
14529 (orc_v4si3): Likewise.
14530 (orc_v2di3): Likewise.
14531 (orc_v4sf3): Likewise.
14532 (orc_v2df3): Likewise.
14534 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14535 allow power8 quad mode in 64-bit.
14536 (rs6000_builtin_vectorized_function): Add support to vectorize
14537 ISA 2.07 count leading zeros, population count builtins.
14538 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14539 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14540 (builtin_function_type): Add vgbbd builtin function which takes an
14542 (altivec_expand_vec_perm_const): Add support for new power8 merge
14545 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14546 that does not include TImdoe for use with 32-bit.
14547 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14549 (UNSPEC_VSX_CVDPSPN): Likewise.
14550 (vsx_xscvdpspn): Likewise.
14551 (vsx_xscvspdpn): Likewise.
14552 (vsx_xscvdpspn_scalar): Likewise.
14553 (vsx_xscvspdpn_directmove): Likewise.
14554 (vsx_and<mode>3): Split logical operations into 32-bit and
14555 64-bit. Add support to do logical operations on TImode as well as
14556 VSX vector types. Allow logical operations to be done in either
14557 VSX registers or in general purpose registers in 64-bit mode. Add
14558 splitters if GPRs were used. For AND, add clobber of CCmode to
14559 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14561 (vsx_and<mode>3_32bit): Likewise.
14562 (vsx_and<mode>3_64bit): Likewise.
14563 (vsx_ior<mode>3): Likewise.
14564 (vsx_ior<mode>3_32bit): Likewise.
14565 (vsx_ior<mode>3_64bit): Likewise.
14566 (vsx_xor<mode>3): Likewise.
14567 (vsx_xor<mode>3_32bit): Likewise.
14568 (vsx_xor<mode>3_64bit): Likewise.
14569 (vsx_one_cmpl<mode>2): Likewise.
14570 (vsx_one_cmpl<mode>2_32bit): Likewise.
14571 (vsx_one_cmpl<mode>2_64bit): Likewise.
14572 (vsx_nor<mode>3): Likewise.
14573 (vsx_nor<mode>3_32bit): Likewise.
14574 (vsx_nor<mode>3_64bit): Likewise.
14575 (vsx_andc<mode>3): Likewise.
14576 (vsx_andc<mode>3_32bit): Likewise.
14577 (vsx_andc<mode>3_64bit): Likewise.
14578 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14579 and xxlorc instructions.
14580 (vsx_eqv<mode>3_64bit): Likewise.
14581 (vsx_nand<mode>3_32bit): Likewise.
14582 (vsx_nand<mode>3_64bit): Likewise.
14583 (vsx_orc<mode>3_32bit): Likewise.
14584 (vsx_orc<mode>3_64bit): Likewise.
14586 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14588 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14590 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14591 (p8_vmrgow): Likewise.
14592 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14593 GPRs to be split under VSX.
14594 (p8v_clz<mode>2): Add power8 count leading zero support.
14595 (p8v_popcount<mode>2): Add power8 population count support.
14596 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14599 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14602 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14604 (vec_nand): Likewise.
14605 (vec_vclz): Likewise.
14606 (vec_vclzb): Likewise.
14607 (vec_vclzd): Likewise.
14608 (vec_vclzh): Likewise.
14609 (vec_vclzw): Likewise.
14610 (vec_vgbbd): Likewise.
14611 (vec_vmrgew): Likewise.
14612 (vec_vmrgow): Likewise.
14613 (vec_vpopcnt): Likewise.
14614 (vec_vpopcntb): Likewise.
14615 (vec_vpopcntd): Likewise.
14616 (vec_vpopcnth): Likewise.
14617 (vec_vpopcntw): Likewise.
14619 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14621 PR rtl-optimization/57468
14622 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14625 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14627 PR rtl-optimization/57459
14628 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14629 type when setting live regs.
14631 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14633 * config/s390/s390.opt (mlra): New option.
14634 * config/s390/s390.c (s390_decompose_address): Check displacement
14635 for all registers for LRA.
14636 (s390_secondary_reload): Don't used secondary reloads for LRA.
14637 (s390_lra_p): New function.
14638 (TARGET_LRA_P): Define.
14639 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14640 of attribute cpu_facility to zarch for the last alternative.
14641 (*cmpmem_short): Ditto.
14643 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14645 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14646 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14647 (arm_expand_prologue): Likewise.
14649 2013-06-06 Teresa Johnson <tejohnson@google.com>
14652 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14653 as this is now done by redirect_edge_and_branch_force.
14654 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14655 barriers, and fix interaction with splitting.
14656 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14657 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14658 reflect changes made in the routine.
14659 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14660 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14661 since this is called in cfglayout mode, and replace partition fixup
14662 with assert as that is now done by force_nonfallthru_and_redirect.
14663 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14664 already be marked with region crossing note.
14665 (insert_section_boundary_note): Make non-static, gate on flag
14666 has_bb_partition, rewrite to also check for multiple partitions.
14667 (rest_of_handle_reorder_blocks): Remove call to
14668 insert_section_boundary_note, now done later during free_cfg.
14669 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14670 * bb-reorder.h (insert_section_boundary_note): Declare.
14671 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14672 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14673 invoke insert_section_boundary_note.
14674 (try_redirect_by_replacing_jump): Remove unnecessary
14675 check for region crossing note.
14676 (fixup_partition_crossing): New function.
14677 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14678 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14679 in non-cfglayout mode.
14680 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14681 remove old code that tried to do this. Emit barrier correctly
14682 when we are in cfglayout mode.
14683 (last_bb_in_partition): New function.
14684 (rtl_split_edge): Correctly fixup partition boundaries.
14685 (commit_one_edge_insertion): Remove old code that tried to
14686 fixup region crossing edge since this is now handled in
14687 split_block, and set up insertion point correctly since
14688 block may now end in a jump.
14689 (verify_hot_cold_block_grouping): Guard against checking when not in
14690 linearized RTL mode.
14691 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14693 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14694 rtl_verify_flow_info, so not called in cfglayout mode.
14695 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14696 (fixup_reorder_chain): Remove old code that attempted to fixup region
14697 crossing note as this is now handled in force_nonfallthru_and_redirect.
14698 (duplicate_insn_chain): Don't duplicate switch section notes.
14699 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14701 * basic-block.h (emit_barrier_after_bb): Declare.
14703 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14705 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14706 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14707 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14709 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14711 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14713 * config/arm/ldmstm.md: Regenerate.
14715 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14717 * config/arm/sync.md (atomic_loaddi_1):
14718 Disable predication for arm_restrict_it.
14719 (arm_load_exclusive<mode>): Likewise.
14720 (arm_load_exclusivesi): Likewise.
14721 (arm_load_exclusivedi): Likewise.
14722 (arm_load_acquire_exclusive<mode>): Likewise.
14723 (arm_load_acquire_exclusivesi): Likewise.
14724 (arm_load_acquire_exclusivedi): Likewise.
14725 (arm_store_exclusive<mode>): Likewise.
14726 (arm_store_exclusive<mode>): Likewise.
14727 (arm_store_release_exclusivedi): Likewise.
14728 (arm_store_release_exclusive<mode>): Likewise.
14730 2013-06-06 Richard Biener <rguenther@suse.de>
14732 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14734 (lto_tag_is_tree_code_p): Adjust.
14735 (lto_tag_is_gimple_code_p): Likewise.
14736 (lto_gimple_code_to_tag): Likewise.
14737 (lto_tag_to_gimple_code): Likewise.
14738 (lto_tree_code_to_tag): Likewise.
14739 (lto_tag_to_tree_code): Likewise.
14740 * data-streamer.h (streamer_write_hwi_in_range): Use
14741 uhwi streaming to stream the normalized range.
14742 (streamer_read_hwi_in_range): Likewise.
14744 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14746 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14747 (predicable_short_it): Likewise.
14748 (predicated): Likewise.
14749 (enabled): Handle above.
14750 (define_cond_exec): Set predicated attribute to yes.
14752 2013-06-05 Mike Stump <mikestump@comcast.net>
14754 * gdbinit.in (__FUNCTION__): Add.
14756 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14758 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14759 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14761 2013-06-05 Jan Hubicka <jh@suse.cz>
14763 * varasm.c (mark_decl_referenced): Revert the removal until targets
14766 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14768 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14769 instead of mark_decl_referenced.
14771 2013-06-05 Jan Hubicka <jh@suse.cz>
14773 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14774 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14775 and symtab_used_from_object_file_p.
14776 (cgraph_make_node_local_1): Clear forced_by_abi.
14777 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14778 * cgraph.h (symtab_node_base): Add forced_by_abi.
14779 (decide_is_variable_needed): Remove.
14780 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14781 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14782 (decide_is_symbol_needed): ... this one; handle symbols in general;
14783 always analyze virtuals; honnor forced_by_abi.
14784 (cgraph_finalize_function): Update.
14785 (varpool_finalize_decl): Update.
14786 (symbol_defined_and_needed): Remove.
14787 (analyze_functions): Update.
14788 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14789 output_refs, input_overwrite_node): Handle forced_by_abi.
14790 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14791 (address_taken_from_non_vtable_p): ... this one.
14792 (comdat_can_be_unshared_p_1): New function.
14793 (cgraph_comdat_can_be_unshared_p): Rename to ...
14794 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14795 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14796 (function_and_variable_visibility): Clear forced_by_abi as needed.
14797 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14798 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
14799 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14800 * varpool.c (decide_is_variable_needed): Remove.
14802 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14804 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14805 (arm_option_override): Override arm_restrict_it where appropriate.
14806 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14807 * config/arm/arm.opt (mrestrict-it): New command-line option.
14808 * doc/invoke.texi: Document -mrestrict-it.
14810 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14812 * tsan.c (tsan_atomic_table): Make const.
14814 2013-06-05 Richard Biener <rguenther@suse.de>
14816 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14817 index associated with the tree we are supposed to replace.
14818 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14819 TREE_ASM_WRITTEN as zero for everything but SSA names.
14821 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14823 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14825 2013-06-04 Jan Hubicka <jh@suse.cz>
14827 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14828 (input_node, input_varpool_node): Handle correctly external same
14830 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14831 nodes at ltrans stage.
14833 2013-06-04 Jan Hubicka <jh@suse.cz>
14835 * ipa-inline.c (update_caller_keys): Fix availability test.
14836 (update_callee_keys): Likewise.
14837 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14838 to follow ELF standard.
14840 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14842 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14843 (mips64r5900el-*-elf*): New configurations.
14844 * config/mips/mips-cpus.def (r5900): New processor.
14845 * config/mips/mips-tables.opt: Regenerate.
14846 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14847 (mips_issue_rate): Handle PROCESSOR_R5900.
14848 (mips_reorg_process_insns): Force reorder mode for the R5900.
14849 * config/mips/mips.h (TARGET_MIPS5900): Define.
14850 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14852 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14854 * config/mips/mips.md (processor): Add r5900.
14855 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14857 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14859 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14860 into function to generate MOVI instruction.
14861 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
14862 (aarch64_preferred_simd_mode): Turn into wrapper.
14863 (aarch64_output_scalar_simd_mov_immediate): New function.
14864 * config/aarch64/aarch64-protos.h: Add prototype for above.
14866 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14868 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14869 element_char member.
14870 (sizetochar): Return signed char.
14871 (aarch64_simd_valid_immediate): Remove elchar and other
14872 unnecessary variables.
14873 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14874 Calculate element_char as required.
14875 * config/aarch64/aarch64-protos.h: Update and move prototype
14876 for aarch64_output_simd_mov_immediate.
14877 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14880 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14882 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14883 information completed by aarch64_simd_valid_immediate.
14884 (aarch64_legitimate_constant_p): Update arguments.
14885 (aarch64_simd_valid_immediate): Work with struct rather than many
14887 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14888 (aarch64_simd_make_constant): Update arguments.
14889 (aarch64_output_simd_mov_immediate): Work with struct rather than
14890 many pointers. Output immediate directly rather than as operand.
14891 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14893 * config/aarch64/constraints.md (Dn): Update arguments.
14895 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14897 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14899 (aarch64_simd_immediate_valid_for_move): Remove.
14900 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14901 (aarch64_simd_make_constant): Update call.
14902 (aarch64_output_simd_mov_immediate): Update call.
14903 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14905 * config/aarch64/constraints.md (Dn): Update call.
14907 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14909 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14910 return type to bool for prototype.
14911 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14912 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14913 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14915 2013-06-04 Catherine Moore <clm@codesourcery.com>
14917 * config/mips/mips.opt (meva): New.
14918 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
14919 (ASM_SPEC): Handle -meva.
14920 * doc/invoke.texi (meva): Document.
14922 2013-06-04 Alan Modra <amodra@gmail.com>
14924 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14927 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14929 * rtl.def: Add extra fourth optional field to define_cond_exec.
14930 * gensupport.c (process_one_cond_exec): Process attributes from
14932 * doc/md.texi: Document fourth field in define_cond_exec.
14934 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14936 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14937 out the processing order as in store_bit_field_1.
14939 2013-06-04 Jan Hubicka <jh@suse.cz>
14941 PR middle-end/57500
14942 * cgraphunit.c (cgraph_process_same_body_aliases): Create
14943 non-VAR_DECL node if it does not exist yet.
14945 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
14947 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
14948 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
14949 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
14950 target_cpu_default setting.
14952 2013-06-03 Teresa Johnson <tejohnson@google.com>
14954 * dumpfile.c (opt_info_switch_p): Change -fopt-info
14955 default to -fopt-info=optimized instead of all.
14956 * doc/invoke.texi: Ditto.
14957 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
14958 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
14959 (execute_vect_slp): Emit BB vectorization success under
14960 MSG_OPTIMIZED_LOCATIONS.
14961 * tree-vect-slp.c (vect_slp_transform_bb): Change
14962 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
14963 * tree-vect-loop.c (vect_transform_loop): Ditto.
14965 2013-06-03 Jason Merrill <jason@redhat.com>
14968 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
14969 Use TARGET_EXPR for C++.
14971 2013-06-03 Jakub Jelinek <jakub@redhat.com>
14973 PR rtl-optimization/57268
14974 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
14975 if DEBUG_INSN_P (insn).
14978 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14980 PR rtl-optimization/57268
14981 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14982 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14984 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14986 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
14987 (ix86_avoid_lea_for_addr): Likewise.
14988 (exact_dependency_1): Likewise.
14989 (ix86_adjust_cost): Likewise.
14990 (swap_top_of_ready_list): Fix formatting and !reload_completed check
14992 (do_reorder_for_imul): Fix typo, formatting and
14993 !reload_completed check removed.
14994 (ix86_sched_reorder): Fix typo and formatting.
14995 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
14998 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15000 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15002 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15004 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15005 <STRING_CST>: Likewise.
15006 <VECTOR_CST>: Likewise.
15008 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15009 Mikael Morin <mikael@gcc.gnu.org>
15011 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15012 * config.in: Regenerated.
15013 * configure: Regenerated.
15015 2013-06-01 Jan Hubicka <jh@suse.cz>
15017 PR middle-end/57366
15018 * cgraphunit.c (compile): When weakref is not supported,
15019 set up transparent aliases before final output pass.
15020 * varasm.c (assemble_alias): Do not try to do it here.
15022 2013-06-01 Jan Hubicka <jh@suse.cz>
15024 PR middle-end/57467
15025 * passes.c (for_per_function): Skip unanalyzed functions.
15027 2013-06-01 Jan Hubicka <jh@suse.cz>
15029 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15030 (lto_symtab_merge_symbols_1): ... this one.
15031 (lto_symtab_merge_cgraph_nodes): Rename to ...
15032 (lto_symtab_merge_symbols): ... this one; simplify.
15033 * cgraph.c (same_body_aliases_done): Rename to ...
15034 (cpp_implicit_aliases_done): ... this one.
15035 (cgraph_create_function_alias): Update.
15036 (cgraph_same_body_alias): Update.
15037 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15038 (verify_edge_corresponds_to_fndecl): Simplify.
15039 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15040 (cgraph_node): Remove same_body_alias.
15041 (varpool_node): Remove alias_of and extra_name_alias.
15042 (same_body_aliases_done): Rename to ..
15043 (cpp_implicit_aliases_done): ... this one.
15044 (symtab_alias_ultimate_target): Add default parameter.
15045 (symtab_resolve_alias): New function.
15046 (fixup_same_cpp_alias_visibility): Declare.
15047 (cgraph_function_node): Add default parameter.
15048 (cgraph_node_asm_name): Likewise.
15049 (cgraph_function_or_thunk_node): Add default parameter; do
15050 not ICE when it is NULL.
15051 (varpool_variable_node): Likewise.
15052 * tree-emutls.c (create_emultls_var): Update.
15053 (ipa_lower_emutls): Update.
15054 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15055 (cgraph_reset_node): Reset alias info.
15056 (cgraph_finalize_function): Update.
15057 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15058 (analyze_function): Simplify.
15059 (cgraph_process_same_body_aliases): Simplify.
15060 (analyze_functions): Fixup same body aliases.
15061 (handle_alias_pairs): Simplify.
15062 (assemble_thunk): Update.
15063 (assemble_thunks_and_aliases): Update.
15064 (output_weakrefs): Rewrite.
15065 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15066 (lto_output_varpool_node): Likewise.
15067 (compute_ltrans_boundary): Remve assert.
15068 (get_alias_symbol): New functoin.
15069 (input_node): Rewrite alias handling.
15070 (input_varpool_node): Likewise.
15071 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15072 * ipa.c (process_references): Handle weakrefs correctly.
15073 (symtab_remove_unreachable_nodes): Likewise.
15074 * trans-mem.c (get_cg_data): Update.
15075 (ipa_tm_create_version_alias): Update.
15076 (ipa_tm_execute): Update.
15077 * symtab.c (dump_symtab_base): Dump aliases.
15078 (verify_symtab_base): Verify aliases.
15079 (symtab_node_availability): New function.
15080 (symtab_alias_ultimate_target): Simplify.
15081 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15082 handle all the fixup cases.
15083 (symtab_resolve_alias): New function.
15084 * passes.c (ipa_write_summaries): Handle weakrefs.
15085 * varpool.c (varpool_analyze_node): Simplify.
15086 (assemble_aliases): Update.
15087 (varpool_create_variable_alias): Simplify.
15088 (varpool_extra_name_alias): Simplify.
15089 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15090 (lto_symtab_merge_symbols): ... this one.
15092 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15095 PR rtl-optimization/57268
15096 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15097 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15099 2013-06-01 Tobias Burnus <burnus@net-b.de>
15101 Partially reverted:
15102 2013-05-31 Tobias Burnus <burnus@net-b.de>
15104 PR middle-end/57073
15105 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15108 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15110 PR rtl-optimization/57268
15111 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15112 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15114 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15116 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15117 unordered comparison operators when -fno-trapping-math is in effect
15119 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15120 and implement unordered comparison operators properly on the e500.
15122 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15124 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15125 for constant scalar integers.
15126 (simplify_relational_operation_1): Likewise.
15128 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15130 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15131 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15134 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15135 Igor Zamyatin <igor.zamyatin@intel.com>
15137 Silvermont (SLM) architecture performance tuning.
15138 * config/i386/i386.h (enum ix86_tune_indices): Add
15139 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15140 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15142 * config/i386/i386.c (initial_ix86_tune_features)
15143 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15144 (ix86_lea_outperforms): Handle Silvermont tuning.
15145 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15147 (ix86_use_lea_for_mov): Likewise.
15148 (ix86_avoid_lea_for_addr): Likewise.
15149 (ix86_lea_for_add_ok): Likewise.
15150 (exact_dependency_1): New function.
15151 (exact_store_load_dependency): Likewise.
15152 (ix86_adjust_cost): Handle Silvermont tuning.
15153 (do_reoder_for_imul): Likewise.
15154 (swap_top_of_ready_list): New function.
15155 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15157 * config/i386/i386.md (peepholes that split memory operand in fp
15160 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15162 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15163 Remove un-necessary braces.
15165 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15167 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15168 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15170 2013-05-31 Tobias Burnus <burnus@net-b.de>
15172 PR middle-end/57073
15173 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15175 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15178 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15179 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15180 * config/arm/neon.md (iordi3_neon): Remove.
15181 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15182 * config/arm/predicates.md (imm_for_neon_logic_operand):
15183 Move to earlier in the file.
15184 (neon_logic_op2): Likewise.
15185 (arm_iordi_operand_neon): New predicate.
15187 2013-05-31 Richard Biener <rguenther@suse.de>
15189 PR tree-optimization/57478
15190 PR tree-optimization/57453
15191 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15194 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15196 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15197 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15199 2013-05-30 Tobias Burnus <burnus@net-b.de>
15200 Thomas Koenig <tkoenig@gcc.gnu.org>
15202 PR middle-end/57073
15203 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15204 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15206 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15208 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15210 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15212 * target.def (register_usage_leveling_p): New hook.
15213 * targhooks.c (default_register_usage_leveling_p): New.
15214 * targhooks.h (default_register_usage_leveling_p): New prototype.
15215 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15216 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15217 * doc/tm.texi: Update.
15218 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15220 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15222 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15223 (*insv_reg<mode>): New define_insn.
15225 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15227 PR rtl-optimization/57439
15228 * postreload.c (move2add_valid_value_p): Check that we have
15229 a zero subreg_regno_offset when accessing the register in
15230 the requested mode.
15232 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15233 Igor Zamyatin <igor.zamyatin@intel.com>
15235 Silvermont (SLM) architecture pipeline model, tuning and
15237 * config.gcc: Add slm config options and target.
15239 * config/i386/slm.md: New.
15241 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15243 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15245 (ix86_target_macros_internal): Likewise.
15247 * config/i386/i386.c (slm_cost): New cost.
15248 (m_SLM): New macro flag.
15249 (initial_ix86_tune_features): Set m_SLM.
15250 (x86_accumulate_outgoing_args): Likewise.
15251 (x86_arch_always_fancy_math_387): Likewise.
15252 (processor_target_table): Add slm cost.
15253 (cpu_names): Add slm cpu name.
15254 (x86_option_override_internal): Set SLM ISA.
15255 (ix86_issue_rate): New case PROCESSOR_SLM.
15256 (ia32_multipass_dfa_lookahead): Likewise.
15257 (fold_builtin_cpu): Add slm.
15259 * config/i386/i386.h (TARGET_SLM): New target macro.
15260 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15261 (processor_type): Add PROCESSOR_SLM.
15263 * config/i386/i386.md (cpu): Add new value "slm".
15264 (slm.md): Include slm.md.
15266 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15267 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15269 * config/arm/arm-protos.h: Add and update function protos.
15270 * config/arm/arm.c (use_simple_return_p): New added.
15271 (thumb2_expand_return): Check simple_return flag.
15272 * config/arm/arm.md: Add simple_return and conditional simple_return.
15273 * config/arm/iterators.md: Add iterator for return and simple_return.
15275 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15277 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15278 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15279 (arm_emit_vfp_multi_reg_pop): Likewise.
15280 (thumb2_emit_ldrd_pop): Likewise.
15281 (arm_expand_epilogue): Add misc REG_CFA notes.
15282 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15284 2013-05-29 Lawrence Crowl <crowl@google.com>
15286 * config/arm/t-arm: Update for below.
15288 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15289 Change type to hash_table. Update dependent calls and types.
15291 * config/i386/t-cygming: Update for below.
15293 * config/i386/t-interix: Update for below.
15295 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15296 Change type to hash_table. Update dependent calls and types.
15297 (i386_find_on_wrapper_list::wrappers): Likewise.
15299 * config/ia64/t-ia64: Update for below.
15301 * config/ia64/ia64.c (bundle_state_table):
15302 Change type to hash_table. Update dependent calls and types.
15304 * config/mips/mips.c (mips_reorg_process_insns::htab):
15305 Change type to hash_table. Update dependent calls and types.
15307 * config/sol2.c (solaris_comdat_htab):
15308 Change type to hash_table. Update dependent calls and types.
15310 * config/t-sol2: Update for above.
15312 2013-05-29 Teresa Johnson <tejohnson@google.com>
15314 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15315 functions are not yet marked as defined.
15317 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15318 Pat Haugen <pthaugen@us.ibm.com>
15319 Peter Bergner <bergner@vnet.ibm.com>
15321 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15325 (vrotl<mode>3): Likewise.
15326 (vashl<mode>3): Likewise.
15327 (vlshr<mode>3): Likewise.
15328 (vashr<mode>3): Likewise.
15330 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15331 support for power8 V2DI builtins.
15333 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15334 power8 V2DI builtins.
15335 (vupkhsw): Likewise.
15336 (vupklsw): Likewise.
15337 (vaddudm): Likewise.
15338 (vminsd): Likewise.
15339 (vmaxsd): Likewise.
15340 (vminud): Likewise.
15341 (vmaxud): Likewise.
15342 (vpkudum): Likewise.
15343 (vpksdss): Likewise.
15344 (vpkudus): Likewise.
15345 (vpksdus): Likewise.
15350 (vsubudm): Likewise.
15351 (vcmpequd): Likewise.
15352 (vcmpgtsd): Likewise.
15353 (vcmpgtud): Likewise.
15354 (vcmpequd_p): Likewise.
15355 (vcmpgtsd_p): Likewise.
15356 (vcmpgtud_p): Likewise.
15357 (vupkhsw): Likewise.
15358 (vupklsw): Likewise.
15359 (vaddudm): Likewise.
15360 (vmaxsd): Likewise.
15361 (vmaxud): Likewise.
15362 (vminsd): Likewise.
15363 (vminud): Likewise.
15364 (vpksdss): Likewise.
15365 (vpksdus): Likewise.
15366 (vpkudum): Likewise.
15367 (vpkudus): Likewise.
15372 (vsubudm): Likewise.
15374 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15375 support for power8 V2DI instructions.
15377 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15378 power8 V2DI instructions. Combine pack and unpack insns to use an
15379 iterator for each mode. Check whether a particular mode supports
15380 Altivec instructions instead of just checking TARGET_ALTIVEC.
15381 (UNSPEC_VPKUWUM): Likewise.
15382 (UNSPEC_VPKSHSS): Likewise.
15383 (UNSPEC_VPKSWSS): Likewise.
15384 (UNSPEC_VPKUHUS): Likewise.
15385 (UNSPEC_VPKSHUS): Likewise.
15386 (UNSPEC_VPKUWUS): Likewise.
15387 (UNSPEC_VPKSWUS): Likewise.
15388 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15389 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15390 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15391 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15392 (UNSPEC_VUPKHSB): Likewise.
15393 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15394 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15395 (UNSPEC_VUPKHSH): Likewise.
15396 (UNSPEC_VUPKLSB): Likewise.
15397 (UNSPEC_VUPKLSH): Likewise.
15399 (VI_char): Likewise.
15400 (VI_scalar): Likewise.
15401 (VI_unit): Likewise.
15403 (VP_small): Likewise.
15404 (VP_small_lc): Likewise.
15405 (VU_char): Likewise.
15406 (add<mode>3): Likewise.
15407 (altivec_vaddcuw): Likewise.
15408 (altivec_vaddu<VI_char>s): Likewise.
15409 (altivec_vadds<VI_char>s): Likewise.
15410 (sub<mode>3): Likewise.
15411 (altivec_vsubcuw): Likewise.
15412 (altivec_vsubu<VI_char>s): Likewise.
15413 (altivec_vsubs<VI_char>s): Likewise.
15414 (altivec_vavgs<VI_char>): Likewise.
15415 (altivec_vcmpbfp): Likewise.
15416 (altivec_eq<mode>): Likewise.
15417 (altivec_gt<mode>): Likewise.
15418 (altivec_gtu<mode>): Likewise.
15419 (umax<mode>3): Likewise.
15420 (smax<mode>3): Likewise.
15421 (umin<mode>3): Likewise.
15422 (smin<mode>3): Likewise.
15423 (altivec_vpkuhum): Likewise.
15424 (altivec_vpkuwum): Likewise.
15425 (altivec_vpkshss): Likewise.
15426 (altivec_vpkswss): Likewise.
15427 (altivec_vpkuhus): Likewise.
15428 (altivec_vpkshus): Likewise.
15429 (altivec_vpkuwus): Likewise.
15430 (altivec_vpkswus): Likewise.
15431 (altivec_vpks<VI_char>ss): Likewise.
15432 (altivec_vpks<VI_char>us): Likewise.
15433 (altivec_vpku<VI_char>us): Likewise.
15434 (altivec_vpku<VI_char>um): Likewise.
15435 (altivec_vrl<VI_char>): Likewise.
15436 (altivec_vsl<VI_char>): Likewise.
15437 (altivec_vsr<VI_char>): Likewise.
15438 (altivec_vsra<VI_char>): Likewise.
15439 (altivec_vsldoi_<mode>): Likewise.
15440 (altivec_vupkhsb): Likewise.
15441 (altivec_vupkhs<VU_char>): Likewise.
15442 (altivec_vupkls<VU_char>): Likewise.
15443 (altivec_vupkhsh): Likewise.
15444 (altivec_vupklsb): Likewise.
15445 (altivec_vupklsh): Likewise.
15446 (altivec_vcmpequ<VI_char>_p): Likewise.
15447 (altivec_vcmpgts<VI_char>_p): Likewise.
15448 (altivec_vcmpgtu<VI_char>_p): Likewise.
15449 (abs<mode>2): Likewise.
15450 (vec_unpacks_hi_v16qi): Likewise.
15451 (vec_unpacks_hi_v8hi): Likewise.
15452 (vec_unpacks_lo_v16qi): Likewise.
15453 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15454 (vec_unpacks_lo_v8hi): Likewise.
15455 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15456 (vec_pack_trunc_v8h): Likewise.
15457 (vec_pack_trunc_v4si): Likewise.
15458 (vec_pack_trunc_<mode>): Likewise.
15460 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15462 (vec_vmaxsd): Likewise.
15463 (vec_vmaxud): Likewise.
15464 (vec_vminsd): Likewise.
15465 (vec_vminud): Likewise.
15466 (vec_vpksdss): Likewise.
15467 (vec_vpksdus): Likewise.
15468 (vec_vpkudum): Likewise.
15469 (vec_vpkudus): Likewise.
15470 (vec_vrld): Likewise.
15471 (vec_vsld): Likewise.
15472 (vec_vsrad): Likewise.
15473 (vec_vsrd): Likewise.
15474 (vec_vsubudm): Likewise.
15475 (vec_vupkhsw): Likewise.
15476 (vec_vupklsw): Likewise.
15478 2013-05-29 Jan Hubicka <jh@suse.cz>
15480 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15481 flags; reorder rest of fields in more consistent way.
15482 (varpool_node): Remove analyzed, finalized and alias.
15483 (cgraph_ndoe): Likewise.
15484 (symtab_alias_ultimate_target): New function.
15485 (cgraph_function_node): Move offline.
15486 (cgraph_reset_node): Declare.
15487 (cgraph_comdat_can_be_unshared_p): Remove.
15488 (varpool_remove_initializer): Declare.
15489 (varpool_first_defined_variable, varpool_next_defined_variable
15490 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15491 (cgraph_function_with_gimple_body_p): Update.
15492 (varpool_all_refs_explicit_p): Update.
15493 (symtab_alias_target): New function.
15494 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15495 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15496 (cgraph_function_or_thunk_node): Simplify using
15497 symtab_alias_ultimate_target.
15498 (varpool_variable_node): Likewise.
15499 * cgraph.c (cgraph_create_function_alias): Update.
15500 (cgraph_add_thunk): Update.
15501 (cgraph_remove_node): Update.
15502 (dump_cgraph_node): Do not dump removed flags.
15503 (cgraph_function_body_availability): Update.
15504 (cgraph_propagate_frequency): Update.
15505 (verify_cgraph_node): Check sanity of local flag.
15506 (cgraph_function_node): Move here from cgraph.h; revamp for
15507 cgraph_function_or_thunk_node.
15508 * lto-symtab.c (lto_varpool_replace_node): Update.
15509 (lto_symtab_resolve_can_prevail_p): Update.
15510 (lto_symtab_merge_cgraph_nodes): Update.
15511 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15512 propagate_constants_accross_call, devirtualization_time_bonus,
15513 ipcp_propagate_stage): Update.
15514 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15515 * ipa-inline-transform.c (clone_inlined_nodes,
15516 preserve_function_body_p): Update.
15517 * ipa-reference.c (propagate): Update.
15518 (write_node_summary_p): Update.
15519 * toplev.c (wrapup_global_declaration_2): Update.
15520 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15521 (analyze_function) ... this one.
15522 (cgraph_process_new_functions): Update.
15523 (cgraph_reset_node): Export.
15524 (cgraph_finalize_function): Update.
15525 (cgraph_add_new_function): Update.
15526 (process_function_and_variable_attributes): Update.
15527 (varpool_finalize_decl): Update.
15528 (symbol_finalized): Remove.
15529 (symbol_finalized_and_needed): Rename to ...
15530 (symbol_defined_and_needed): ... update.
15531 (cgraph_analyze_functions): Update.
15532 (handle_alias_pairs): Update.
15533 (mark_functions_to_output): Update.
15534 (assemble_thunk): Update.
15535 (output_in_order): Update.
15536 (output_weakrefs): Update.
15537 (finalize_compilation_unit): Update.
15538 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15539 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15540 input_node, input_varpool_node): Update.
15541 * dbxout.c (dbxout_expand_expr): Update.
15542 * cgraphclones.c (cgraph_clone_node): Update.
15543 (cgraph_copy_node_for_versioning): Update.
15544 (cgraph_materialize_clone): Update.
15545 (cgraph_materialize_all_clones): Update.
15546 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15547 propagate_pure_const, propagate_nothrow): Update.
15548 * lto-streamer-out.c (lto_output, write_symbol): Update.
15549 * ipa-utils.c (ipa_reverse_postorder): Update.
15550 * ipa-inline.c (can_inline_edge_p): Update.
15551 (update_caller_keys, ipa_inline): Update.
15552 * dwarf2out.c (reference_to_unused,
15553 premark_types_used_by_global_vars_helper): Update.
15554 * tree-eh.c (tree_could_trap_p): Update.
15555 * ipa-split.c (consider_split, execute_split_functions): Update.
15556 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15557 has_addr_references_p): Update; move ahead in file for better
15559 (process_references): Simplify.
15560 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15561 bodies are removed.
15562 (cgraph_comdat_can_be_unshared_p): Make static.
15563 (cgraph_externally_visible_p): Update.
15564 (varpool_externally_visible_p): Update.
15565 (function_and_variable_visibility): Update.
15566 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15567 ipa_tm_mark_force_output_node): Update.
15568 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15569 estimate_edge_devirt_benefit, inline_generate_summary,
15570 inline_write_summary): Update.
15571 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15572 * ipa-prop.c (ipa_compute_jump_functions): Update.
15573 (ipa_print_node_params, ipa_prop_read_section,
15574 ipa_update_after_lto_read, read_replacements_section): Update.
15575 * varasm.c (mark_decl_referenced): Update.
15576 (assemble_alias, dump_tm_clone_pairs): Update.
15577 * tree-inline.c (copy_bb): Update.
15578 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15580 * symtab.c (dump_symtab_base): Print new flags.
15581 (verify_symtab_base): Verify new flags.
15582 (symtab_alias_ultimate_target): New function.
15583 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15584 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15586 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15588 * i386.c (ix86_get_function_versions_dispatcher,
15589 ix86_generate_version_dispatcher_body): Update.
15590 (fold_builtin_cpu): Use varpool_add_new_variable.
15591 * varpool.c (varpool_remove_initializer): Break out from ...
15592 (varpool_remove_node): ... this one.
15593 (dump_varpool_node, varpool_node_for_asm,
15594 cgraph_variable_initializer_availability, varpool_analyze_node,
15595 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15596 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15599 2013-05-29 Jan Hubicka <jh@suse.cz>
15601 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15603 2013-05-29 Easwaran Raman <eraman@google.com>
15605 PR tree-optimization/57442
15606 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15607 when control exits the main loop.
15609 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15611 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15613 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15614 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15615 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15618 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15620 PR tree-optimization/57441
15621 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15622 Don't limit size of incr_vec to number of candidates.
15624 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15626 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15627 and mips16 directories.
15628 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15629 (MULTILIB_DIRNAMES): Ditto.
15630 (MULTILIB_EXCEPTIONS): Add new exceptions.
15631 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15632 (MULTILIB_DIRNAMES): Ditto.
15633 (MULTILIB_EXCEPTIONS): Add new exceptions.
15635 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15636 Marcus Shawcroft <marcus.shawcroft@arm.com>
15638 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15639 SYMBOL_TINY_ABSOLUTE.
15640 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15641 SYMBOL_TINY_ABSOLUTE.
15642 (aarch64_expand_mov_immediate): Likewise.
15643 (aarch64_classify_symbol): Likewise.
15644 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15645 Permit SYMBOL_TINY_ABSOLUTE.
15646 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15648 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15649 Marcus Shawcroft <marcus.shawcroft@arm.com>
15651 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15652 Refactor if/switch. Replace gcc_assert with if.
15654 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15656 * config/i386/i386.c (initial_ix86_tune_features): Enable
15657 FP Reassociation for AMD bdver1 and bdver2.
15659 2013-05-29 Martin Jambor <mjambor@suse.cz>
15661 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15662 and IMAGPART_EXPR do not occur within other handled_components.
15664 2013-05-29 Richard Biener <rguenther@suse.de>
15666 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15667 access on whether the use is in the BB we currently try to
15669 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15670 vectorize to vect_bb_slp_scalar_cost.
15672 2013-05-29 Richard Biener <rguenther@suse.de>
15674 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15675 computing scalar cost offsetted by stmts that are kept live
15677 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15678 for computation of scalar cost.
15680 2013-05-28 Steve Ellcey <sellcey@mips.com>
15682 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15684 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15686 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15687 array notation built-in reduction functions.
15688 * doc/passes.texi (Passes): Added documentation about changes done
15690 * doc/invoke.texi (C Dialect Options): Added documentation about
15691 the -fcilkplus flag.
15692 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15693 (BUILTINS_DEF): Depend on cilkplus.def.
15694 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15695 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15696 * cilkplus.def: New file.
15698 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15700 PR rtl-optimization/57439
15701 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15703 2013-05-28 Easwaran Raman <eraman@google.com>
15705 PR tree-optimization/57337
15706 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15707 (find_insert_point): Correctly identify the insertion point
15708 when two statements with the same UID is compared.
15710 2013-05-28 Richard Biener <rguenther@suse.de>
15712 PR tree-optimization/56787
15713 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15714 from the list of data references.
15715 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15717 (vect_analyze_loop_operations): Likewise.
15718 (vect_transform_loop): Remove clobbers.
15720 2013-05-28 Martin Jambor <mjambor@suse.cz>
15722 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15723 and REALPART_EXPRs have scalar type.
15725 2013-05-28 Richard Biener <rguenther@suse.de>
15727 PR tree-optimization/57411
15728 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15730 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15731 virtual operand propagation.
15733 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15735 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15736 destination register for bmasksi_vis.
15737 (vector_init_bshuffle): Likewise.
15738 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15740 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15742 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15743 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15744 mode if the instruction isn't available in the original mode.
15745 * config/sparc/sparc.opt (mfix-ut699): New option.
15746 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15747 (divdf3): Turn into expander.
15748 (divdf3_nofix): New insn.
15749 (divdf3_fix): Likewise.
15750 (divsf3): Disable if -mfix-ut699.
15751 (sqrtdf2): Turn into expander.
15752 (sqrtdf2_nofix): New insn.
15753 (sqrtdf2_fix): Likewise.
15754 (sqrtsf2): Disable if -mfix-ut699.
15756 2013-05-27 Richard Biener <rguenther@suse.de>
15758 PR middle-end/57412
15759 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15760 block for the new loop.
15762 2013-05-27 Richard Biener <rguenther@suse.de>
15764 PR tree-optimization/57343
15765 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15766 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15767 (number_of_iterations_cond): Do not build the folded tree.
15769 2013-05-27 Richard Biener <rguenther@suse.de>
15772 PR middle-end/57381
15773 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15774 OEP_CONSTANT_ADDRESS_OF retained.
15776 PR tree-optimization/57417
15777 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15778 for unchanged base.
15779 (set_ssa_val_to): Compare addresses using
15780 get_addr_base_and_unit_offset.
15782 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15784 PR rtl-optimization/56833
15785 * postreload.c (move2add_record_mode): New function.
15786 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15787 (move2add_use_add2_insn): Use move2add_record_sym_value.
15788 (move2add_use_add3_insn): Likewise.
15789 (reload_cse_move2add): Use move2add_valid_value_p and
15790 move2add_record_mode. Invalidate call-clobbered and REG_INC
15791 affected regs by setting reg_mode to VOIDmode.
15792 (move2add_note_store): Don't pretend the inside of a SUBREG is
15793 the actual destination. Invalidate single/leading registers by
15794 setting reg_mode to VOIDmode.
15795 Use move2add_record_sym_value, move2add_valid_value_p and
15796 move2add_record_mode.
15798 2013-05-27 Richard Biener <rguenther@suse.de>
15800 PR tree-optimization/57396
15801 * tree-affine.c (double_int_constant_multiple_p): Properly
15802 return false for val == 0 and div != 0.
15804 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15806 * config/mips/mips.h: Use #elif in preprocessor conditions.
15808 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15811 * config/mips/constraints.md (kl): New constraint.
15812 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15813 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15814 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15815 from LO for MIPS16.
15816 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15818 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15821 * config/mips/mips.c (mips_can_inline_p): New function.
15822 (TARGET_CAN_INLINE_P): Define.
15824 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15826 * sched-int.h (ds_t, dw_t): Make unsigned int.
15827 Fix documentation that describes how all the ds_t bits are used.
15828 Reserve the last bit for delayed-branch scheduling.
15829 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15830 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15831 (gen_dep_weak_1): Remove prototype.
15832 * sched-deps.c (get_dep_weak_1): Make static.
15833 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15834 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15835 * doc/tm.texi: Regenerate.
15836 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15838 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15841 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15843 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15844 Sandra Loosemore <sandra@codesourcery.com>
15846 * config.gcc (powerpc-*): Allow native for with-cpu.
15848 2013-05-24 Jeff Law <law@redhat.com>
15850 PR tree-optimization/57124
15851 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15852 conversion feeding a condition if the range has an overflow
15853 if -fstrict-overflow. Add warnings for when we do make the
15856 2013-05-24 Dehao Chen <dehao@google.com>
15858 * tree-cfg.c (locus_discrim_map): Fix the typo.
15859 (locus_discrim_hasher): Likewise.
15860 (locus_discrim_hasher::hash): Likewise.
15861 (locus_discrim_hasher::equal): Likewise.
15863 2013-05-24 Martin Jambor <mjambor@suse.cz>
15865 PR tree-optimization/57294
15866 * cgraph.h (ipa_record_stmt_references): Declare.
15867 * cgraphbuild.c (ipa_record_stmt_references): New function.
15868 (build_cgraph_edges): Use ipa_record_stmt_references.
15869 (rebuild_cgraph_edges): Likewise.
15870 (cgraph_rebuild_references): Likewise.
15871 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15872 associated with the old statement and build references from the
15873 newly built statements.
15874 * ipa-ref.c (ipa_remove_stmt_references): New function.
15875 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15877 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15879 * lra-constraints.c (emit_spill_move): Use smaller mode for
15881 (check_and_process_move): Consider mem-reg moves for secondary
15883 (curr_insn_transform): Don't lose insns emitted before for
15884 secondary memory moves.
15885 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15886 reg set up in the current insn.
15888 2013-05-24 Dehao Chen <dehao@google.com>
15890 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
15892 (locus_descrim_hasher::equal): Likewise.
15893 (build_gimple_cfg): New discriminator assignment algorithm.
15894 (make_edges): Likewise.
15895 (next_discriminator_for_locus): Likewise.
15896 (same_line_p): Likewise.
15897 (assign_discriminators): Likewise.
15898 (make_cond_expr_edges): Likewise.
15899 (make_gimple_switch_edges): Likewise.
15900 (make_goto_expr_edges): Likewise.
15901 (make_gimple_asm_edges): Likewise.
15903 2013-05-24 Ian Bolton <ian.bolton@arm.com>
15905 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15906 X format specifier to only display bottom 16 bits.
15907 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15908 immediate to match for operand 2, since it will be masked.
15910 2013-05-24 Richard Biener <rguenther@suse.de>
15912 PR tree-optimization/57287
15913 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15914 all SSA names that occur in abnormal PHIs.
15916 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15919 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15920 that index is not negative.
15922 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15924 PR rtl-optimization/55177
15925 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15926 (simplify_byte_swapping_operation): New.
15927 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15928 (simplify_relational_operation_1): Deal with BSWAP.
15930 2013-05-23 Richard Henderson <rth@redhat.com>
15933 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15934 (ix86_reorg): Call it.
15936 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
15939 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
15940 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
15941 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
15943 2013-05-23 Christian Bruel <christian.bruel@st.com>
15946 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
15948 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15949 Marcus Shawcroft <marcus.shawcroft@arm.com>
15951 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
15952 * config/aarch64/constraints.md (Usa): Remove.
15953 * doc/md.texi (AArch64 Usa): Remove.
15955 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15956 Marcus Shawcroft <marcus.shawcroft@arm.com>
15958 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
15959 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
15960 * config/aarch64/predicates.md (aarch64_const_address): Remove.
15961 (aarch64_mov_operand): Use aarch64_mov_operand_p.
15963 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
15965 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
15966 instruction (AdvSIMD).
15967 * config/aarch64/aarch64-builtins.c
15968 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
15969 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
15971 2013-05-23 Martin Jambor <mjambor@suse.cz>
15973 PR middle-end/57347
15974 * tree.h (contains_bitfld_component_ref_p): Declare.
15975 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
15976 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
15978 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
15979 not access a bit-field. Assert all final offsets are byte-aligned.
15981 2013-05-23 Richard Biener <rguenther@suse.de>
15983 PR tree-optimization/57380
15984 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
15985 least one invariant or re-used load.
15986 * passes.c (init_optimization_passes): Move pass_phiprop before
15989 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
15991 * config/aarch64/aarch64-simd.md
15992 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
15994 2013-05-23 Richard Biener <rguenther@suse.de>
15996 PR middle-end/57381
15997 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15998 OEP_CONSTANT_ADDRESS_OF retained.
16000 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16002 PR middle-end/57344
16003 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16004 don't lower unit. Handle unit not being always BITS_PER_WORD.
16006 2013-05-23 Richard Biener <rguenther@suse.de>
16008 PR rtl-optimization/57341
16009 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16010 instead of true_dependence.
16012 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16014 * bb-reorder.c (branch_threshold): Make const.
16015 (exec_threshold): Ditto.
16017 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16018 Pat Haugen <pthaugen@us.ibm.com>
16019 Peter Bergner <bergner@vnet.ibm.com>
16021 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16022 documentation for the power8 crypto builtins.
16024 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16026 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16027 macros for defining power8 builtin functions.
16028 (BU_P8V_AV_2): Likewise.
16029 (BU_P8V_AV_P): Likewise.
16030 (BU_P8V_VSX_1): Likewise.
16031 (BU_P8V_OVERLOAD_1): Likewise.
16032 (BU_P8V_OVERLOAD_2): Likewise.
16033 (BU_CRYPTO_1): Likewise.
16034 (BU_CRYPTO_2): Likewise.
16035 (BU_CRYPTO_3): Likewise.
16036 (BU_CRYPTO_OVERLOAD_1): Likewise.
16037 (BU_CRYPTO_OVERLOAD_2): Likewise.
16038 (XSCVSPDP): Fix typo, point to the correct instruction.
16039 (VCIPHER): Add power8 crypto builtins.
16040 (VCIPHERLAST): Likewise.
16041 (VNCIPHER): Likewise.
16042 (VNCIPHERLAST): Likewise.
16043 (VPMSUMB): Likewise.
16044 (VPMSUMH): Likewise.
16045 (VPMSUMW): Likewise.
16046 (VPERMXOR_V2DI): Likewise.
16047 (VPERMXOR_V4SI: Likewise.
16048 (VPERMXOR_V8HI: Likewise.
16049 (VPERMXOR_V16QI: Likewise.
16050 (VSHASIGMAW): Likewise.
16051 (VSHASIGMAD): Likewise.
16052 (VPMSUM): Likewise.
16053 (VPERMXOR): Likewise.
16054 (VSHASIGMA): Likewise.
16056 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16057 __CRYPTO__ if the crypto instructions are available.
16058 (altivec_overloaded_builtins): Add support for overloaded power8
16061 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16062 support for power8 crypto builtins.
16063 (builtin_function_type): Likewise.
16064 (altivec_init_builtins): Add support for builtins that take vector
16065 long long (V2DI) arguments.
16067 * config/rs6000/crypto.md: New file, define power8 crypto
16070 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16071 Pat Haugen <pthaugen@us.ibm.com>
16072 Peter Bergner <bergner@vnet.ibm.com>
16074 * doc/invoke.texi (Option Summary): Add power8 options.
16075 (RS/6000 and PowerPC Options): Likewise.
16077 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16078 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16079 wm, wn, wr documentation.
16081 * config/rs6000/constraints.md (wm): New constraint for VSX
16082 registers if direct move instructions are enabled.
16083 (wn): New constraint for no registers.
16084 (wq): New constraint for quad word even GPR registers.
16085 (wr): New constraint if 64-bit instructions are enabled.
16086 (wv): New constraint if power8 vector instructions are enabled.
16087 (wQ): New constraint for quad word memory locations.
16089 * config/rs6000/predicates.md (const_0_to_15_operand): New
16090 constraint for 0..15 for crypto instructions.
16091 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16092 as GPR and floating point registers.
16093 (int_reg_operand): New predicate to match only GPR registers.
16094 (base_reg_operand): New predicate to match base registers.
16095 (quad_int_reg_operand): New predicate to match even GPR registers
16096 for quad memory operations.
16097 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16098 operations in both GPR and VSX registers.
16099 (quad_memory_operand): New predicate for quad memory operations.
16100 (reg_or_indexed_operand): New predicate for direct move support.
16102 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16103 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16104 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16105 (POWERPC_MASKS): Add power8 options.
16106 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16109 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16110 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16112 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16113 (-mpower8-fusion): New power8 options.
16114 (-mpower8-fusion-sign): Likewise.
16115 (-mpower8-vector): Likewise.
16116 (-mcrypto): Likewise.
16117 (-mdirect-move): Likewise.
16118 (-mquad-memory): Likewise.
16120 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16121 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16123 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16124 (rs6000_debug_vector_unit): Add p8_vector.
16125 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16126 definitions. Also print fusion state.
16127 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16128 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16129 (rs6000_option_override_internal): Add support for power8.
16130 (rs6000_common_init_builtins): Add debugging for skipped builtins
16131 if -mdebug=builtin.
16132 (rs6000_adjust_cost): Add power8 support.
16133 (rs6000_issue_rate): Likewise.
16134 (insn_must_be_first_in_group): Likewise.
16135 (insn_must_be_last_in_group): Likewise.
16136 (force_new_group): Likewise.
16137 (rs6000_register_move_cost): Likewise.
16138 (rs6000_opt_masks): Likewise.
16140 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16141 power8 capable assembler, default to power7 options.
16142 (TARGET_DIRECT_MOVE): Likewise.
16143 (TARGET_CRYPTO): Likewise.
16144 (TARGET_P8_VECTOR): Likewise.
16145 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16146 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16147 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16148 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16149 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16150 (TARGET_XSCVDPSPN): Likewise.
16151 (TARGET_XSCVSPDPN): Likewsie.
16152 (TARGET_SYNC_HI_QI): Likewise.
16153 (TARGET_SYNC_TI): Likewise.
16154 (MASK_CRYPTO): Likewise.
16155 (MASK_DIRECT_MOVE): Likewise.
16156 (MASK_P8_FUSION): Likewise.
16157 (MASK_P8_VECTOR): Likewise.
16158 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16159 temporary used by some of the direct move instructions to get two FP
16160 temporary registers does not force creation of a stack frame.
16161 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16162 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16163 that any VSX registers are tieable, even if they are also an
16164 Altivec vector mode.
16165 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16166 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16167 (RS6000_BTM_CRYPTO): Likewise.
16168 (RS6000_BTM_COMMON): Likewise.
16170 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16171 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16172 (enum rs6000_vector): Add power8 vector support.
16174 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16178 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16179 (any_sibcall_could_use_r3): this and handle indirect calls.
16180 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16182 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16184 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16186 2013-05-22 Richard Biener <rguenther@suse.de>
16188 PR middle-end/57349
16189 * profile.c (branch_prob): Do not split blocks that are
16190 abnormally receiving from ECF_RETURNS_TWICE functions.
16192 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16194 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16195 address mode. Move pointer mode initialization to the same place.
16197 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16199 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16200 while it has any effect.
16202 2013-05-21 Easwaran Raman <eraman@google.com>
16204 PR tree-optimization/57322
16205 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16206 UID of the statement added to the BB to be 1.
16208 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16210 PR tree-optimization/57331
16211 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16212 of conversion from pointer type to integral type with integer.
16214 2013-05-21 Martin Jambor <mjambor@suse.cz>
16217 * ipa-prop.c (ipa_read_node_info): Process param_used and
16218 controlled_uses in the same order as when writing.
16220 2013-05-21 Magnus Granberg <baldrick@free.fr>
16223 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16225 2013-05-21 Richard Biener <rguenther@suse.de>
16227 PR tree-optimization/57318
16228 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16229 estimate stmts with side-effects as likely eliminated.
16231 2013-05-21 Richard Biener <rguenther@suse.de>
16233 PR tree-optimization/57330
16234 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16235 preserve the call stmts fntype.
16237 2013-05-21 Richard Biener <rguenther@suse.de>
16239 PR tree-optimization/57303
16240 * tree-ssa-sink.c (statement_sink_location): Improve killing
16241 stmt detection and properly handle self-assignments.
16243 2013-05-21 Christian Bruel <christian.bruel@st.com>
16245 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16246 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16247 registers. Set register size out of the PARALLEL loop.
16249 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16252 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16253 (*fmasf4, *fmasf4_media): New insns.
16255 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16257 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16258 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16259 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16260 (mips_idiv_insns): Update the comments to say that the returned
16261 instruction counts are in units of BASE_INSN_LENGTH.
16262 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16263 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16264 using 2 rather than 4 as the length of indirect MIPS16 and
16265 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16266 length of a NOP. Don't divide MIPS16 lengths by 2.
16267 (mips16_split_long_branches): Assume a branch is long if the
16268 length is greater than 4 rather than 8.
16269 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16270 rather than multiplying them by 2. Multiply instruction counts
16271 by BASE_INSN_LENGTH rather than 4.
16272 (*jump_mips16, tls_get_tp_mips16_<mode>)
16273 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16275 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16277 * config/mips/mips.md (extended_mips16): Remove branch case.
16278 (length): Remove duplicated extended_mips16 test.
16280 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16282 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16284 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16286 * recog.h (Recog_data): Rename to...
16287 (recog_data_d): ...this.
16288 (recog_data): Update accordingly.
16289 * recog.c (recog_data): Likewise.
16290 * reload.c (save_recog_data): Likewise.
16291 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16292 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16294 2013-05-17 Julian Brown <julian@codesourcery.com>
16296 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16297 found in a REG_EQUAL note, invalidate it.
16299 2013-05-17 Easwaran Raman <eraman@google.com>
16301 * tree-ssa-reassoc.c (find_insert_point): New function.
16302 (insert_stmt_after): Likewise.
16303 (get_def_stmt): Likewise.
16304 (ensure_ops_are_available): Likewise.
16305 (not_dominated_by): Likewise.
16306 (rewrite_expr_tree): Do not move statements beyond what is
16307 necessary. Remove call to swap_ops_for_binary_stmt...
16308 (reassociate_bb): ... and move it here.
16309 (build_and_add_sum): Assign UIDs for new statements.
16310 (linearize_expr): Likewise.
16311 (do_reassoc): Renumber gimple statement UIDs.
16313 2013-05-17 Jan Hubicka <jh@suse.cz>
16315 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16317 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16318 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16320 (output_weakrefs): Update.
16322 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16323 Martin Jambor <mjambor@suse.cz>
16325 PR middle-end/57276
16326 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16327 value that corresponds to the given aggval is found in values vector.
16329 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16331 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16332 sse, sse2, sse3, ssse3 and sse4a flags to options.
16334 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16336 * gengtype-state.c: (s_expr_writer): New class, to handle
16337 prettifying of output layout of s-expressions.
16338 (state_writer): New class, to write out gtype.state.
16339 (state_written_type_count): Move this variable into member data of
16341 (s_expr_writer::s_expr_writer): New code: constructor for new class
16342 (state_writer::state_writer(): ditto
16343 (s_expr_writer::write_new_line): New function
16344 (s_expr_writer::write_any_indent): ditto
16345 (s_expr_writer::begin_s_expr): ditto
16346 (s_expr_writer::end_s_expr): ditto
16347 (write_state_fileloc): convert to method of state_writer...
16348 (state_writer:: write_state_fileloc): ...and use methods of
16349 s_expr_writer to write indentation into the gtype.state output file
16350 to visually represent the hierarchical structure of the list
16352 (write_state_fields): ditto, renaming to...
16353 (state_writer::write_state_fields)
16354 (write_state_a_string): ditto, renaming to...
16355 (state_writer::write_state_a_string)
16356 (write_state_string_option): ditto, renaming to...
16357 (state_writer::write_state_string_option)
16358 (write_state_type_option): ditto, renaming to...
16359 (state_writer::write_state_type_option)
16360 (write_state_nested_option): ditto, renaming to...
16361 (state_writer::write_state_nested_option)
16362 (write_state_option): ditto, renaming to...
16363 (state_writer::write_state_option)
16364 (write_state_options): ditto, renaming to...
16365 (state_writer::write_state_options)
16366 (write_state_lang_bitmap): ditto, renaming to...
16367 (state_writer::write_state_lang_bitmap)
16368 (write_state_version): ditto, renaming to...
16369 (state_writer::write_state_version)
16370 (write_state_scalar_type): ditto, renaming to...
16371 (state_writer::write_state_scalar_type)
16372 (write_state_string_type): ditto, renaming to...
16373 (state_writer::write_state_string_type)
16374 (write_state_undefined_type): ditto, renaming to...
16375 (state_writer::write_state_undefined_type)
16376 (write_state_struct_union_type): ditto, renaming to...
16377 (state_writer::write_state_struct_union_type)
16378 (write_state_struct_type): ditto, renaming to...
16379 (state_writer::write_state_struct_type)
16380 (write_state_user_struct_type): ditto, renaming to...
16381 (state_writer::write_state_user_struct_type)
16382 (write_state_lang_struct_type): ditto, renaming to...
16383 (state_writer::write_state_lang_struct_type)
16384 (write_state_param_struct_type): ditto, renaming to...
16385 (state_writer::write_state_param_struct_type)
16386 (write_state_pointer_type): ditto, renaming to...
16387 (state_writer::write_state_pointer_type)
16388 (write_state_array_type): ditto, renaming to...
16389 (state_writer::write_state_array_type)
16390 (write_state_gc_used): ditto, renaming to...
16391 (state_writer::write_state_gc_used)
16392 (write_state_common_type_content): ditto, renaming to...
16393 (state_writer::write_state_common_type_content)
16394 (write_state_type): ditto, renaming to...
16395 (state_writer::write_state_type)
16396 (write_state_pair_list): ditto, renaming to...
16397 (state_writer::write_state_pair_list)
16398 (write_state_pair): ditto, renaming to...
16399 (state_writer::write_state_pair)
16400 (write_state_typedefs): ditto, renaming to...
16401 (state_writer::write_state_typedefs)
16402 (write_state_structures): ditto, renaming to...
16403 (state_writer::write_state_structures)
16404 (write_state_param_structs): ditto, renaming to...
16405 (state_writer::write_state_param_structs)
16406 (write_state_variables): ditto, renaming to...
16407 (state_writer::write_state_variables)
16408 (write_state_srcdir): ditto, renaming to...
16409 (state_writer::write_state_srcdir)
16410 (write_state_files_list): ditto, renaming to...
16411 (state_writer::write_state_files_list)
16412 (write_state_languages): ditto, renaming to...
16413 (state_writer::write_state_languages)
16414 (write_state): create a state_writer instance and use it when
16415 writing out the state file
16417 2013-05-17 Mike Stump <mikestump@comcast.net>
16419 PR rtl-optimization/57304
16420 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16421 accessing DF_REF_REAL_LOC.
16423 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16425 PR rtl-optimization/57281
16426 PR rtl-optimization/57300
16427 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16428 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16429 what the other splitter did if the registers are dead.
16431 2013-05-17 Richard Biener <rguenther@suse.de>
16433 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16436 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16438 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16441 2013-05-17 Marek Polacek <polacek@redhat.com>
16443 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16444 length when doing non-zero store of storing '\0' to '\0'.
16446 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16448 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16449 vect_external_def oprnd1 with loop_vinfo, try to emit
16450 optional cast, negation and and stmts on the loop preheader
16451 edge instead of into the pattern def seq.
16453 PR tree-optimization/57051
16454 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16455 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16457 2013-05-16 Nick Clifton <nickc@redhat.com>
16459 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16460 (rl78_is_naked_func): New function.
16461 (rl78_expand_prologue): Skip prologue generation for naked functions.
16462 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16463 * doc/extend.texi (naked): Add RL78 to the list of processors
16464 that supports this attribute.
16466 2013-05-16 Jeff Law <law@redhat.com>
16468 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16470 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16472 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16473 cache parameters using detect_caches_amd also for CYRIX,
16474 NSC and TM2 signatures.
16476 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16477 Dzianis Kahanovich <mahatma@eu.by>
16481 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16482 VIA/Centaur processors and determine their cache parameters
16483 using detect_caches_amd.
16485 2013-05-16 Teresa Johnson <tejohnson@google.com>
16487 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16488 (rtl_verify_edges): New function.
16489 (rtl_verify_bb_insns): Ditto.
16490 (rtl_verify_bb_pointers): Ditto.
16491 (rtl_verify_bb_insn_chain): Ditto.
16492 (rtl_verify_fallthru): Ditto.
16493 (rtl_verify_bb_layout): Ditto.
16494 (rtl_verify_flow_info_1): Outline checks into new functions.
16495 (rtl_verify_flow_info): Ditto.
16497 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16499 * cfghooks.c (copy_bbs): Add update_dominance argument.
16500 * cfghooks.h (copy_bbs): Update prototype.
16501 * tree-cfg.c (gimple_duplicate_sese_region):
16502 Add update_dominance argument.
16503 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16504 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16505 gimple_duplicate_sese_region call.
16506 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16507 Update copy_bbs call.
16508 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16509 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16511 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16513 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16514 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16515 vect_recog_rotate_pattern.
16516 (vect_recog_rotate_pattern): New function.
16518 2013-05-16 Jason Merrill <jason@redhat.com>
16520 * Makefile.in (LLINKER): New variable.
16521 (mostlyclean): Remove link mutex.
16522 * configure.ac: Handle --enable-link-mutex.
16523 * lock-and-run.sh: New script.
16525 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16528 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16531 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16533 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16535 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16537 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16538 * config/arm/arm.c (next_consecutive_mem): New function.
16539 (gen_movmem_ldrd_strd): Likewise.
16540 * config/arm/arm.md (movmemqi): Update condition and code.
16541 (unaligned_loaddi, unaligned_storedi): New patterns.
16543 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16545 * config.gcc: Obsolete *-*-solaris2.9*.
16546 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16548 2013-05-16 Richard Biener <rguenther@suse.de>
16550 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16551 earlier, after GRAPHITE transforms and IV canonicalization.
16553 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16555 * omp-low.c (extract_omp_for_data): For collapsed loops,
16556 if at least one of the loops is known at compile time to
16557 iterate zero times, set count to 0.
16558 (expand_omp_regimplify_p): New function.
16559 (expand_omp_for_generic): For collapsed loops, if at least
16560 one of the loops isn't known to iterate at least once,
16561 add runtime check with setting count to 0.
16562 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16563 For unsigned types if it isn't known at compile time that
16564 the loop will iterate at least once, add runtime check to bypass
16565 the whole loop if initial condition isn't true.
16567 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16569 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16571 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16573 PR middle-end/57286
16574 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16575 transformations to avoid an infinite loop.
16577 2013-05-16 Marek Polacek <polacek@redhat.com>
16579 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16581 2013-05-15 Leif Ekblad <leif@rdos.net>
16583 * config/i386/i386.c (ix86_decompose_address): Use
16584 DEFAULT_TLS_SEG_REG to access TLS segment register.
16585 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16586 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16587 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16589 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16592 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16593 sibling calls to functions that would normally be lazily bound,
16594 unless $gp is call-clobbered.
16596 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16598 * config/i386/i386.c (ix86_option_override_internal): Update
16599 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16600 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16601 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16602 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16604 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16606 2013-05-15 Andreas Schwab <schwab@suse.de>
16608 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16609 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16612 2013-05-15 Teresa Johnson <tejohnson@google.com>
16614 * loop-unroll.c (report_unroll_peel): Check decision before
16615 emitting unroll/peel message.
16617 2013-05-15 Teresa Johnson <tejohnson@google.com>
16619 * function.h (has_bb_partition): New rtl_data flag.
16620 (bb_reorder_complete): Ditto.
16621 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16622 instead of flag_reorder_blocks_and_partition.
16623 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16624 with some enhancements.
16625 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16626 * bb-reorder.c (connect_traces): Check for has_bb_partition
16627 instead of flag_reorder_blocks_and_partition.
16628 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16629 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16630 verify_hot_cold_block_grouping.
16631 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16633 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16636 * config/arm/predicates.md (call_insn_operand): New predicate.
16637 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16638 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16639 if insn is not a tail call.
16640 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16642 * config/arm/arm.h (enum reg_class): New caller save register class.
16643 (REG_CLASS_NAMES): Likewise.
16644 (REG_CLASS_CONTENTS): Likewise.
16645 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16648 2013-05-15 Richard Biener <rguenther@suse.de>
16650 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16651 of MSG_OPTIMIZED_LOCATIONS.
16652 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16653 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16655 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16656 of MSG_OPTIMIZED_LOCATIONS.
16657 (execute_vect_slp): Likewise.
16658 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16659 (vect_create_cond_for_alias_checks): Likewise.
16660 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16661 (vect_recog_widen_mult_pattern): Likewise.
16662 (vect_recog_widen_sum_pattern): Likewise.
16663 (vect_recog_over_widening_pattern): Likewise.
16664 (vect_recog_widen_shift_pattern): Likewise.
16665 (vect_recog_vector_vector_shift_pattern): Likewise.
16666 (vect_recog_divmod_pattern): Likewise.
16667 (vect_recog_mixed_size_cond_pattern): Likewise.
16668 (vect_recog_bool_pattern): Likewise.
16669 (vect_pattern_recog_1): Likewise.
16671 2013-05-15 Martin Jambor <mjambor@suse.cz>
16673 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16674 non-functions to builtin_unreachable.
16675 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16676 correct when new direct edges were discovered.
16678 2013-05-15 Martin Jambor <mjambor@suse.cz>
16680 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16681 header, print symbol order instead of node uid, print more information
16682 about indirect edge targets.
16683 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16685 (ipa_make_edge_direct_to_target): Likewise.
16686 (remove_described_reference): Likewise.
16687 (propagate_controlled_uses): Likewise.
16688 (ipa_print_node_params): Also print symbol order.
16689 (ipcp_transform_function): Print symbol order instead of node uids.
16690 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16691 (cgraph_get_create_real_symbol_node): Likewise.
16692 * ipa-cp.c (print_lattice): Likewise.
16693 (print_all_lattices): Likewise.
16694 (determine_versionability): Likewise.
16695 (initialize_node_lattices): Likewise.
16696 (estimate_local_effects): Likewise.
16697 (update_profiling_info): Likewise.
16698 (create_specialized_node): Likewise.
16699 (perhaps_add_new_callers): Likewise.
16700 (decide_about_value): Likewise.
16701 (decide_whether_version_node): Likewise.
16702 (identify_dead_nodes): Likewise.
16703 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16704 (dump_inline_summary): Likewise.
16705 (estimate_node_size_and_time): Likewise.
16706 (inline_analyze_function): Likewise.
16707 * ipa-inline.c (report_inline_failed_reason): Likewise.
16708 (want_early_inline_function_p): Likewise.
16709 (edge_badness): Likewise.
16710 (update_edge_key): Likewise.
16711 (inline_small_functions): Likewise. Add dumping of order to two other
16713 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16714 instead of node uids.
16715 (propagate_pure_const): Likewise.
16716 (propagate_pure_const): Likewise.
16717 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16718 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16719 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16721 * tree-pretty-print.c (dump_function_header): Likewise.
16722 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16723 Print symbol order instead of node uids.
16725 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16727 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16728 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16730 2013-05-15 Richard Biener <rguenther@suse.de>
16732 PR tree-optimization/57275
16733 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16734 return value for fail to do runtime alias checks for gather loads.
16736 2013-05-15 Jan Hubicka <jh@suse.cz>
16740 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16741 weakrefs are not external.
16742 (lto_symtab_merge_decls): Fix thinko when dealing with
16743 non-lto_symtab decls.
16744 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16745 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16746 * varpool.c (dump_varpool_node): Dump more flags.
16748 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16750 * config/i386/i386.c (processor_alias_table): Add instruction
16751 FSGSBASE for AMD bdver3 architecture.
16753 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16755 * tree.c (warn_deprecated_use): Print file:line using locus color.
16756 * diagnostic.c (diagnostic_report_current_module): Print file:line
16757 and file:line:column using locus color.
16759 2013-05-14 Mike Stump <mikestump@comcast.net>
16761 * gdbinit.in: Add __null.
16763 2013-05-14 Mike Stump <mikestump@comcast.net>
16765 * recog.h: Rename struct recog_data to Recog_data.
16766 * recog.c: Likewise.
16767 * reload.c (can_reload_into): Likewise.
16768 * config/picochip/picochip.c: Likewise.
16770 2013-05-14 Mike Stump <mikestump@comcast.net>
16772 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16774 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16776 * resource.h (struct resources): Remove unch_memory member.
16777 (CLEAR_RESOURCE): Don't clear unch_memory.
16778 * resource.c (mark_referenced_resources): Don't set it.
16779 (mark_set_resources): Likewise.
16780 (mark_target_live_regs): Don't clear it.
16781 (init_resource_info): Likewise.
16782 * reorg.c (resource_conflicts_p): Don't compare it.
16783 (redundant_insn): Don't set it.
16785 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16787 * emit-rtl.c (next_label): Remove unused function.
16788 (skip_consecutive_labels, link_cc0_insns): Move to ...
16789 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16790 only place where these functions are used, and make them static.
16792 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16794 * fold-const.c (fold_negate_expr): Handle vectors.
16795 (fold_truth_not_expr): Make it static.
16796 (fold_invert_truthvalue): New static function.
16797 (invert_truthvalue_loc): Handle vectors. Do not call
16798 fold_truth_not_expr directly.
16799 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16800 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16801 (fold_comparison): Handle vector constants.
16802 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16803 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16804 * tree.h (fold_truth_not_expr): Remove declaration.
16806 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16808 * config/aarch64/aarch64-simd.md
16809 (aarch64_vcond_internal<mode>): Rename to...
16810 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16811 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16812 float modes. Clarify all iterator modes.
16813 (vcond<mode><mode>): Use new name for vcond expanders.
16814 (vcond<v_cmp_result><mode>): Likewise.
16815 (vcondu<mode><mode>: Likewise.
16816 * config/aarch64/iterators.md (VDQF_COND): New.
16818 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16821 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16822 variable for the shift amount. Check that we shift by non-negative
16825 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16828 * config/arm/arm.h (EPILOGUE_USES): Only return true
16829 for LR_REGNUM after epilogue_completed.
16831 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
16833 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
16834 is error_mark_node.
16836 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16839 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16840 and Solaris 11+/x86 with gld.
16841 * configure: Regenerate.
16843 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16845 * expmed.c (expand_shift_1): Canonicalize rotates by
16846 constant bitsize / 2 to bitsize - 1.
16847 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
16848 case ROTATERT>: Likewise.
16851 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16853 * config/i386/i386.md (rotateinv): New code attr.
16854 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16855 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16856 roll $31, %eax, etc.
16858 2013-05-14 Richard Biener <rguenther@suse.de>
16860 PR middle-end/57235
16861 * tree-eh.c (sink_clobbers): Give up for successors with
16862 multiple predecessors and no virtual uses.
16864 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16866 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16867 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16869 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16871 PR middle-end/57251
16872 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16873 the case when both op0 and op1 have VOIDmode.
16875 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16877 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16878 in multiply-accumulate mode.
16880 2013-05-13 Guozhi Wei <carrot@google.com>
16882 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16884 2013-05-13 Kai Tietz <ktietz@redhat.com>
16887 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
16888 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
16889 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
16890 * config/i386/i386.c (ix86_option_override_internal): Likewise.
16891 (ix86_expand_prologue): Likewise.
16892 (ix86_expand_split_stack_prologue): Likewise.
16893 (legitimate_pic_address_disp_p): Likewise.
16894 (legitimize_pic_address): Likewise.
16895 (legitimize_tls_address): Likewise.
16896 (legitimize_pe_coff_symbol): Likewise.
16897 (output_pic_addr_const): Likewise.
16898 (construct_plt_address): Likewise.
16899 (ix86_expand_call): Likewise.
16900 (x86_output_mi_thunk): Likewise.
16901 (x86_function_profiler): Likewise.
16903 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16905 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16906 similar switch cases.
16907 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16908 (aarch64_simd_mov_to_<mode>low): Delete.
16909 (aarch64_simd_mov_to_<mode>high): Delete.
16910 (move_lo_quad_<mode>): Add w<-r alternative.
16911 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16912 (aarch64_simd_mov_from_*): Update type attribute.
16913 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16916 2013-05-13 Jan Hubicka <jh@suse.cz>
16918 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16919 * config/i386/i386.c (ix86_compute_frame_layout,
16920 ix86_expand_epilogue, emit_i387_cw_initialization,
16921 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16922 ix86_local_alignment): Fix use of size/speed predicates.
16924 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16926 PR tree-optimization/45216
16927 PR tree-optimization/57157
16928 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16929 the (-Y) & (B - 1) variant if OP is |.
16930 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16931 return shifted. Use (-op1) & (prec - 1) as other_amount
16932 instead of prec - op1.
16934 2013-05-13 Martin Jambor <mjambor@suse.cz>
16936 PR middle-end/42371
16937 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
16938 (ipa_constant_data): New type.
16939 (ipa_jump_func): Use ipa_constant_data to hold information about
16940 constant jump functions.
16941 (ipa_get_jf_constant): Adjust to jump function type changes.
16942 (ipa_get_jf_constant_rdesc): New function.
16943 (ipa_param_descriptor): New field controlled_uses.
16944 (ipa_get_controlled_uses): New function.
16945 (ipa_set_controlled_uses): Likewise.
16946 * ipa-ref.h (ipa_find_reference): Declare.
16947 * ipa-prop.c (ipa_cst_ref_desc): New type.
16948 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
16950 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
16951 New parameter cs. Adjust all callers.
16952 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
16953 (remove_described_reference): New function.
16954 (jfunc_rdesc_usable): Likewise.
16955 (try_make_edge_direct_simple_call): Decrement controlled use count,
16956 attempt to remove reference if it hits zero.
16957 (combine_controlled_uses_counters): New function.
16958 (propagate_controlled_uses): Likewise.
16959 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
16960 (ipa_edge_duplication_hook): Duplicate reference descriptions.
16961 (ipa_print_node_params): Print described use counter.
16962 (ipa_write_jump_function): Adjust to jump function type changes.
16963 (ipa_read_jump_function): New parameter CS, pass it to
16964 ipa_set_jf_constant. Adjust caller.
16965 (ipa_write_node_info): Stream controlled use count
16966 (ipa_read_node_info): Likewise.
16967 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
16969 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
16970 count. Remove cloning-added reference if it reaches zero.
16971 * ipa-ref.c (ipa_find_reference): New function.
16973 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16975 * config/i386/i386.c (processor_target_table): Modified default
16976 alignment values for AMD BD and BT architectures.
16978 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
16980 * tree-vect-generic.c (uniform_vector_p): Move ...
16981 * tree.c (uniform_vector_p): ... here.
16982 * tree.h (uniform_vector_p): Declare it.
16983 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
16986 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16988 PR tree-optimization/57230
16989 * tree-ssa-strlen.c (handle_char_store): Record length for
16990 array store from STRING_CST.
16992 PR tree-optimization/57230
16993 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
16996 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
16998 * config/epiphany/epiphany.c (epiphany_init): Check size of
16999 NUM_MODES_FOR_MODE_SWITCHING.
17000 (epiphany_expand_prologue):
17001 Remove CONFIG_REGNUM initial value handling code.
17002 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17003 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17004 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17005 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17006 Don't return 1 for FP_MODE_NONE.
17007 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17008 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17009 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17010 * config/epiphany/epiphany.md (save_config): New pattern.
17012 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17014 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17016 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17018 * config/i386/i386.md (memory): Handle sseishft1.
17019 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17020 (*vec_extractv2di_1): Ditto.
17022 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17024 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17027 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17029 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17030 Add mthumb/march=armv7-a multilib.
17031 Add mthumb/march=armv7-r multilib.
17032 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17034 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17036 * config/v850/t-rtems: Add more multilibs.
17038 2013-05-10 Richard Biener <rguenther@suse.de>
17040 PR tree-optimization/57214
17041 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17042 not propagate from SSA names that occur in abnormal PHI nodes.
17044 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17046 * stor-layout.c (element_precision): New function.
17047 * machmode.h (element_precision): Declare it.
17048 * tree.c (build_minus_one_cst): New function.
17049 (element_precision): Likewise.
17050 * tree.h (build_minus_one_cst): Declare new function.
17051 (element_precision): Likewise.
17052 * fold-const.c (operand_equal_p): Use element_precision.
17053 (fold_binary_loc): Handle vector types.
17054 * convert.c (convert_to_integer): Use element_precision.
17055 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17058 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17060 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17061 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17062 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17063 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17064 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17065 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17066 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17067 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17068 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17069 (Uuw8): New constraints.
17070 (Usb4): Move into alphabetical order.
17071 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17072 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17073 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17074 previously unnamed patterns.
17075 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17076 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17077 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17078 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17079 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17080 of specific lengths.
17082 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17084 * config/i386/i386.md (rotateinv): New code attr.
17085 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17086 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17087 roll $31, %eax, etc.
17089 PR tree-optimization/45216
17090 PR tree-optimization/57157
17091 * tree-ssa-forwprop.c (simplify_rotate): New function.
17092 (ssa_forward_propagate_and_combine): Call it.
17094 2013-05-10 Richard Biener <rguenther@suse.de>
17096 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17097 disable peeling when we version for aliasing.
17098 (vector_alignment_reachable_p): Honor explicit user alignment.
17099 (vect_supportable_dr_alignment): Likewise.
17100 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17101 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17102 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17103 then peeling to arrange for the cost-model check to come first.
17105 2013-05-10 Alan Modra <amodra@gmail.com>
17107 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17108 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17109 * configure: Regenerate.
17111 2013-05-10 Alan Modra <amodra@gmail.com>
17114 * varasm.c (default_elf_select_section): Move !DECL_P check..
17115 (get_named_section): ..to here before calling get_section_name.
17117 (default_section_type_flags): Add DECL_P check.
17118 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17119 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17121 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17123 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17124 When using gen_stack_adjust_str with a register offset, add a
17125 REG_FRAME_RELATED_EXPR note.
17127 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17129 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17130 (*vec_extractv4si_zext_mem): Ditto.
17131 (*vec_extractv2di): Add 0->x and x->x alternatives.
17132 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17133 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17135 2013-05-09 Jason Merrill <jason@redhat.com>
17137 N3639 C++1y VLA support
17138 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17141 * tree.c (build_constructor_va): New.
17142 * tree.h: Declare it.
17144 2013-05-09 Martin Jambor <mjambor@suse.cz>
17147 * gimple-fold.c (canonicalize_constructor_val): Call
17148 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17150 2013-05-09 Jan Hubicka <jh@suse.cz>
17151 Richard Biener <rguenther@suse.de>
17154 * symtab.c (symtab_make_decl_local): Do not add private names.
17156 2013-05-09 Jan Hubicka <jh@suse.cz>
17159 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17160 (unlink_from_assembler_name_hash): Likewise.
17161 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17162 symtab_unregister_node, symtab_initialize_asm_name_hash,
17163 change_decl_assembler_name): Update.
17165 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17167 * config/aarch64/aarch64.md: New movtf split.
17168 (*movtf_aarch64): Update.
17169 (aarch64_movdi_tilow): Handle TF modes and rename to
17170 aarch64_movdi_<mode>low.
17171 (aarch64_movdi_tihigh): Handle TF modes and rename to
17172 aarch64_movdi_<mode>high
17173 (aarch64_movtihigh_di): Handle TF modes and rename to
17174 aarch64_mov<mode>high_di
17175 (aarch64_movtilow_di): Handle TF modes and rename to
17176 aarch64_mov<mode>low_di
17177 (aarch64_movtilow_tilow): Remove spurious whitespace.
17178 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17180 (aarch64_print_operand): Update.
17182 2013-05-09 Alan Modra <amodra@gmail.com>
17184 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17186 * configure: Regenerate.
17188 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17190 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17191 splitter preparation statements.
17192 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17193 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17196 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17198 * gimple-ssa-strength-reduction.c (count_candidates): Change
17199 return value to int.
17200 (analyze_candidates_and_replace): Change type of length to int.
17202 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17204 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17205 (*vec_extract<mode>): Use VI12_128 mode iterator.
17206 (*vec_extract<mode>_mem): Ditto.
17207 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17210 2013-05-08 Diego Novillo <dnovillo@google.com>
17215 2012-08-17 Diego Novillo <dnovillo@google.com>
17218 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17219 * config.in: Regenerate.
17220 * configure: Regenerate.
17221 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17223 2013-05-08 Jan Hubicka <jh@suse.cz>
17226 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17227 * cgraph.h (symtab_node_base): Add unique_name.
17228 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17229 input_overwrite_node, input_varpool_node): Stream unique_name.
17230 * cgraphclones.c (cgraph_create_virtual_clone,
17231 cgraph_function_versioning): Set unique_name.
17232 * ipa.c (function_and_variable_visibility): Set unique_name.
17234 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17236 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17237 (alloc_cand_and_find_basis): Restrict conditional candidate
17238 processing to CAND_MULTs.
17240 2013-05-08 Jan Hubicka <jh@suse.cz>
17243 lto-symtab.c (lto_symtab_symbol_p): New function.
17244 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17245 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17246 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17247 Skip static symbols.
17249 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17251 PR tree-optimization/57200
17252 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17253 Only call inform if the preceding warning_at returns true.
17255 2013-05-07 Han Shen <shenhan@google.com>
17257 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17258 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17259 * common.opt (fstack-protector-strong): New option.
17260 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17261 * doc/invoke.texi (Optimization Options): Document
17262 "-fstack-protector-strong".
17263 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17265 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17267 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17269 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17271 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17272 it is not lhs of the insn.
17273 (setup_reg_equiv): Remove insn having equiv memory which it is not
17275 * lra-constraints.c (process_address): Try to improve generation
17276 code for address base + disp.
17277 (lra_constraints): Make correct the code for checking insn setting
17278 up backward equivalence. Remove insn only if it is in the init
17280 * lra-eliminations.c (update_reg_eliminate): Change return value.
17281 (lra_eliminate): Use the result.
17283 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17285 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17286 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17287 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17288 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17289 register target operands.
17290 (*vec_extractv8hi_sse2): New pattern.
17291 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17292 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17293 (*vec_extract<mode>_mem): New insn and split pattern.
17295 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17297 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17298 (TARGET_ASAN_SHADOW_OFFSET): Define.
17299 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17300 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17302 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17304 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17305 (incr_vec_index): Return -1 if increment not found.
17306 (create_add_on_incoming_edge): Assert if increment not found.
17307 (record_increment): Limit number of increments recorded.
17308 (all_phi_incrs_profitable): Return false if an increment not found.
17309 (replace_profitable_candidates): Don't process increments that were
17311 (analyze_candidates_and_replace): Limit size of incr_vec.
17313 2013-05-07 Richard Biener <rguenther@suse.de>
17315 * calls.c (special_function_p): setjmp-like functions are leaf.
17316 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17317 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17319 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17321 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17322 (aarch64_simd_mov<mode>): New expander.
17323 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17324 (aarch64_simd_mov_to_<mode>high): Likewise.
17325 (aarch64_simd_mov_from_<mode>low): Likewise.
17326 (aarch64_simd_mov_from_<mode>high): Likewise.
17327 (aarch64_dup_lane<mode>): Update.
17328 (aarch64_dup_lanedi): New instruction pattern.
17329 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17330 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17332 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17334 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17335 (replace_mult_candidate): Remove unnecessary argument; remove
17336 unnecessary parameter from call to introduce_cast_before_cand.
17337 (replace_unconditional_candidate): Remove unnecessary parameter
17338 from call to replace_mult_candidate.
17339 (replace_conditional_candidate): Likewise.
17340 (insert_initializers): Use make_temp_ssa_name.
17341 (introduce_cast_before_cand): Remove unnecessary argument; use
17342 make_temp_ssa_name.
17343 (replace_one_candidate): Remove unnecessary argument; remove
17344 unnecessary parameter from calls to introduce_cast_before_cand.
17345 (replace_profitable_candidates): Remove unnecessary parameters
17346 from calls to replace_one_candidate.
17348 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17350 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17351 phi def as possibly hiding a basis for a CAND_ADD whose operands
17352 have been commuted in the analysis.
17353 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17355 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17357 * config/aarch64/aarch64.md
17358 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17359 shift value between 0-4.
17361 2013-05-07 Richard Biener <rguenther@suse.de>
17363 * double-int.h (rshift): New overload.
17364 * double-int.c (rshift): New function.
17365 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17366 (create_reference_ops_from_ref): Remove.
17367 (vn_reference_insert): Use shared ops for constructing the
17368 reference and copy it.
17370 2013-05-07 Richard Biener <rguenther@suse.de>
17372 PR middle-end/57190
17373 * tree-eh.c (sink_clobbers): Properly propagate
17374 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17376 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17378 PR tree-optimization/57149
17379 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17380 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17381 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17382 uninit_undefined_value_p instead of ssa_undefined_value_p.
17385 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17386 for modifier == EXPAND_INITIALIZER.
17388 2013-05-07 Anton Blanchard <anton@samba.org>
17390 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17391 for powerpc64 little endian.
17392 * configure: Regenerate.
17394 2013-05-06 Graham Stott <grahams@btinternet.com>
17396 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17397 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17398 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17399 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17401 2013-05-06 Graham Stott <grahams@btinternet.com>
17403 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17404 codes which allow non-lvalues.
17406 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17408 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17409 components are all 1s.
17410 (integer_minus_onep): New function.
17411 * tree.h (integer_minus_onep): Declare it.
17412 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17413 integer_minus_onep instead of integer_all_onesp.
17415 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17418 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17419 variations of these patterns.
17421 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17423 * config/i386/i386.md (isa): Add x64_sse4 member.
17424 (enabled): Handle x64_sse4.
17425 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17426 instruction for 64bit SSE4_1 targets. Update insn attributes.
17427 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17428 instruction for SSE4_1 targets. Update insn attributes.
17429 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17430 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17431 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17433 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17434 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17435 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17437 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17440 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17442 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17444 * final.c (do_assembler_dialects): Don't handle curly braces and
17445 vertical bar escaped by % as dialect delimiters.
17446 (output_asm_insn): Print curly braces and vertical bar if escaped
17447 by % and ASSEMBLER_DIALECT defined.
17448 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17449 * doc/tm.texi: Regenerated.
17451 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17453 * config/mips/mips.c: Include tree-pass.h.
17454 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17455 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17456 (pass_mips_machine_reorg2): New machine specific pass.
17457 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17458 (mips_option_override): Register the new pass.
17459 * rtl.h (cleanup_barriers): Remove prototype.
17460 (dbr_schedule): Likewise.
17461 * jump.c (cleanup_barriers): Make static.
17462 * reorg.c (dbr_schedule): Likewise.
17464 2013-05-06 Richard Biener <rguenther@suse.de>
17466 PR tree-optimization/57185
17467 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17468 SSA names properly.
17470 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17473 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17474 between "first_arg" and "insn", not "last" and "insn".
17476 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17478 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17479 (find_candidates_in_block): Re-enable slsr_process_phi.
17480 (create_phi_basis): Fix double counting of candidate adjustment.
17482 2013-05-06 Richard Biener <rguenther@suse.de>
17484 PR middle-end/57147
17485 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17486 the edge is also fallthru, preserve it and just clear the
17488 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17489 also complex, preserve that and just clear the fallthru flag.
17490 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17491 update virtual operands.
17493 2013-05-06 Alan Modra <amodra@gmail.com>
17495 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17496 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17497 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17498 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17499 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17500 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17502 2013-05-06 Alan Modra <amodra@gmail.com>
17504 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17505 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17506 (DEFAULT_ASM_ENDIAN): Define.
17507 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17508 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17509 Update -K PIC clause from sysv4.h.
17510 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17511 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17513 2013-05-06 Alan Modra <amodra@gmail.com>
17515 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17516 twice for little-endian.
17517 (ashrdi3_no_power, ashrdi3): Support little-endian.
17519 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17522 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17523 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17525 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17526 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17527 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17530 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17531 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17533 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17534 * config/pa/pa.opt: Make mbig-switch a no-op.
17535 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17536 (CASE_VECTOR_MODE): Always return SImode.
17537 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17538 for the !TARGET_BIG_SWITCH case.
17539 * config/pa/pa-linux.h: Likewise.
17540 * config/pa/pa-openbsd.h: Likewise.
17541 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17542 * config/pa/pa.md (short_jump): Remove define_insn.
17543 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17544 (casesi0): Remove define_insn.
17545 (type): Remove btable_branch.
17546 (pa_combine_type): Likewise.
17547 (in_nullified_branch_delay): Likewise.
17548 (in_call_delay): Likewise.
17549 (define_delay): Likewise.
17550 (define_insn_reservation "Z3"): Likewise.
17551 (define_insn_reservation "Z4"): Likewise.
17552 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17553 (pa_adjust_insn_length): Remove adjustment for btable branches.
17554 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17557 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17559 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17560 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17561 Add m->r,x alternatives.
17562 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17563 splitters using SWI48x mode iterator.
17564 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17565 TARGET_64BIT. Add m->x alternative.
17566 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17567 Add o->x alternative. Enable for TARGET_SSE.
17568 (sse_storeq): Remove expander.
17569 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17570 with memory input operand.
17571 (*vec_extractv2di_1 splitter): New.
17572 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17573 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17575 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17577 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17578 (INT_LOWPART): Delete.
17579 (extract_MB): Adjust.
17580 (extract_ME): Adjust.
17581 (print_operand): Adjust.
17583 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17585 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17586 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17587 (reg_or_logical_cint_operand, easy_fp_constant,
17588 logical_const_operand): Delete "CONST_DOUBLE" case.
17589 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17590 "HOST_BITS_PER_WIDE_INT == 64" test.
17591 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17592 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17593 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17594 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17596 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17597 CONST_DOUBLE DImode/VOIDmode case.
17598 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17599 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17600 CONST_DOUBLE VOIDmode case.
17601 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17602 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17603 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17604 Delete CONST_DOUBLE case.
17605 (splitters for mov FMOVE64 const_double): Delete
17606 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17607 "HOST_BITS_PER_WIDE_INT >= 64" test.
17608 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17610 (mov DI const_double): Delete.
17612 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17614 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17615 on op shows all bits zero in mode of a lowpart subreg, return zero.
17617 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17620 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17621 to save TFmode registers and DImode to save TImode registers for
17622 caller save operations.
17623 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17624 mark being partially clobbered since they only use the first
17627 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17628 and TDmode only use the upper 64-bits of each VSX register.
17630 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17632 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17633 (find_candidates_in_block): Disable slsr_process_phi.
17635 2013-05-03 Guozhi Wei <carrot@google.com>
17637 * coverage.c (coverage_obj_init): Move the construction of gcov
17639 (build_init_ctor): ... here.
17641 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17643 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17644 (slsr_cand_d): Redefine def_phi.
17645 (stride_status, phi_adjust_status, count_phis_status): New enums.
17646 (find_phi_def): New.
17647 (find_basis_for_base_expr): New.
17648 (find_basis_for_candidate): Handle hidden bases.
17649 (alloc_cand_and_find_basis): Handle phi candidates.
17650 (slsr_process_phi): New.
17651 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17652 (create_mul_imm_cand): Likewise.
17653 (create_add_ssa_cand): Exclude phi base candidates.
17654 (create_add_imm_cand): Likewise.
17655 (slsr_process_cast): Likewise.
17656 (slsr_process_copy): Likewise.
17657 (find_candidates_in_block): Handle phi candidates.
17658 (dump_candidate): Likewise.
17659 (unconditional_cands): Delete.
17660 (unconditional_cands_with_known_stride_p): Delete.
17661 (phi_dependent_cand_p): New.
17662 (cand_increment): Handle phi-dependent candidates.
17663 (replace_dependent): Delete.
17664 (replace_mult_candidate): New.
17665 (replace_unconditional_candidate): New.
17666 (incr_vec_index): Move to avoid forward reference.
17667 (create_add_on_incoming_edge): New.
17668 (create_phi_basis): New.
17669 (replace_dependents): Delete.
17670 (replace_conditional_candidate): New.
17671 (phi_add_costs): New.
17672 (replace_uncond_cands_and_profitable_phis): New.
17673 (record_increment): Handle phi adjustments.
17674 (record_phi_increments): New.
17675 (record_increments): Handle phi adjustments.
17676 (phi_incr_cost): New.
17677 (lowest_cost_path): Handle phis.
17678 (total_savings): Likewise.
17679 (analyze_increments): Likewise.
17680 (ncd_with_phi): New.
17681 (ncd_of_cand_and_phis): New.
17682 (nearest_common_dominator_for_cands): Handle phi increments.
17683 (all_phi_incrs_profitable): New.
17684 (replace_profitable_candidates): Handle phi-dependent candidates.
17685 (analyze_candidates_and_replace): Likewise.
17687 2013-05-03 Teresa Johnson <tejohnson@google.com>
17690 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17691 do not exceed REG_BR_PROB_BASE.
17693 2013-05-03 Jeff Law <law@redhat.com>
17695 PR tree-optimization/57144
17696 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17697 operand of the condition will bit into the new type when eliminating
17698 a cast feeding a condition.
17700 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17702 PR rtl-optimization/57130
17703 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17704 of COMPARE as in_code to the recursive call if needed.
17706 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17708 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17709 (enabled): Handle new members.
17710 * config/i386/sse.md (*vec_concatv2si): Merge from
17711 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17712 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17714 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17716 PR tree-optimization/57027
17717 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17718 for fnms opportunity, check we got the prerequisite kind
17719 of tree / gimple before using accessor functions.
17721 2013-05-03 Richard Biener <rguenther@suse.de>
17723 * double-int.h (lshift): New overload without precision
17724 and arith argument.
17725 (operator *=, operator +=, operator -=): Move ...
17726 * double-int.c (operator *=, operator +=, operator -=): ... here
17727 and implement more efficiently.
17728 (mul_double_with_sign): Remove.
17729 (lshift_double): Adjust to take unsinged shift argument, push
17730 dispatching code to callers.
17731 (mul_double_wide_with_sign): Add early out for callers that
17732 are not interested in high parts or overflow.
17733 (lshift): New function.
17734 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17735 dispatch code here.
17736 (lrotate, rrotate): Use logical shifts.
17737 * expr.c (get_inner_reference): Use lshift.
17738 * fixed-value.c (do_fixed_divide): Likewise.
17739 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17740 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17741 (indirect_refs_may_alias_p): Likewise.
17742 (stmt_kills_ref_p_1): Likewise.
17744 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17746 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17748 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17750 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17751 scalar form of FABD instruction.
17753 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17755 * lra-constraints.c (process_alt_operands): Add checking alt
17756 number to choose the best alternative.
17758 2013-05-02 Richard Biener <rguenther@suse.de>
17760 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17761 bitmap and its handling.
17762 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17764 2013-05-02 Richard Biener <rguenther@suse.de>
17766 PR middle-end/57140
17767 * tree-inline.c (copy_loops): Properly handle removed loops.
17768 (copy_cfg_body): Mark destination loops for fixup if source
17769 loops needed fixup.
17771 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17774 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17775 generating simple_return for naked functions.
17777 2013-05-02 Martin Jambor <mjambor@suse.cz>
17779 PR middle-end/56988
17780 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17781 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17783 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17784 ipa_agg_replacement_value structures.
17785 (known_aggs_to_agg_replacement_list): Likewise.
17786 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17787 (read_agg_replacement_chain): Likewise.
17788 (ipcp_transform_function): Also check that by_ref flags match.
17790 2013-05-02 Richard Biener <rguenther@suse.de>
17792 * graphds.h (struct graph): Add obstack member.
17793 * graphds.c (new_graph): Initialize obstack and allocate
17795 (add_edge): Allocate edge from the obstack.
17796 (free_graph): Free the obstack instead of all edges and vertices.
17798 2013-05-02 Teresa Johnson <tejohnson@google.com>
17800 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17802 * cfg.c (update_bb_profile_for_threading): Ditto.
17803 * tree-inline.c (copy_bb): Ditto.
17804 (copy_edges_for_bb): Ditto.
17805 (initialize_cfun): Ditto.
17806 (copy_cfg_body): Ditto.
17807 (expand_call_inline): Ditto.
17808 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17809 (estimate_node_size_and_time): Ditto.
17810 (inline_merge_summary): Ditto.
17811 * cgraphclones.c (cgraph_clone_edge): Ditto.
17812 (cgraph_clone_node): Ditto.
17813 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17814 (compute_trg_info): Ditto.
17816 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17818 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17819 S reg when fp attribute set.
17820 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17822 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17824 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17826 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17827 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17828 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17830 2013-05-02 Richard Biener <rguenther@suse.de>
17832 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17833 (struct instantiate_cache_entry): New type.
17834 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17835 (struct instantiate_cache_type): New type.
17836 (set_instantiated_value, get_instantiated_value): Remove.
17837 (get_instantiated_value_entry): New function.
17838 (instantiate_scev_name): Use the new cache and adjust.
17839 (instantiate_scev_poly): Adjust.
17840 (instantiate_scev_binary): Likewise.
17841 (instantiate_array_ref): Likewise.
17842 (instantiate_scev_convert): Likewise.
17843 (instantiate_scev_not): Likewise.
17844 (instantiate_scev_3): Likewise.
17845 (instantiate_scev_2): Likewise.
17846 (instantiate_scev_r): Likewise.
17847 (instantiate_scev): Likewise.
17848 (resolve_mixers): Likewise.
17850 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17853 * lra-constraints.c (best_small_class_operands_num): Remove.
17854 (process_alt_operands): Remove small_class_operands_num. Take
17855 small classes operands into losers and only if the operand is not
17856 matched. Modify debugging output.
17857 (curr_insn_transform): Remove best_small_class_operands_num.
17860 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17862 * config/aarch64/aarch64-builtins.c
17863 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17864 * config/aarch64/aarch64-simd-builtins.def
17865 (reduc_splus_): Add new modes.
17866 (reduc_uplus_): New.
17867 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17868 (reduc_uplus_v4sf): Likewise.
17869 (reduc_splus_v4sf): Likewise.
17870 (aarch64_addv<mode>): Likewise.
17871 (reduc_uplus_<mode>): Likewise.
17872 (reduc_splus_<mode>): Likewise.
17873 (aarch64_addvv2di): Likewise.
17874 (reduc_uplus_v2di): Likewise.
17875 (reduc_splus_v2di): Likewise.
17876 (aarch64_addvv2si): Likewise.
17877 (reduc_uplus_v2si): Likewise.
17878 (reduc_splus_v2si): Likewise.
17879 (reduc_<sur>plus_<mode>): New.
17880 (reduc_<sur>plus_v2di): Likewise.
17881 (reduc_<sur>plus_v2si): Likewise.
17882 (reduc_<sur>plus_v4sf): Likewise.
17883 (aarch64_addpv4sf): Likewise.
17884 * config/aarch64/arm_neon.h
17885 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17886 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17887 add UNSPEC_SADDV, UNSPEC_UADDV.
17889 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17891 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17893 * config/aarch64/arm_neon.h
17894 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17896 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17898 * config/aarch64/aarch64-builtins
17899 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17901 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17903 * config/aarch64/aarch64-simd-builtins.def
17904 (reduc_smax_): New.
17905 (reduc_smin_): Likewise.
17906 (reduc_umax_): Likewise.
17907 (reduc_umin_): Likewise.
17908 (reduc_smax_nan_): Likewise.
17909 (reduc_smin_nan_): Likewise.
17912 (smax): Update for V2SF, V4SF and V2DF modes.
17915 (smin_nan): Likewise.
17916 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17917 (<su><maxmin><mode>3): ...This, refactor.
17918 (s<maxmin><mode>3): New.
17919 (<maxmin_uns><mode>3): Likewise.
17920 (reduc_<maxmin_uns>_<mode>): Refactor.
17921 (reduc_<maxmin_uns>_v4sf): Likewise.
17922 (reduc_<maxmin_uns>_v2si): Likewise.
17923 (aarch64_<fmaxmin><mode>: Remove.
17924 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17926 (vmin<q>_f<32,64>): Likewise.
17927 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17929 (su): Add mappings for smax, smin, umax, umin.
17931 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17932 (FMAXMIN): Rename as...
17933 (FMAXMIN_UNS): ...This.
17935 (fmaxminv): Likewise.
17936 (fmaxmin): Likewise.
17938 (maxmin_uns_op): Likewise.
17940 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17942 * config/aarch64/arm_neon.h
17943 (vac<ge, gt><sd>_f<32, 64>): Rename to...
17944 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
17945 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
17947 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17949 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
17950 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
17952 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17954 * config/aarch64/aarch64-simd.md
17955 (vcond<mode>_internal): Handle special cases for constant masks.
17956 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
17957 (vcondu<mode><mode>): Likewise.
17958 (vcond<v_cmp_result><mode>): New.
17960 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17962 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
17963 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
17964 * config/aarch64/aarch64-simd-builtins.def
17965 (cmeq): Update to BUILTIN_VALLDI.
17970 * config/aarch64/arm_neon.h
17971 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
17972 to builtins or C as appropriate.
17974 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17976 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
17978 (cmhi): Rename to...
17980 * config/aarch64/aarch64-simd.md
17981 (simd_mode): Add SF.
17982 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
17983 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
17984 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
17985 (cstore<mode>_neg): ...This.
17986 * config/aarch64/iterators.md
17988 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
17989 (COMPARISONS): New.
17990 (UCOMPARISONS): Likewise.
17991 (optab): Add missing comparisons.
17997 (VCMP_S): Likewise.
17998 (VCMP_U): Likewise.
17999 (V_cmp_result): Add DF, SF modes.
18000 (v_cmp_result): Likewise.
18002 (vmtype): Likewise.
18003 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18005 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18007 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18008 define_insn to define_insn_and_split.
18009 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18010 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18011 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18012 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18013 (thumb2_negscc): Likewise.
18015 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18017 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18019 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18021 * config/arm/thumb2.md: Remove trailing whitespaces.
18023 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18025 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18026 Use gen_int_mode rather than GEN_INT.
18028 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18030 * value-prof.c (stream_in_histogram_value): Remove the strayed
18033 2013-04-30 Richard Biener <rguenther@suse.de>
18035 PR middle-end/57122
18036 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18038 2013-04-30 Richard Biener <rguenther@suse.de>
18040 PR middle-end/57107
18041 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18043 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18045 PR rtl-optimization/56957
18046 PR rtl-optimization/57105
18047 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18048 variable. Use just INSN_UID for determining whether an insn
18049 should be only disconnected from the insn stream.
18050 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18052 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18054 PR tree-optimization/57104
18055 * tsan.c (instrument_expr): Don't instrument accesses to
18056 DECL_HARD_REGISTER VAR_DECLs.
18058 2013-04-30 Richard Biener <rguenther@suse.de>
18060 * function.h (loops_for_fn): New inline function.
18061 (set_loops_for_fn): Likewise.
18062 * cfgloop.h (place_new_loop): Add struct function parameter.
18063 (get_loop): Likewise.
18064 (get_loops): Likewise.
18065 (number_of_loops): Likewise.
18066 (fel_next): Adjust.
18067 (fel_init): Likewise.
18068 * cfg.c (get_loop_copy): Adjust.
18069 * cfgloop.c (flow_loops_dump): Likewise.
18070 (record_loop_exits): Likewise.
18071 (verify_loop_structure): Likewise.
18072 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18073 (estimate_reg_pressure_cost): Likewise.
18074 (mark_loop_exit_edges): Likewise.
18075 * cfgloopmanip.c (place_new_loop): Likewise.
18076 (add_loop): Likewise.
18077 (duplicate_loop): Likewise.
18078 * graph.c (draw_cfg_nodes): Likewise.
18079 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18080 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18081 (extract_affine_chrec): Likewise.
18082 (build_scop_iteration_domain): Likewise.
18083 * graphite.c (graphite_initialize): Likewise.
18084 * ira-build.c (create_loop_tree_nodes): Likewise.
18085 (more_one_region_p): Likewise.
18086 (rebuild_regno_allocno_maps): Likewise.
18087 (mark_loops_for_removal): Likewise.
18088 (mark_all_loops_for_removal): Likewise.
18089 (remove_unnecessary_regions): Likewise.
18090 (ira_build): Likewise.
18091 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18092 * loop-init.c (fix_loop_structure): Likewise.
18093 (gate_rtl_move_loop_invariants): Likewise.
18094 (gate_rtl_unswitch): Likewise.
18095 (gate_rtl_unroll_and_peel_loops): Likewise.
18096 (rtl_doloop): Likewise.
18097 * lto-streamer-in.c (input_cfg): Likewise.
18098 * lto-streamer-out.c (output_cfg): Likewise.
18099 * modulo-sched.c (sms_schedule): Likewise.
18100 * predict.c (tree_estimate_probability): Likewise.
18101 (tree_estimate_probability_driver): Likewise.
18102 (estimate_loops): Likewise.
18103 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18104 (move_sese_region_to_fn): Likewise.
18105 (debug_loop_num): Likewise.
18106 * tree-chrec.c (chrec_evaluate): Likewise.
18107 (hide_evolution_in_other_loops_than_loop): Likewise.
18108 (chrec_component_in_loop_num): Likewise.
18109 (reset_evolution_in_loop): Likewise.
18110 (evolution_function_is_invariant_rec_p): Likewise.
18111 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18112 * tree-inline.c (copy_loops): Likewise.
18113 (copy_cfg_body): Likewise.
18114 (tree_function_versioning): Likewise.
18115 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18116 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18118 (add_to_evolution_1): Likewise.
18119 (scev_const_prop): Likewise.
18120 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18121 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18122 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18123 (tree_ssa_lim_initialize): Likewise.
18124 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18125 (verify_loop_closed_ssa): Likewise.
18126 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18127 (tree_ssa_loop_im): Likewise.
18128 (tree_ssa_loop_unswitch): Likewise.
18129 (tree_vectorize): Likewise.
18130 (check_data_deps): Likewise.
18131 (tree_ssa_loop_ivcanon): Likewise.
18132 (tree_ssa_loop_bounds): Likewise.
18133 (tree_complete_unroll): Likewise.
18134 (tree_complete_unroll_inner): Likewise.
18135 (tree_parallelize_loops): Likewise.
18136 (tree_ssa_loop_prefetch): Likewise.
18137 (tree_ssa_loop_ivopts): Likewise.
18138 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18139 * tree-vectorizer.c (vectorize_loops): Likewise.
18141 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18143 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18144 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18145 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18146 with EABI_LINK_SPEC.
18148 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18151 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18154 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18157 * lra-constraints.c (process_alt_operands): Discourage a bit more
18158 using memory for pseudos. Print cost dump for alternatives.
18159 Modify cost values for conflicts with early clobbers.
18160 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18162 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18165 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18167 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18169 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18170 from/to S register.
18171 (movdi_aarch64): Support LDR/STR from/to D register.
18173 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18175 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18176 or higher by default.
18178 2013-04-29 Richard Biener <rguenther@suse.de>
18180 PR middle-end/57075
18181 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18182 even if not adding abnormal edges for calls that can make
18185 2013-04-29 Richard Biener <rguenther@suse.de>
18187 PR middle-end/57103
18188 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18190 (move_stmt_r): Remove redundant checking.
18192 2013-04-29 Teresa Johnson <tejohnson@google.com>
18195 * basic-block.h (apply_scale): New function.
18196 (apply_probability): Use apply_scale.
18197 * gimple-streamer-in.c (input_bb): Ditto.
18198 * lto-streamer-in.c (input_cfg): Ditto.
18199 * lto-cgraph.c (merge_profile_summaries): Ditto.
18200 * tree-optimize.c (execute_fixup_cfg): Ditto.
18201 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18202 (copy_edges_for_bb): Ditto.
18203 (copy_cfg_body): Ditto.
18205 2013-04-29 Tom de Vries <tom@codesourcery.com>
18207 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18208 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18209 (tail_merge_optimize): Handle current_loops == NULL.
18211 2013-04-26 Jeff Law <law@redhat.com>
18213 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18214 (simplify_cond_using_ranges): Generalize code to simplify
18215 COND_EXPRs where one argument is a constant and the other
18216 is an SSA_NAME created by an integral type conversion.
18218 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18220 * config/arm/arm.md (store_minmaxsi): Use only when
18221 optimize_insn_for_size_p.
18223 2013-04-29 Christian Bruel <christian.bruel@st.com>
18226 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18228 2013-04-29 Richard Biener <rguenther@suse.de>
18230 PR middle-end/57089
18231 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18232 loop tree make sure to schedule a fixup for the child as well.
18233 (expand_omp_for_generic): Properly add loops.
18234 (expand_omp_for_static_nochunk): Likewise.
18235 (expand_omp_for_static_chunk): Likewise.
18236 (expand_omp_for): For the degenerate case fixup loops.
18237 (expand_omp_sections): Fix default bb placement in loops.
18238 (expand_omp_atomic_pipeline): Properly add loops.
18240 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18242 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18244 2013-04-29 Tom de Vries <tom@codesourcery.com>
18246 * tree-ssa-tail-merge.c: Update header comment.
18248 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18250 * config/aarch64/arm_neon.h
18251 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18252 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18253 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18254 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18255 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18256 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18257 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18258 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18260 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18262 * config/aarch64/aarch64-simd.md
18263 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18264 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18265 fix_trunc, fixuns_trunc.
18266 (ftrunc<VDQF:mode>2): New.
18267 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18268 (fix_trunc_optab): New.
18270 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18272 * config/aarch64/aarch64-builtins.c
18273 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18274 iceilf, lround, iroundf.
18276 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18279 * config/i386/i386.h (enum ix86_tune_indices)
18280 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18281 New, split from X86_TUNE_INTER_UNIT_MOVES.
18282 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18283 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18284 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18285 (TARGET_INTER_UNIT_MOVES): Remove.
18286 * config/i386/i386.c (initial_ix86_tune_features): Update.
18287 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18288 (ix86_expand_convert_uns_didf_sse): Use
18289 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18290 (ix86_expand_vector_init_one_nonzero): Ditto.
18291 (ix86_expand_vector_init_interleave): Ditto.
18292 (inline_secondary_memory_needed): Return true for moves from SSE class
18293 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18294 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18295 * config/i386/constraints.md (Yi, Ym): Depend on
18296 TARGET_INTER_UNIT_MOVES_TO_VEC.
18297 (Yj, Yn): New constraints.
18298 * config/i386/i386.md (*movdi_internal): Change constraints of
18299 operand 1 from Yi to Yj and from Ym to Yn.
18300 (*movsi_internal): Ditto.
18301 (*movdf_internal): Ditto.
18302 (*movsf_internal): Ditto.
18303 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18304 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18305 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18306 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18307 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18308 * config/i386/sse.md (movdi_to_sse): Ditto.
18309 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18310 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18311 TARGET_INTER_UNIT_MOVES.
18312 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18313 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18314 instead of TARGET_INTER_UNIT_MOVES.
18315 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18316 operand 1 from Yi to Yj and from Ym to Yn.
18318 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18320 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18321 (float_truncate_hi_): Likewise.
18322 (float_extend_lo_): Likewise.
18323 (float_truncate_lo_): Likewise.
18324 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18325 (aarch64_float_extend_lo_v2df): Likewise.
18326 (vec_unpacks_hi_v4sf): Likewise.
18327 (aarch64_float_truncate_lo_v2sf): Likewise.
18328 (aarch64_float_truncate_hi_v4sf): Likewise.
18329 (vec_pack_trunc_v2df): Likewise.
18330 (vec_pack_trunc_df): Likewise.
18332 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18334 * config/aarch64/aarch64-builtins.c
18335 (aarch64_fold_builtin): Fold float conversions.
18336 * config/aarch64/aarch64-simd-builtins.def
18337 (floatv2si, floatv4si, floatv2di): New.
18338 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18339 * config/aarch64/aarch64-simd.md
18340 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18341 * config/aarch64/iterators.md (FLOATUORS): New.
18342 (optab): Add float, floatuns.
18343 (su_optab): Likewise.
18345 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18347 * config/aarch64/aarch64-builtins.c
18348 (aarch64_builtin_vectorized_function): Use new names for
18350 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18351 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18352 (fcvtzu): Split as...
18353 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18354 (fcvtas): Split as...
18355 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18356 (fcvtau): Split as...
18357 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18358 (fcvtps): Split as...
18359 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18360 (fcvtpu): Split as...
18361 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18362 (fcvtms): Split as...
18363 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18364 (fcvtmu): Split as...
18365 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18366 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18367 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18368 (lfrintnusf, lfrintnudf): Likewise.
18369 * config/aarch64/aarch64-simd.md
18370 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18372 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18373 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18374 (fcvt_pattern): Likewise.
18376 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18378 * config/aarch64/aarch64-simd.md
18379 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18380 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18382 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18384 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18385 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18386 (vrnd<a,m,n,p>_f32): Implement using builtins.
18387 (vrnd<i,x><q>_f<32, 64>): New.
18389 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18391 * config/aarch64/aarch64-builtins.c
18392 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18393 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18394 (frintz): Rename to...
18396 (frintp): Rename to...
18398 (frintm): Rename to...
18400 (frinti): Rename to...
18401 (nearbyint): ...this.
18402 (frintx): Rename to...
18404 (frinta): Rename to...
18406 * config/aarch64/aarch64-simd.md
18407 (aarch64_frint<frint_suffix><mode>): Delete.
18408 (<frint_pattern><mode>2): Convert to insn.
18409 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18410 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18411 (frint_pattern): Likewise.
18412 (frint_suffix): Likewise.
18414 2013-04-29 Richard Biener <rguenther@suse.de>
18416 PR tree-optimization/57081
18417 * loop-init.c: Include tree-flow.h.
18418 (loop_optimizer_finalize): Free number of iteration estimates.
18419 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18421 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18423 PR tree-optimization/57083
18424 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18425 non-singleton shift count range, zero extend low_bound for uns case.
18427 * config/i386/predicates.md (general_vector_operand): New predicate.
18428 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18429 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18430 if they aren't nonimmediate operands. If their original values
18431 satisfy const_vector_equal_evenodd_p, don't shift them.
18432 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18433 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18434 if not nonimmediate_operand.
18435 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18436 instead of register_operand.
18437 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18439 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18441 * stor-layout.c (finalize_size_functions): Allocate a structure and
18442 reset cfun before dumping the functions.
18444 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18446 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18449 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18450 use xop_pmacsdqh if uns_p.
18451 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18452 the immediate rotate count.
18454 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18456 * rtl.h (struct rtx_def): Add comment for field jump.
18457 (LRA_SUBREG_P): New macro.
18458 * recog.c (register_operand): Check LRA_SUBREG_P.
18459 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18461 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18462 (lra_final_code_change): Skip subreg change for operators.
18463 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18464 if there are no operand changes.
18465 * lra-constraints.c (curr_insn_set): New.
18466 (match_reload): Set LRA_SUBREG_P.
18467 (emit_spill_move): Ditto.
18468 (check_and_process_move): Use curr_insn_set. Process only single
18469 set insns. Don't initialize sec_mem_p and change_p.
18470 (simplify_operand_subreg): Use LRA_SUBREG_P.
18471 (reg_in_class_p): New function.
18472 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18473 of #ifdef. Add code to remove cycling.
18474 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18475 non-null disp. Reload inner instead of disp when base and index
18476 are null. Try to put lo_sum into register.
18477 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18478 (check_and_process_move): Move code for move cost check to
18479 simple_move_p. Remove equiv_substitution.
18480 (simple_move_p): New function.
18481 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18482 curr_insn_set. Call check_and_process_move only for single set
18483 insns. Use the new function. Move call of check_and_process_move
18484 after operand equiv substitution and address process.
18486 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18489 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18490 with nonlocal goto receivers or returns twice calls, ignore
18491 unininitialized values from abnormal edges to nl goto receiver
18492 or returns twice call.
18494 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18496 PR tree-optimization/57051
18497 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18498 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18501 2013-04-26 Richard Biener <rguenther@suse.de>
18503 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18504 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18505 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18506 (fixup_loop_arrays_after_move): New function.
18507 (move_sese_region_to_fn): Properly outline the loop tree parts
18508 of the SESE region.
18510 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18512 * config/i386/i386.md (type, unit): Fix long lines.
18514 2013-04-26 Richard Biener <rguenther@suse.de>
18516 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18517 (lto-streamer-out.o): Likewise.
18518 * cfgloop.c (init_loops_structure): Export, add struct function
18519 argument and adjust.
18520 (flow_loops_find): Adjust.
18521 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18522 (init_loops_structure): Declare.
18523 * lto-streamer-in.c: Include cfgloop.h.
18524 (input_cfg): Input the loop tree.
18525 * lto-streamer-out.c: Include cfgloop.h.
18526 (output_cfg): Output the loop tree.
18527 (output_struct_function_base): Do not drop PROP_loops.
18529 2013-03-26 Richard Biener <rguenther@suse.de>
18531 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18532 (pass_build_cfg): Provide PROP_loops.
18533 (move_sese_region_to_fn): Remove loops that are outlined into fn
18535 * tree-inline.c: Include cfgloop.h.
18536 (initialize_cfun): Do not drop PROP_loops.
18537 (copy_loops): New function.
18538 (copy_cfg_body): Copy loop structure.
18539 (tree_function_versioning): Initialize destination loop tree.
18540 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18541 (pass_parallelize_loops): Do IL verification.
18542 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18543 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18544 the CFG make sure we fixup loops as well.
18545 * tree-ssa-tail-merge.c: Include cfgloop.h.
18546 (replace_block_by): When merging loop latches mark loops for fixup.
18547 * lto-streamer-out.c (output_struct_function_base): Drop
18548 PROP_loops for now.
18549 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18550 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18551 * ipa-split.c: Include cfgloop.h.
18552 (split_function): Add the new return block to the loop tree root.
18553 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18554 whether we have removed the forwarder block.
18555 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18556 * cfgloop.h (place_new_loop): Declare.
18557 * cfgloopmanip.c (place_new_loop): Export.
18558 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18559 (tree-switch-conversion.o): Likewise.
18560 (tree-complex.o): Likewise.
18561 (tree-inline.o): Likewise.
18562 (tree-ssa-tailmerge.o): Likewise.
18563 (ipa-split.o): Likewise.
18564 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18565 (tree-ssa-copy.o): Likewise.
18566 * tree-switch-conversion.c: Include cfgloop.h
18567 (process_switch): If we emit a bit-test cascade, schedule loops
18569 * tree-complex.c: Include cfgloop.h.
18570 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18571 * asan.c: Include cfgloop.h.
18572 (create_cond_insert_point): Properly add new basic-blocks to
18573 loops, schedule loop fixup.
18574 * cfgloop.c (verify_loop_structure): Check that looks are not
18576 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18578 (expand_omp_for_generic): Likewise.
18579 (expand_omp_sections): Likewise.
18580 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18581 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18582 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18583 is initialized, not when loops are present.
18584 * tree-parloops.c (parallelize_loops): Remove checking here.
18585 * passes.c (init_optimization_passes): Schedule a copy-propagation
18586 pass before complete unrolling of inner loops.
18588 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18590 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18591 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18592 (colorize_init): Add argument to _WIN32 version.
18593 * toplev.c: Include diagnostic-color.h.
18594 (process_options): Default to -fdiagnostics-color=auto if
18595 GCC_COLORS env var is in the environment.
18596 * common.opt (fdiagnostics-color=): Add Var and Init.
18597 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18598 env var is in the environment, the default is auto rather than never.
18600 * diagnostic.h (file_name_as_prefix): Add context argument.
18601 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18602 the string as locus.
18603 * langhooks.c (lhd_print_error_function): Adjust caller.
18605 2013-04-25 Lawrence Crowl <crowl@google.com>
18607 * var-tracking.c (shared_hash_def::htab):
18608 Change type to hash_table. Update dependent calls and types.
18610 2013-04-25 Lawrence Crowl <crowl@google.com>
18612 * Makefile.in: Update as needed below.
18614 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18615 Move declaration to after the type's method definitons.
18617 * attribs.c (htab_t scoped_attributes::attribute_hash):
18618 Change type to hash_table. Update dependent calls and types.
18620 * bitmap.c (htab_t bitmap_desc_hash):
18621 Change type to hash_table. Update dependent calls and types.
18623 * cselib.c (htab_t cselib_hash_table):
18624 Change type to hash_table. Update dependent calls and types.
18626 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18627 (hash_string_slot_node): Move implementation into lto-streamer.h
18628 struct string_slot_hasher.
18629 (eq_string_slot_node): Likewise.
18631 * data-streamer-out.c: Update output_block::string_hash_table
18632 dependent calls and types.
18634 * dwarf2cfi.c (htab_t trace_index):
18635 Change type to hash_table. Update dependent calls and types.
18637 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18638 Change type to hash_table. Update dependent calls and types.
18639 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18640 (htab_t optimize_external_refs::map): Likewise.
18641 (htab_t output_comp_unit::extern_map): Likewise.
18642 (htab_t output_comdat_type_unit::extern_map): Likewise.
18643 (htab_t output_macinfo::macinfo_htab): Likewise.
18644 (htab_t optimize_location_lists::htab): Likewise.
18645 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18647 * except.c (htab_t ehspec_hash_type):
18648 Change type to hash_table. Update dependent calls and types.
18649 (assign_filter_values::ttypes): Likewise.
18650 (assign_filter_values::ehspec): Likewise.
18651 (sjlj_assign_call_site_values::ar_hash): Likewise.
18652 (convert_to_eh_region_ranges::ar_hash): Likewise.
18654 * gcse.c (htab_t pre_ldst_table):
18655 Change type to hash_table. Update dependent calls and types.
18657 * ggc-common.c (htab_t saving_htab):
18658 Change type to hash_table. Update dependent calls and types.
18659 (htab_t loc_hash): Likewise.
18660 (htab_t ptr_hash): Likewise.
18661 (call_count): Rename ggc_call_count.
18662 (call_alloc): Rename ggc_call_alloc.
18663 (loc_descriptor): Rename make_loc_descriptor.
18664 (add_statistics): Rename ggc_add_statistics.
18666 * ggc-common.c (saving_htab):
18667 Change type to hash_table. Update dependent calls and types.
18669 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18670 (push_gimplify_context): Likewise.
18671 (pop_gimplify_context): Likewise.
18672 (struct gimple_temp_hash_elt): Added.
18673 (struct gimplify_hasher): Likewise.
18674 (struct gimplify_ctx.temp_htab):
18675 Change type to hash_table. Update dependent calls and types.
18677 * gimple-fold.c: Include gimplify-ctx.h.
18679 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18680 Change type to hash_table. Update dependent calls and types.
18681 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18682 avoid potential global name collision.
18684 * gimplify.c: Include gimplify-ctx.h.
18685 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18686 (htab_t gimplify_ctx::temp_htab):
18687 Update dependent calls and types for new type hash_table.
18688 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18689 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18691 * gimplify-ctx.h: New.
18692 (struct gimple_temp_hash_elt): Move from gimplify.c.
18693 (class gimplify_hasher): New.
18694 (struct gimplify_ctx): Move from gimple.h.
18695 (htab_t gimplify_ctx::temp_htab):
18696 Change type to hash_table. Update dependent calls and types.
18698 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18699 (htab_t ivs_params::newivs_index):
18700 Change type to hash_table. Update dependent calls and types.
18701 (htab_t ivs_params::params_index): Likewise.
18702 (htab_t print_generated_program::params_index): Likewise.
18703 (htab_t gloog::newivs_index): Likewise.
18704 (htab_t gloog::params_index): Likewise.
18706 * graphite.c: Include graphite-htab.h.
18707 4htab_t graphite_transform_loops::bb_pbb_mapping):
18708 Change type to hash_table. Update dependent calls and types.
18710 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18711 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18712 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18714 * graphite-dependences.c: Include graphite-htab.h.
18715 (loop_is_parallel_p): Change hash table type of parameter.
18717 * graphite-htab.h: New.
18718 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18719 (extern find_pbb_via_hash): Move from graphite-poly.h.
18720 (extern loop_is_parallel_p): Move from graphite-poly.h.
18721 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18723 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18724 (extern loop_is_parallel_p): Move to graphite-htab.h.
18725 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18727 * haifa-sched.c (htab_t delay_htab):
18728 Change type to hash_table. Update dependent calls and types.
18729 (htab_t delay_htab_i2): Likewise.
18731 * ira-color.c (htab_t allocno_hard_regs_htab):
18732 Change type to hash_table. Update dependent calls and types.
18734 * ira-costs.c (htab_t cost_classes_htab):
18735 Change type to hash_table. Update dependent calls and types.
18737 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18738 Change type to hash_table. Update dependent calls and types.
18740 * loop-iv.c (htab_t bivs):
18741 Change type to hash_table. Update dependent calls and types.
18743 * loop-unroll.c (htab_t opt_info::insns_to_split):
18744 Change type to hash_table. Update dependent calls and types.
18745 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18747 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18748 (struct string_slot_hasher): New.
18749 (htab_t output_block::string_hash_table):
18750 Change type to hash_table. Update dependent calls and types.
18752 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18753 (htab_t file_name_hash_table):
18754 Change type to hash_table. Update dependent calls and types.
18756 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18759 * lto-streamer.c (htab_t tree_htab):
18760 Change type to hash_table. Update dependent calls and types.
18762 * omp-low.c: Include gimplify-ctx.h.
18764 * passes.c (htab_t name_to_pass_map):
18765 Change type to hash_table. Update dependent calls and types.
18766 (pass_traverse): Rename to passes_pass_traverse.
18768 * plugin.c (htab_t event_tab):
18769 Change type to hash_table. Update dependent calls and types.
18771 * postreload-gcse.c (htab_t expr_table):
18772 Change type to hash_table. Update dependent calls and types.
18773 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18775 * sese.c (debug_rename_map_1): Make extern.
18776 (htab_t copy_bb_and_scalar_dependences::rename_map):
18777 Change type to hash_table. Update dependent calls and types.
18779 * sese.h (extern debug_rename_map): Move to .c file.
18781 * store-motion.c (htab_t store_motion_mems_table):
18782 Change type to hash_table. Update dependent calls and types.
18784 * trans-mem.c (htab_t tm_new_mem_hash):
18785 Change type to hash_table. Update dependent calls and types.
18787 * tree-browser.c (htab_t TB_up_ht):
18788 Change type to hash_table. Update dependent calls and types.
18790 * tree-cfg.c (htab_t discriminator_per_locus):
18791 Change type to hash_table. Update dependent calls and types.
18793 * tree-complex.c: Include tree-hasher.h
18794 (htab_t complex_variable_components):
18795 Change type to hash_table. Update dependent calls and types.
18797 * tree-eh.c (htab_t finally_tree):
18798 Change type to hash_table. Update dependent calls and types.
18800 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18801 struct int_tree_hasher.
18802 (extern int_tree_map_eq): Likewise.
18803 (uid_decl_map_hash): Removed.
18804 (extern decl_tree_map_eq): Likewise.
18806 * tree-hasher.h: New.
18807 (struct int_tree_hasher): New.
18808 (typedef int_tree_htab_type): New.
18810 * tree-inline.c: Include gimplify-ctx.h.
18812 * tree-mudflap.c: Include gimplify-ctx.h.
18814 * tree-parloops.c: Include tree-hasher.h.
18815 (htab_t eliminate_local_variables_stmt::decl_address):
18816 Change type to hash_table. Update dependent calls and types.
18817 (htab_t separate_decls_in_region::decl_copies): Likewise.
18819 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18820 Change type to hash_table. Update dependent calls and types.
18822 * tree-sra.c (candidates):
18823 Change type to hash_table. Update dependent calls and types.
18825 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18827 (int_tree_map_hash): Likewise.
18829 * tree-ssa-dom.c (htab_t avail_exprs):
18830 Change type to hash_table. Update dependent calls and types.
18832 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18833 Change type to hash_table. Update dependent calls and types.
18835 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18836 Change type to hash_table. Update dependent calls and types.
18838 * tree-ssa-phiopt.c (seen_ssa_names):
18839 Change type to hash_table. Update dependent calls and types.
18841 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18842 Change type to hash_table. Update dependent calls and types.
18844 * tree-ssa-uncprop.c (equiv):
18845 Change type to hash_table. Update dependent calls and types.
18847 2013-04-25 Jakub Jelinek <jakub@redhat.com>
18849 PR rtl-optimization/57003
18850 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18851 call note_stores with kill_clobbered_value callback again after
18852 killing regs_invalidated_by_call.
18854 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18856 * config/aarch64/aarch64-simd.md
18857 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18858 (aarch64_simd_bsl<mode>): Likewise.
18859 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18861 2013-04-25 Marek Polacek <polacek@redhat.com>
18863 PR tree-optimization/57066
18864 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
18866 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18868 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18870 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18872 * config/aarch64/aarch64-builtins.c
18873 (aarch64_fold_builtin): New.
18874 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18875 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18876 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18877 * config/aarch64/arm_neon.h
18878 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18880 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18881 Tejas Belagod <tejas.belagod@arm.com>
18883 * config/aarch64/aarch64-builtins.c
18884 (aarch64_gimple_fold_builtin): New.
18885 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18886 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18887 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18888 (addvv4sf): Update.
18889 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18891 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18893 * config/aarch64/aarch64.md
18894 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18896 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18898 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18899 (*ngcsi_uxtw): New pattern.
18901 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18902 Julian Brown <julian@codesourcery.com>
18904 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18905 (TB_DREG): Add T_V4HF.
18906 (v4hf_UP): New macro.
18907 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18908 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18909 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18911 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18912 (arm_vector_mode_supported_p): Handle V4HF.
18913 (arm_mangle_map): Handle V4HFmode.
18914 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18915 * config/arm/arm_neon_builtins.def: Add entries for
18916 vcvtv4hfv4sf, vcvtv4sfv4hf.
18917 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18918 (neon_vcvtv4hfv4sf): Likewise.
18919 * config/arm/neon-gen.ml: Handle half-precision floating point
18921 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18922 * config/arm/arm_neon.h: Regenerate.
18923 * config/arm/neon.ml (type elts): Add F16.
18924 (type vectype): Add T_float16x4, T_floatHF.
18925 (type vecmode): Add V4HF.
18926 (type features): Add Requires_FP_bit feature.
18927 (elt_width): Handle F16.
18928 (elt_class): Likewise.
18929 (elt_of_class_width): Likewise.
18930 (mode_of_elt): Refactor.
18931 (type_for_elt): Handle F16, fix error messages.
18932 (vectype_size): Handle T_float16x4.
18933 (vcvt_sh): New function.
18934 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18935 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18936 (string_of_mode): Handle V4HF.
18937 * doc/arm-neon-intrinsics.texi: Regenerate.
18939 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18941 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
18942 format specifier in 'X' case.
18944 2013-04-25 Alan Modra <amodra@gmail.com>
18947 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
18948 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
18949 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
18950 Repeat for many other rotate/shift and mask patterns using subregs.
18951 Name lshiftrt insns.
18952 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
18953 on WORDS_BIG_ENDIAN.
18955 2013-04-25 Alan Modra <amodra@gmail.com>
18957 * config.gcc: Support little-endian powerpc-linux targets.
18958 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
18959 (LINK_OS_LINUX_SPEC): Define.
18960 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
18961 Preserve MASK_LITTLE_ENDIAN.
18962 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
18963 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
18964 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
18965 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
18966 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
18967 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
18968 Correct fp word order for little-endian. Don't shift toc entries
18969 smaller than a word for little-endian.
18970 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
18971 (bswapdi2 splits): Correct low-part subreg for little-endian.
18972 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
18973 low/high where such is correct only for be.
18974 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
18975 little-endian for -mcall-aixdesc.
18977 2013-04-25 Alan Modra <amodra@gmail.com>
18979 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
18980 replace_equiv_address_nv.
18982 2013-04-25 Alan Modra <amodra@gmail.com>
18984 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
18986 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18989 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18990 * rtl.h (struct rtx_def): ...
18992 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18994 PR rtl-optimizations/57046
18995 * lra-constraints (split_reg): Set up lra_risky_transformations_p
18996 for multi-reg splits.
18998 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19000 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19002 2013-04-24 Sterling Augustine <saugustine@google.com>
19004 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19005 (comp_dir_string, debug_str_dwo_section): New.
19006 (DEBUG_STR_DWO_SECTION): Rename to ...
19007 (DEBUG_DWO_STR_SECTION): ... this.
19008 (DEBUG_NORM_STR_SECTION): Delete.
19009 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19010 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19011 (find_AT_string): Move most logic to ...
19012 (find_AT_string_in_table): ... here. New.
19013 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19014 add_skeleton_AT_string. Delete logic.
19015 (output_skeleton_debug_sections): Remove call to
19016 add_top_level_skeleton_die_attrs.
19017 (add_comp_dir_attribute): Move logic to comp_dir_string.
19018 (dwarf2out_init): Initialize debug_str_dwo_section.
19019 (output_indirect_string): Call find_string_form.
19020 (output_indirect_strings): Rewrite.
19021 (prune_unused_types): Empty skeleton_debug_str_hash.
19022 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19023 (dwarf2out_finish): Call output_indirect_strings.
19025 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19027 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19029 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19031 * rtl.h (struct rtx_def): Add comment for field jump.
19032 (LRA_SUBREG_P): New macro.
19033 * recog.c (register_operand): Check LRA_SUBREG_P.
19034 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19036 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19037 (lra_final_code_change): Skip subreg change for operators.
19038 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19039 if there are no operand changes.
19040 * lra-constraints.c (curr_insn_set): New.
19041 (match_reload): Set LRA_SUBREG_P.
19042 (emit_spill_move): Ditto.
19043 (check_and_process_move): Use curr_insn_set. Process only single
19044 set insns. Don't initialize sec_mem_p and change_p.
19045 (simplify_operand_subreg): Use LRA_SUBREG_P.
19046 (reg_in_class_p): New function.
19047 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19048 of #ifdef. Add code to remove cycling.
19049 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19050 non-null disp. Reload inner instead of disp when base and index
19051 are null. Try to put lo_sum into register.
19052 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19053 (check_and_process_move): Move code for move cost check to
19054 simple_move_p. Remove equiv_substitution.
19055 (simple_move_p): New function.
19056 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19057 curr_insn_set. Call check_and_process_move only for single set
19058 insns. Use the new function. Move call of check_and_process_move
19059 after operand equiv substitution and address process.
19061 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19063 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19064 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19065 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19067 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19069 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19071 2013-04-24 Marek Polacek <polacek@redhat.com>
19073 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19074 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19075 (select_loops_exit_conditions): Likewise.
19076 (number_of_iterations_for_all_loops): Likewise.
19077 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19078 (scev_analysis): Likewise.
19080 2013-04-02 Catherine Moore <clm@codesourcery.com>
19081 Chao-ying Fu <fu@mips.com>
19083 * config/mips/micromips.md (jraddiusp): New pattern.
19084 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19085 instruction if possible.
19087 2013-04-24 Alan Modra <amodra@gmail.com>
19089 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19091 2013-04-24 Julian Brown <julian@codesourcery.com>
19092 Chung-Lin Tang <cltang@codesourcery.com>
19094 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19095 dependency behavior in enumeration type DIE generation. Add TODO note
19096 to comments about future DW_FORM_sdata/udata re-work of related code.
19098 2013-04-23 Lawrence Crowl <crowl@google.com>
19100 * Makefile.in: Update as needed below.
19102 * hash-table.h (class hash_table):
19103 Correct many methods with parameter types compare_type to the correct
19104 value_type. (Correct code was unlikely to notice the change.)
19105 (hash_table::elements_with_deleted) New.
19106 (class hashtable::iterator): New.
19107 (hashtable::begin()): New.
19108 (hashtable::end()): New.
19109 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19111 * statistics.c (statistics_hashes):
19112 Change type to hash_table. Update dependent calls and types.
19114 * tree-into-ssa.c (var_infos):
19115 Change type to hash_table. Update dependent calls and types.
19117 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19118 Change type to hash_table. Update dependent calls and types.
19120 * tree-ssa-loop-im.c (struct mem_ref.refs):
19121 Change type to hash_table. Update dependent calls and types.
19123 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19124 Change type to hash_table. Update dependent calls and types.
19126 * tree-ssa-sccvn.c (vn_tables_s::nary):
19127 Change type to hash_table. Update dependent calls and types.
19128 (vn_tables_s::phis): Likewise.
19129 (vn_tables_s::references): Likewise.
19131 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19132 (vn_reference_eq): Update parameter and return types.
19134 * tree-ssa-structalias.c (pointer_equiv_class_table):
19135 Change type to hash_table. Update dependent calls and types.
19136 (location_equiv_class_table): Likewise.
19138 * tree-vect-data-refs.c: Consequential changes for making
19139 peeling a hash_table.
19141 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19142 (destroy_loop_vec_info): Dependent hash_table update.
19144 * tree-vectorizer.h (peeling_htab):
19145 Change type to hash_table. Update dependent calls and types.
19147 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19149 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19150 to check the register content is equal or not.
19151 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19152 to assign register content record.
19153 * lra-eliminations.c (update_reg_eliminate): Use
19154 lra_update_reg_val_offset to update register content offset.
19155 * lra-int.h (struct lra_reg): Add offset member.
19156 (lra_reg_val_equal_p): New static inline function.
19157 (lra_update_reg_val_offset): New static inline function.
19158 (lra_assign_reg_val): New static inline function.
19159 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19160 to assign register content record.
19161 (initialize_lra_reg_info_element): Initial offset to zero.
19163 2013-04-23 Catherine Moore <clm@codesourcery.com>
19165 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19166 operands. Record compression.
19168 2013-04-23 Xinliang David Li <davidxl@google.com>
19170 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19172 2013-04-23 Richard Biener <rguenther@suse.de>
19174 PR middle-end/57036
19175 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19176 parameter, only add abnormal goto edges from the copied body
19177 if the call could perform abnormal gotos.
19178 (copy_cfg_body): Adjust.
19180 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19182 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19184 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19186 * coretypes.h (gimple_stmt_iterator): Add struct to make
19189 2013-04-23 Richard Biener <rguenther@suse.de>
19191 PR tree-optimization/57026
19192 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19193 from SSA names occuring in abnormal PHI nodes.
19195 2013-04-22 Andi Kleen <ak@linux.intel.com>
19197 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19199 2013-04-22 Andi Kleen <ak@linux.intel.com>
19201 * lto/lto.c (print_lto_report_1): Declare early.
19202 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19204 2013-04-22 Andi Kleen <ak@linux.intel.com>
19206 * common.opt (-flto-report-wpa): Add.
19207 * doc/invoke.texi (-flto-report-wpa): Add.
19208 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19211 2013-04-22 Xinliang David Li <davidxl@google.com>
19213 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19214 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19215 * Makefile.in: New dependency
19217 David Daney <ddaney.cavm@gmail.com>
19219 * configure.ac (gcc_cv_as_micromips_support): Use the
19220 --fatal-warnings option.
19221 * configure: Regenerate.
19223 2013-04-22 Marek Polacek <polacek@redhat.com>
19226 * tsan.c (instrument_expr): Don't instrument expression
19227 in case its size is zero.
19229 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19233 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19235 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19237 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19239 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19240 (gimple_stmt_iterator): New typedef.
19241 * gimple.h (gimple_stmt_iterator): Rename to...
19242 (gimple_stmt_iterator_d): ... This.
19243 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19244 trees be valid for GIMPLE and GENERIC.
19245 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19246 * gimple-fold.c (gimple_fold_call): Call target hook
19247 gimple_fold_builtin.
19248 * hooks.c (hook_bool_gsiptr_false): New.
19249 * hooks.h (hook_bool_gsiptr_false): New.
19250 * target.def (fold_stmt): New.
19251 * doc/tm.texi: Regenerate.
19253 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19256 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19257 a set sp if no stack realignment.
19259 2013-04-22 Nick Clifton <nickc@redhat.com>
19261 * config.gcc (tilegx-linux): Extend extra_objs rather than
19263 (tilepro-linux): Likewise.
19265 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19267 * config/aarch64/aarch64-builtins.c
19269 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19270 (VAR<1-12>): Add MAP parameter.
19271 (BUILTIN_*): Likewise.
19272 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19273 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19274 (aarch64_ushl_n<mode>): Likewise.
19275 (aarch64_sshr_n<mode>): Likewise.
19276 (aarch64_ushr_n<mode>): Likewise.
19277 (aarch64_<maxmin><mode>): Likewise.
19278 (aarch64_sqrt<mode>): Likewise.
19279 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19280 (vshr<q>_n_*): Likewise.
19282 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19284 * config/aarch64/aarch64-builtins.c
19285 (aarch64_simd_builtin_type_mode): Handle SF types.
19287 (BUILTIN_GPF): Define.
19288 (aarch64_init_simd_builtins): Handle SF types.
19289 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19290 (frecps): Likewise.
19291 (frecpx): Likewise.
19292 * config/aarch64/aarch64-simd.md
19293 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19294 (aarch64_frecpe<mode>): New.
19295 (aarch64_frecps<mode>): Likewise.
19296 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19297 (v8type): Add frecp<esx>.
19298 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19299 (aarch64_frecps<mode>): Likewise.
19300 * config/aarch64/iterators.md (FRECP): New.
19301 (frecp_suffix): Likewise.
19302 * config/aarch64/arm_neon.h
19303 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19305 2013-04-22 Christian Bruel <christian.bruel@st.com>
19308 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19309 (REG_CLASS_NAMES): Idem.
19310 (REG_CLASS_CONTENTS): Idem.
19311 (REGCLASS_HAS_FP_REG): Idem.
19312 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19313 (sh_conditional_register_usage): Idem.
19315 2013-04-21 Jeff Law <law@redhat.com>
19317 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19318 (ssa_forward_propagate_and_combine): Use it.
19320 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19322 * lra.c: Update the flow chart diagram.
19324 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19326 PR rtl-optimization/56847
19327 * lra-constraints.c (process_alt_operands): Discourage alternative
19328 with non-matche doffsettable memory constraint fro memory with
19331 2013-04-19 Richard Biener <rguenther@suse.de>
19333 PR tree-optimization/56982
19334 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19336 * gimplify.c (gimplify_call_expr): Notice special calls.
19337 (gimplify_modify_expr): Likewise.
19338 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19339 abnormal control flow receivers.
19340 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19341 in the same way as cfun->has_nonlocal_labels.
19342 (gimple_purge_dead_abnormal_call_edges): Likewise.
19343 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19344 receivers start a basic-block.
19346 2013-04-19 Richard Biener <rguenther@suse.de>
19348 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19350 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19351 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19352 (SLP_TREE_LOAD_PERMUTATION): Add.
19353 (vect_transform_slp_perm_load): Adjust prototype.
19354 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19355 (vect_free_slp_instance): Likewise.
19356 (vect_create_new_slp_node): Likewise.
19357 (vect_supported_slp_permutation_p): Remove.
19358 (vect_slp_rearrange_stmts): Adjust.
19359 (vect_supported_load_permutation_p): Likewise. Inline
19360 vect_supported_slp_permutation_p here.
19361 (vect_analyze_slp_instance): Compute load permutations per
19362 slp node instead of per instance.
19363 (vect_get_slp_defs): Adjust.
19364 (vect_transform_slp_perm_load): Likewise.
19365 (vect_schedule_slp_instance): Remove redundant code.
19366 (vect_schedule_slp): Remove hack for PR56270, add it ...
19367 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19368 CSE loads for SLP. Adjust.
19370 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19372 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19373 spelling in two comments.
19375 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19378 * config/arm/arm.c (load_multiple_sequence): Require SP
19379 as base register for loads if SP is in the register list.
19381 2013-04-19 Martin Jambor <mjambor@suse.cz>
19383 PR tree-optimization/56718
19384 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19385 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19386 and made public. Adjusted all callers.
19387 (ipa_intraprocedural_devirtualization): New function.
19388 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19389 (ipa_intraprocedural_devirtualization): Likewise.
19390 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19392 2013-04-19 Richard Biener <rguenther@suse.de>
19394 PR tree-optimization/57000
19395 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19397 2013-04-19 Terry Guo <terry.guo@arm.com>
19399 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19401 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19402 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19403 (cortex_m4_fmacs): Use new reservations.
19404 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19406 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19408 PR rtl-optimization/56999
19409 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19411 (lra_coalesce): Remove split_origin_bitmap and related code.
19412 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19413 ranges if necessary.
19415 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19417 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19419 (ix86_expand_call): Remove clobbered_registers array and use
19420 x86_64_ms_sysv_extra_clobbered_registers instead.
19421 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19423 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19425 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19426 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19427 (*call_value_rex64_ms_sysv): Ditto.
19429 2013-04-18 Cary Coutant <ccoutant@google.com>
19431 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19432 parent instead of die_mark.
19434 2013-04-18 Diego Novillo <dnovillo@google.com>
19436 * gimple.c (create_gimple_tmp): New.
19437 (get_expr_type): New.
19438 (build_assign): New.
19439 (build_type_cast): New.
19440 * gimple.h (enum ssa_mode): Define.
19441 (gimple_seq_set_location): New.
19442 * asan.c (build_check_stmt): Change some gimple_build_* calls
19443 to use build_assign and build_type_cast.
19445 2013-04-18 Richard Biener <rguenther@suse.de>
19447 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19448 handle negative step. Remove redundant checks.
19449 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19450 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19451 for negative step and grouped loads fail to vectorize.
19453 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19455 * emit-rtl.c (reset_insn_used_flags): New function.
19456 (reset_all_used_flags): Use it.
19457 (verify_insn_sharing): New function.
19458 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19460 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19462 PR tree-optimization/56984
19463 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19464 and (x >> M) >= N don't register any assertion if N << M is the
19467 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19469 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19470 do not delete_insn insns that have not yet been emitted, only
19471 unlink them with remove_insn.
19472 * df-scan.c (df_insn_delete): Revert r197492.
19474 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19476 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19477 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19479 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19481 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19482 define_insn_and_split.
19483 (and_scc,ior_scc,negscc): Likewise.
19484 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19486 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19488 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19489 can benefit from using a sequence of LDRD instructions in epilogue
19490 instead of a single LDM instruction.
19492 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19495 * doc/extend.texi: Fix typo.
19497 2013-04-17 Richard Biener <rguenther@suse.de>
19499 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19500 (vect_build_slp_tree): ... here.
19501 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19502 match. Remove special-casing of mismatched complex loads.
19503 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19504 re-try the match with swapped commutative operands.
19505 (vect_supported_load_permutation_p): Remove special-casing of
19506 mismatched complex loads.
19507 (vect_analyze_slp_instance): Adjust.
19509 2013-04-17 Richard Biener <rguenther@suse.de>
19511 PR rtl-optimization/56921
19512 * cfgloop.h (struct loop): Add simple_loop_desc member.
19513 (struct niter_desc): Mark with GTY(()).
19514 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19515 * loop-iv.c (get_simple_loop_desc): Likewise.
19516 (free_simple_loop_desc): Likewise.
19519 2013-04-16 Richard Biener <rguenther@suse.de>
19521 PR rtl-optimization/56921
19522 * loop-init.c (pass_rtl_move_loop_invariants): Add
19523 TODO_do_not_ggc_collect to todo_flags_finish.
19524 (pass_rtl_unswitch): Same.
19525 (pass_rtl_unroll_and_peel_loops): Same.
19526 (pass_rtl_doloop): Same.
19528 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19530 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19531 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19532 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19534 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19535 * tree-streamer.c (record_common_node): Adjust reference in comment.
19537 2013-04-17 Terry Guo <terry.guo@arm.com>
19539 * config/arm/cortex-m4.md: Add a new bypass.
19541 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19543 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19545 (*subs_<optab><mode>_multp2): New pattern.
19546 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19547 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19549 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19551 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19552 (*subs_mul_imm_<mode>): New pattern.
19554 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19557 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19558 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19559 (vsx_movti_32bit): Change j->wa to O->wa.
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 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19572 * config/arm/arm.c (emit_multi_reg_push): New declaration
19573 for an existing function.
19574 (arm_emit_strd_push): New function.
19575 (arm_expand_prologue): Used here.
19576 (arm_emit_ldrd_pop): New function.
19577 (arm_expand_epilogue): Used here.
19578 (arm_get_frame_offsets): Update condition.
19579 (arm_emit_multi_reg_pop): Add a special case for load of a single
19580 register with writeback.
19582 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19584 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19587 2013-04-16 Richard Biener <rguenther@suse.de>
19589 PR tree-optimization/56756
19590 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19591 (first_mem_ref_loc): New.
19592 (execute_sm): Place the load temporarily before a previous
19593 access instead of in the latch edge to ensure its SSA dependencies
19594 are defined at points dominating the load.
19596 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19598 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19599 correct fix by moving header and footer insn to the footer of
19600 the merged basic block. Clear BB_END of the merged-away block.
19602 PR middle-end/43631
19603 * emit-rtl.c (make_note_raw): New function.
19604 (link_insn_into_chain): New static inline function.
19605 (add_insn): Use it.
19606 (add_insn_before, add_insn_after): Factor insn chain linking code...
19607 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19608 using link_insn_into_chain.
19609 (note_outside_basic_block_p): New helper function for emit_note_after
19610 and emit_note_before.
19611 (emit_note_after): Use nobb variant of add_insn_after if the note
19612 should not be contained in a basic block.
19613 (emit_note_before): Use nobb variant of add_insn_before if the note
19614 should not be contained in a basic block.
19615 (emit_note_copy): Use make_note_raw.
19616 (emit_note): Likewise.
19617 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19618 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19619 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19620 the moved barrier the tail of the basic block it follows.
19621 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19623 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19625 PR tree-optimization/56962
19626 * gimple-ssa-strength-reduction.c (record_increment): Only set
19627 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19628 either rhs1 or rhs2 is equal to c->base_expr.
19630 2013-04-15 Richard Biener <rguenther@suse.de>
19632 PR tree-optimization/56933
19633 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19635 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19636 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19637 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19638 dependence check ...
19639 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19641 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19642 GROUP_READ_WRITE_DEPENDENCE.
19644 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19646 * emit-rtl.c (reset_all_used_flags): New function.
19647 (verify_rtl_sharing): Call reset_all_used_flags before and after
19648 performing the checks.
19650 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19652 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19653 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19654 * config/arm/constraints.md (De): New constraint.
19655 * config/arm/neon.md (anddi3_neon): Delete.
19656 (neon_vand<mode>): Expand to standard anddi3 pattern.
19657 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19658 Move earlier in the file.
19659 (neon_inv_logic_op2): Likewise.
19660 (arm_anddi_operand_neon): New predicate.
19662 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19664 * configure.ac (gcc_cv_ld_as_needed): Set
19665 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19666 Use -z ignore, -z record on *-*-solaris2*.
19667 (HAVE_LD_AS_NEEDED): Update comment.
19668 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19669 * configure: Regenerate.
19670 * config.in: Regenerate.
19671 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19672 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19673 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19674 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19675 equivalents. Fix markup.
19676 * doc/tm.texi: Regenerate.
19678 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19680 * config/i386/i386.opt: New option mstack-protector-guard=.
19681 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19682 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19683 TARGET_SSP_TLS_GUARD.
19684 * config/i386/i386.c (ix86_option_override_internal): Set
19685 ix86_stack_protector_guard.
19686 * config/i386/i386.md (stack_protect_set): Enable for
19687 TARGET_SSP_TLS_GUARD only.
19688 (stack_protect_set_<mode>): Ditto.
19689 (stack_protect_test): Ditto.
19690 (stack_protect_test_<mode>): Ditto.
19691 * doc/invoke.texi (i386 Option): Document.
19693 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19696 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19697 (S_MODES): Set H_MODE bit.
19698 (SF_MODES): Set only S_MODE and SF_MODE bits.
19699 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19700 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19701 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19702 <MODE_FLOAT>: Likewise.
19704 2013-04-15 Joey Ye <joey.ye@arm.com>
19706 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19708 2013-04-15 Joey Ye <joey.ye@arm.com>
19710 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19712 (thumb_far_jump_used_p): Count instruction size and set
19715 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19717 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19718 * resource.c (mark_target_live_regs): Reformat conditional block.
19720 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19722 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19723 notes, they are emitted only just before final.
19724 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19726 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19728 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19729 * cfgrtl.c (delete_insn): Call it here instead.
19730 * lra-spills.c (lra_final_code_change): Use delete_insn.
19731 * haifa-sched.c (sched_remove_insn): Likewise.
19732 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19733 returning to the nop pool.
19734 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19735 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19737 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19739 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19740 * doc/tm.texi: Regenerated.
19742 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19744 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19747 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19749 * df-core.c (df_find_def): Compare register numbers.
19750 (df_find_use): Likewise.
19752 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19755 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19756 lra_in_progress for return.
19758 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19760 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19761 define_insn into define_insn_and_split and emit movsicc patterns.
19763 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19765 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19767 2013-04-12 Richard Biener <rguenther@suse.de>
19769 * tree-pass.h (TODO_do_not_ggc_collect): New.
19770 * passes.c (execute_one_ipa_transform_pass): Honor
19771 TODO_do_not_ggc_collect.
19772 (execute_one_pass): Likewise.
19775 2013-04-10 Richard Biener <rguenther@suse.de>
19777 * passes.c (init_optimization_passes): Remove reload pass.
19778 * ira.c (do_reload): Merge into ...
19780 (rest_of_handle_reload): Remove.
19781 (pass_reload): Likewise.
19782 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19783 of reload for vzeroupper pass placement.
19785 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19787 PR tree-optimization/56918
19788 PR tree-optimization/56920
19789 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19790 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19791 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19792 use wide_mul_with_sign method.
19794 2013-04-12 Richard Biener <rguenther@suse.de>
19796 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19797 not be considered a gimple constant.
19799 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
19801 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19802 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19805 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19806 Jakub Jelinek <jakub@redhat.com>
19808 * opts.c: Include diagnostic-color.h.
19809 (common_handle_option): Handle OPT_fdiagnostics_color_.
19810 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19811 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19812 (diagnostic-color.o): New.
19813 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19814 (diagnostic_color_rule): New enum.
19815 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19816 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19817 the location string.
19818 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19819 either NULL, or color kind.
19820 * diagnostic-color.c: New file.
19821 * diagnostic-color.h: New file.
19822 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19824 * doc/invoke.texi (-fdiagnostics-color): Document.
19825 * pretty-print.h (pp_show_color): Define.
19826 (struct pretty_print_info): Add show_color field.
19827 * diagnostic.c: Include diagnostic-color.h.
19828 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19829 macros. Colorize error:, warning: etc. strings and also the location
19831 (diagnostic_show_locus): Colorize the caret line.
19832 * pretty-print.c: Include diagnostic-color.h.
19833 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19834 inside of %< %> quotes or quoted through q format modifier.
19836 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19838 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
19840 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19842 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19843 code in CC_NZ mode.
19844 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19847 2013-04-11 Marek Polacek <polacek@redhat.com>
19849 PR tree-optimization/48184
19850 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
19852 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19854 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19855 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19856 (skip_simple_arithmetic): Tidy up.
19857 * tree.h (skip_simple_constant_arithmetic): Declare.
19859 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19861 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19863 2013-04-11 Richard Biener <rguenther@suse.de>
19865 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19866 generate vector constants.
19868 2013-04-11 Richard Biener <rguenther@suse.de>
19870 PR tree-optimization/56878
19871 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19872 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19874 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19875 Prefer to align the DR with the most invariant base address.
19877 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19879 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19882 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19884 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19885 floating-point vector comparisons against 0.
19887 2013-04-11 Jakub Jelinek <jakub@redhat.com>
19889 PR tree-optimization/56899
19890 * fold-const.c (extract_muldiv_1): Apply distributive law
19891 only if TYPE_OVERFLOW_WRAPS (ctype).
19893 2013-04-11 Bin Cheng <bin.cheng@arm.com>
19896 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19897 loading has side effect.
19899 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19901 * config/sparc/sparc.c: Include tree-pass.h.
19902 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19903 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19904 head of file. Change return type. Split off gate function.
19905 (sparc_gate_work_around_errata): New function.
19906 (pass_work_around_errata): New pass definition.
19907 (insert_pass_work_around_errata) New pass insert definition to
19908 insert pass_work_around_errata just after delayed-branch scheduling.
19909 (sparc_option_override): Insert the pass.
19910 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19912 2013-04-10 David S. Miller <davem@davemloft.net>
19914 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19915 or -mcpu=hypersparc.
19917 * target.def (cstore_mode): New hook.
19918 * target.h: Include insn-codes.h
19919 * targhooks.c: Likewise.
19920 (default_cstore_mode): New function.
19921 * targhooks.h: Declare it.
19922 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19923 * doc/tm.texi: Rebuild.
19924 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19925 target hook, rather than inspecting the insn_data.
19926 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19927 (TARGET_CSTORE_MODE): Redefine.
19928 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19930 * config/sparc/predicates.md (cstore_result_operand): New special
19932 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19933 Use it for operand 0.
19934 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19935 (*snesi_special): Likewise.
19936 (*snesi_zero): Likewise.
19937 (*seqsi_zero): Likewise.
19938 (*sltu_insn): Likewise.
19939 (*sgeu_insn): Likewise.
19940 (*seqdi_special): Make operand 0 and comparison operation be of
19942 (*snedi_special): Likewise.
19943 (*snedi_special_vis3): Likewise.
19944 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
19945 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
19946 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
19947 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
19948 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
19949 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
19950 (*sltu_extend_sp64): Likewise.
19951 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
19952 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
19953 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
19954 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
19955 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
19957 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
19959 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
19960 (aarch64_start_file): Use the new function.
19962 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19963 Jason Merrill <jason@redhat.com>
19965 * common.opt: Add -gdwarf.
19966 * opts.c (common_handle_option): Handle it.
19967 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
19969 2013-04-10 Richard Biener <rguenther@suse.de>
19971 * passes.c (execute_todo): Do not call ggc_collect conditional here.
19972 (execute_one_ipa_transform_pass): But unconditionally here.
19973 (execute_one_pass): And here.
19974 (init_optimization_passes): Remove reload pass.
19975 * tree-pass.h (TODO_ggc_collect): Remove.
19976 (pass_reload): Likewise.
19977 * ira.c (do_reload): Merge into ...
19979 (rest_of_handle_reload): Remove.
19980 (pass_reload): Likewise.
19981 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19982 of reload for vzeroupper pass placement.
19983 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
19984 and todo_flags_finish of all passes.
19986 2013-04-10 Richard Biener <rguenther@suse.de>
19988 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
19989 first_const_oprnd field, rename first_def_type to first_op_type.
19990 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
19991 (vect_get_and_check_slp_defs): Always use the type of the
19992 operand. Allow mixed vect_external_def, vect_constant_def types.
19993 (vect_get_constant_vectors): Handle mixed vect_external_def,
19994 vect_constant_def types.
19996 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
19998 PR tree-optimization/55524
19999 * tree-ssa-math-opts.c
20000 (convert_mult_to_fma): Don't use an fms construct
20001 when we don't have an fms operation, but fnma, and it looks
20002 likely that we'll be able to use the latter.
20004 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20006 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20008 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20009 inline fail caused by overwritable functions.
20011 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20013 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20014 unnecessary bits in the constant power of two case.
20016 2013-04-10 Richard Biener <rguenther@suse.de>
20018 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20019 broken code swapping operands.
20020 (vect_build_slp_tree): Do not compute load permutations here.
20021 (vect_analyze_slp_instance): Compute load permutations here,
20022 after building the SLP tree.
20024 2013-04-09 Christian Bruel <christian.bruel@st.com>
20026 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20027 of next/prev_real_insn.
20029 2013-04-09 Jan Hubicka <jh@suse.cz>
20031 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20032 Drop aliased parameter.
20033 (function_and_variable_visibility): Do not handle alias pairs.
20034 * cgraph.c (varpool_externally_visible_p): Update prototype.
20035 * varpool.c (varpool_add_new_variable): Update.
20037 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20039 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20041 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20043 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20045 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20047 2013-04-09 Marek Polacek <polacek@redhat.com>
20049 PR tree-optimization/48762
20050 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20052 2013-04-09 Richard Biener <rguenther@suse.de>
20054 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20056 (vect_build_slp_tree): Likewise.
20057 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20058 calculating the cost of a SLP instance.
20059 (vect_analyze_slp_instance): Use it from here, after building
20062 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20064 PR middle-end/56883
20065 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20066 expand_omp_for_static_chunk): Use simple_p = true in
20067 force_gimple_operand_gsi calls when assigning to addressable decls.
20069 2013-04-09 Jeff Law <law@redhat.com>
20071 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20072 when the boolean was created by converting a wider object which
20073 had a boolean range.
20075 2013-04-09 Richard Biener <rguenther@suse.de>
20077 * tree-vectorizer.h (slp_void_p): Remove.
20078 (slp_tree): Typedef before _slp_tree declaration.
20079 (struct _slp_tree): Use a vector of slp_tree as children.
20080 (vect_get_place_in_interleaving_chain): Remove.
20081 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20083 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20085 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20086 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20087 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20088 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20089 Use slp_node instead of slp_void_p and adjust.
20091 2013-04-09 Richard Biener <rguenther@suse.de>
20093 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20094 work that is not necessary.
20096 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20098 PR tree-optimization/56854
20099 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20100 forward into clobber stmts if it would change MEM_REF lhs into
20103 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20105 * tree.c (type_hash_lookup, type_hash_add): Make static.
20106 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20108 2013-04-09 Richard Biener <rguenther@suse.de>
20110 * tree.h (unsave_expr_now): Remove.
20111 * tree-inline.c (mark_local_for_remap_r): Remove.
20112 (unsave_expr_1): Likewise.
20113 (unsave_r): Likewise.
20114 (unsave_expr_now): Likewise.
20115 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20116 (propagate_tree_value): Likewise.
20118 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20120 * doc/rtl.texi (sequence): Rewrite documentation to match the
20121 current use of SEQUENCE rtl objects.
20122 * rtl.def (SEQUENCE): Likewise.
20124 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20125 Update documentation.
20126 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20127 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20129 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20131 2013-04-08 Teresa Johnson <tejohnson@google.com>
20133 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20134 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20136 (estimate_edge_size_and_time): Add comment to suggest using rounding
20138 (estimate_node_size_and_time): Ditto.
20139 (remap_edge_change_prob): Use helper rounding divide methods.
20140 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20141 (gimple_mod_pow2_value_transform): Ditto.
20142 (gimple_mod_subtract_transform): Ditto.
20143 (gimple_ic_transform): Ditto.
20144 (gimple_stringops_transform): Ditto.
20145 * stmt.c (conditional_probability): Ditto.
20146 (emit_case_dispatch_table): Ditto.
20147 * lto-cgraph.c (merge_profile_summaries): Ditto.
20148 * tree-optimize.c (execute_fixup_cfg): Ditto.
20149 * cfgcleanup.c (try_forward_edges): Ditto.
20150 * cfgloopmanip.c (scale_loop_profile): Ditto.
20152 (duplicate_loop_to_header_edge): Ditto.
20153 (lv_adjust_loop_entry_edge): Ditto.
20154 * tree-vect-loop.c (vect_transform_loop): Ditto.
20155 * profile.c (compute_branch_probabilities): Ditto.
20156 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20157 * lto-streamer-in.c (input_cfg): Ditto.
20158 * gimple-streamer-in.c (input_bb): Ditto.
20159 * ipa-cp.c (update_profiling_info): Ditto.
20160 (update_specialized_profile): Ditto.
20161 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20162 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20164 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20165 (compute_trg_info): Ditto.
20166 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20167 (purge_dead_edges): Ditto.
20168 * loop-unswitch.c (unswitch_loop): Ditto.
20169 * cgraphclones.c (cgraph_clone_edge): Ditto.
20170 (cgraph_clone_node): Ditto.
20171 * tree-inline.c (copy_bb): Ditto.
20172 (copy_edges_for_bb): Ditto.
20173 (initialize_cfun): Ditto.
20174 (copy_cfg_body): Ditto.
20175 (expand_call_inline): Ditto.
20177 2013-04-08 Kai Tietz <ktietz@redhat.com>
20179 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20180 TARGET_CYGWIN64 by TARGET_64BIT.
20182 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20184 * config/epiphany/epiphany.md (GPR_1): New constant.
20185 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20186 * config/epiphany/epiphany.c (gen_compare_reg):
20187 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20188 is already in place.
20189 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20190 Don't require being called during rtl expansion; If y operlaps r0,
20192 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20193 (epiphany_expand_epilogue): Likewise.
20195 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20196 Don't use CC_FPmode for ORDERED / UNORDERED.
20197 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20199 * config/epiphany/constraints.md (CnL): New constraint.
20200 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20201 * config/epiphany/predicates.md (add_operand): Allow 1024.
20203 * config/epiphany/epiphany.md (logical_op): New code iterator.
20204 (op_mnc): New code attribute.
20205 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20206 (mov_f+1, mov_f+2): New peephole2 patterns.
20208 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20209 (cstoresi4): Also allow re-use of zero result when doing a NE
20210 comparison to a non-zero operand.
20211 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20213 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20214 Use gen_addsi3_i / gen_subsi3_i.
20216 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20220 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20221 contain anything but clobbers, at most one __builtin_stack_restore,
20222 optionally debug stmts and final resx, and if it has at least one
20223 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20224 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20225 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20226 which isn't defaut definition, remove them.
20227 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20228 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20229 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20230 with MEM_REF LHS with SSA_NAME address.
20232 2013-04-08 Jeff Law <law@redhat.com>
20234 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20236 2013-04-08 Richard Biener <rguenther@suse.de>
20238 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20240 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20241 determined vector type.
20242 (vect_analyze_data_refs): Likewise.
20243 (vect_get_new_vect_var): Adjust.
20244 (vect_create_destination_var): Preserve SSA name versions.
20245 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20246 not dump anything here.
20248 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20250 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20251 Add member lr_slot_known.
20252 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20254 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20255 Remove code that sets lr_slot_offset according to what a previous
20256 version of epiphany_emit_save_restore used to do.
20257 (epiphany_emit_save_restore): When doing an lr save or restore,
20258 set/verify lr_slot_known and lr_slot_offset.
20260 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20263 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20266 2013-04-08 Richard Biener <rguenther@suse.de>
20268 * alias.c (find_base_term): Fix thinko in previous change.
20270 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20272 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20273 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20274 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20275 if possible to compute val.
20276 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20277 For QImode integers don't require anything about precision. Use
20278 const_with_all_bytes_same to find out if the constant doesn't have
20279 repeated bytes in it.
20281 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20283 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20286 2013-04-08 Marek Polacek <polacek@redhat.com>
20288 PR rtl-optimization/48182
20289 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20292 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20295 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20296 nuses, make sure we have a label.
20298 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20301 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20302 (rs6000_emit_swdiv_low_precision): Remove.
20303 (rs6000_emit_swdiv): Rewrite to handle between one and four
20304 iterations of Newton-Raphson generally; modify required number of
20305 iterations for some cases.
20306 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20308 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20310 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20311 set-but-unused variable.
20313 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20314 basic blocks of released function bodies garbage-collectable.
20316 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20317 (struct rtl_opt_pass): Add TODO_df_finish.
20319 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20321 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20323 * config/arm/constraints.md (q): New constraint.
20324 * config/arm/ldrdstrd.md: New file.
20325 * config/arm/arm.md (ldrdstrd.md) New include.
20326 (arm_movdi): Use "q" instead of "r" constraint
20327 for double-word memory access.
20328 (movdf_soft_insn): Likewise.
20329 * config/arm/vfp.md (movdi_vfp): Likewise.
20330 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20331 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20332 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20333 (mem_ok_for_ldrd_strd): Likewise.
20334 (output_move_double): Update assertion.
20336 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20338 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20340 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20342 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20343 define_insn_and_split.
20344 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20346 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20348 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20349 define_insn_and_split.
20350 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20351 (shiftsi3_compare): New pattern.
20352 (rrx): New pattern.
20353 * config/arm/unspecs.md (UNSPEC_RRX): New.
20355 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20357 * config/arm/arm.md (negdi_extendsidi): New pattern.
20358 (negdi_zero_extendsidi): Likewise.
20360 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20362 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20363 define_insn_and_split.
20364 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20365 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20367 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20369 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20370 define_insn_and_split.
20371 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20372 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20374 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20376 * config/arm/arm.md (subsi3_carryin): New pattern.
20377 (subsi3_carryin_const): Likewise.
20378 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20379 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20381 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20383 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20385 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20387 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20388 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20390 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20392 * config/arm/arm.c (arm_expand_builtin): Change fcode
20393 type to unsigned int.
20395 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20397 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20399 2013-04-04 Ian Lance Taylor <iant@google.com>
20401 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20404 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20406 PR middle-end/56729
20407 * df-scan.c (df_insn_delete): Disable failing assert.
20409 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20411 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20412 New function prototype.
20413 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20414 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20415 (arm_builtin_vectorized_function): New function.
20417 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20419 * config/arm/arm_neon_builtins.def: New file.
20420 * config/arm/arm.c (neon_builtin_data): Move contents to
20421 arm_neon_builtins.def.
20422 (enum arm_builtins): Include neon builtin definitions.
20423 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20424 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20426 2013-04-04 Marek Polacek <polacek@redhat.com>
20428 PR tree-optimization/48186
20429 * predict.c (maybe_hot_frequency_p): Return false if
20430 HOT_BB_FREQUENCY_FRACTION is 0.
20431 (cgraph_maybe_hot_edge_p): Likewise.
20433 2013-04-04 Richard Biener <rguenther@suse.de>
20435 PR tree-optimization/56826
20436 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20439 2013-04-04 Richard Biener <rguenther@suse.de>
20441 PR tree-optimization/56213
20442 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20443 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20445 2013-04-04 Richard Biener <rguenther@suse.de>
20447 PR tree-optimization/56837
20448 * tree-loop-distribution.c (classify_partition): For non-zero
20449 values require that the value has the same precision as its
20450 mode to be useful as memset value.
20452 2013-04-03 Nick Clifton <nickc@redhat.com>
20454 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20455 (fmssf4): Use fmsf.s on E3V5 architectures.
20456 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20457 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20459 2013-04-03 Jeff Law <law@redhat.com>
20461 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20462 (lra-eliminations.o): Likewise.
20464 2013-04-03 Teresa Johnson <tejohnson@google.com>
20466 * gcov-io.c (compute_working_sets): Moved most of body of old
20467 compute_working_sets here from profile.c.
20468 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20469 (gcov_working_set_t): Moved typedef here from basic-block.h
20470 (compute_working_set): Declare.
20471 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20472 (get_working_sets): Renamed from compute_working_set,
20473 replace most of body with call to new compute_working_sets.
20474 (get_exec_counts): Replace call to compute_working_sets
20475 to get_working_sets.
20476 * profile.h (get_working_sets): Renamed from compute_working_set.
20477 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20478 to get_working_sets.
20479 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20480 * gcov-dump.c (dump_working_sets): New function.
20482 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20484 * hwint.c (sext_hwi, zext_hwi): New functions.
20485 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20486 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20487 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20488 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20489 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20490 (sext_hwi, zext_hwi): New functions.
20492 2013-04-03 Jeff Law <law@redhat.com>
20494 PR tree-optimization/56799
20495 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20496 back test for widening conversion erroneously dropped in prior change.
20498 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20501 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20502 instead of next_real_insn.
20504 2013-04-03 Marek Polacek <polacek@redhat.com>
20507 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20509 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20512 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20514 (thumb1_output_casesi): Likewise.
20515 (thumb2_output_casesi): Likewise.
20517 2013-04-03 Richard Biener <rguenther@suse.de>
20519 PR tree-optimization/56817
20520 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20522 (tree_unroll_loops_completely_1): ... new function to manually
20523 walk the loop tree, properly defering outer loops of unrolled
20524 loops to later iterations.
20526 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20528 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20529 (vectorizable_load): Likewise.
20530 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20531 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20533 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20535 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20538 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20540 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20542 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20544 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20546 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20548 PR tree-optimization/56790
20549 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20552 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20554 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20556 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20557 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20560 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20563 * tree.h (force_folding_builtin_constant_p): New decl.
20564 * builtins.c (force_folding_builtin_constant_p): New variable.
20565 (fold_builtin_constant_p): Fold immediately also if
20566 force_folding_builtin_constant_p.
20568 2013-04-03 Richard Biener <rguenther@suse.de>
20570 PR tree-optimization/56812
20571 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20572 DRs of the same interleaving chain are independent.
20574 2013-04-02 Jason Merrill <jason@redhat.com>
20576 * gdbinit.in (pbb): Use debug fn.
20578 2013-04-02 Lawrence Crowl <crowl@google.com>
20580 * sese.h (struct ivtype_map_elt_s): Remove unused.
20581 (extern debug_ivtype_map): Remove unused.
20582 (extern eq_ivtype_map_elts): Remove unused.
20583 * sese.c (debug_ivtype_map): Removed unused.
20584 (debug_ivtype_map_1): Removed unused.
20585 (debug_ivtype_elt): Remove unused.
20586 (eq_ivtype_map_elts): Remove unused.
20589 2013-04-02 Kai Tietz <ktietz@redhat.com>
20592 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20593 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20594 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20596 (legitimize_pe_coff_symbol): Likewise.
20597 (is_imported_p): New helper-function.
20598 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20599 for Windows x64 targets.
20600 (ix86_expand_prologue): Optimize for pe-coff targets.
20601 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20602 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20603 medium/large code-model.
20604 (legitimize_pic_address): Likewise.
20605 (legitimize_tls_address): Likewise.
20606 (ix86_expand_call): Likewise.
20607 (x86_output_mi_thunk): Likewise.
20608 (get_dllimport_decl): Add new beimport argument.
20609 (construct_plt_address): Don't assert for x64 pe-coff targets.
20610 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20612 (SYMBOL_FLAG_STUBVAR): New macro.
20613 (SYMBOL_REF_STUBVAR_P): Likewise.
20614 * config/i386/winnt.c (stub_list): New structure.
20615 (stub_head): New local variable.
20616 (i386_pe_record_stub): New function.
20617 (i386_pe_file_end): Emit refptr-stubs.
20619 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20621 PR rtl-optimization/56745
20622 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20623 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20626 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20627 and both of them are MEM_REFs, just compare first argument for
20628 equality and attempt to deal even with differing offsets.
20631 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20632 of gimple_clobber_p to be MEM_REF.
20633 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20634 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20635 after gimplification.
20636 * asan.c (get_mem_ref_of_assignment): Don't instrument
20637 gimple_clobber_p stmts.
20638 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20639 gimple_clobber_p stmt if they have MEM_REF lhs and
20640 are dead because of another gimple_clobber_p stmt.
20641 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20642 gimple_clobber_p stmts like debug stmts.
20643 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20644 that refer to unused VAR_DECLs or uninitialized values.
20645 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20646 gimple_clobber_p stmts if they refer to removed parameters.
20647 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20650 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20652 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20653 using SWI48 mode attribute.
20655 2013-04-02 Wei Mi <wmi@google.com>
20657 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20658 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20659 *<rotate_insn><mode>3_mask in i386.md.
20661 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20663 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20665 2013-04-02 Richard Biener <rguenther@suse.de>
20667 PR tree-optimization/56778
20668 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20669 Runtime alias tests are not supported for gather loads.
20670 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20671 stmts referenced from SSA operands before updating SSA form.
20673 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20674 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20676 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20677 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20678 * config/arm/cortex-a53.md: New file.
20679 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20680 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20681 * config/arm/arm.c (arm_issue_rate): Likewise.
20682 * config/arm/arm-tune.md: Regenerate
20683 * config/arm/arm-tables.opt: Regenerate.
20684 * config/arm/arm-cores.def: Add cortex-a53.
20686 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20688 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20691 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20693 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20694 scalar load/store operations using B/H registers.
20695 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20697 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20699 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20701 * config/aarch64/aarch64.c
20702 (aarch64_simd_scalar_immediate_valid_for_move): New.
20703 * config/aarch64/aarch64-protos.h
20704 (aarch64_simd_scalar_immediate_valid_for_move): New.
20705 * config/aarch64/constraints.md (Dh, Dq): New.
20706 * config/aarch64/iterators.md (hq): New.
20708 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20710 * reorg.c (get_branch_condition): Deal with conditional returns.
20711 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20713 2013-04-01 Wei Mi <wmi@google.com>
20715 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20716 Truncate operand 2 using %b asm operand modifier.
20717 (*<shift_insn><mode>3_mask): Ditto.
20718 (*<rotate_insn><mode>3_mask): Ditto.
20720 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20722 PR middle-end/56798
20723 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20725 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20727 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20729 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20731 2013-03-30 Lawrence Crowl <crowl@google.com>
20733 * dse.c (clear_alias_sets): Remove never set.
20734 (disqualified_clear_alias_sets): Remove never set.
20735 (clear_alias_mode_pool): Remove never set.
20736 (dse_step0): Remove condition that is never true.
20737 (canon_address): Remove condition that is never true.
20738 (dse_step7): Remove condition that is never true.
20739 (rest_of_handle_dse): Remove condition that is never true.
20740 (rest_of_handle_dse::did_global): Remove never read from above.
20741 (dse_step2_spill): Remove never called from above.
20742 (dse_step5_spill): Remove never called from above.
20744 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20746 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20747 JUMP_TABLE_DATA changes.
20748 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20749 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20750 (Insns) <jump_table_data>: New entry.
20751 * doc/tm.texi: Regenerate.
20753 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20755 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20756 for table jump at the end of a basic block using tablejump_p.
20757 * targhooks.c (default_invalid_within_doloop): Likewise.
20758 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20759 target hook implementation that is identical to the default hook.
20760 (rs6000_invalid_within_doloop): Remove.
20762 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20763 unused variable from tablejump_p call.
20765 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20766 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20767 (INSN_DELETED_P): Likewise.
20768 (emit_jump_table_data): New prototype.
20769 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20770 after 4th as unused.
20771 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20772 * sched-vis.c (print_insn): Likewise.
20773 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20774 insn for compatibility with back ends that use next_active_insn to
20775 identify jump table data.
20776 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20777 (remove_insn): Likewise.
20778 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20780 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20781 (emit_jump_table_data): New function.
20783 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20784 basic block, a JUMP_TABLE_DATA never is.
20785 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20786 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20787 off from code handling real insns.
20788 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20789 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20790 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20791 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20792 is not a NONDEBUG_INSN_P.
20793 * ira-costs.c (scan_one_insn): Likewise.
20794 * jump.c (mark_all_labels): Likewise.
20795 (mark_jump_label_1): Likewise.
20796 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20797 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20799 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20800 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20801 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20802 JUMP_TABLE_DATA_P insns.
20803 (calculate_elim_costs_all_insns): Likewise.
20804 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20805 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20806 (delete_output_reload): Code style fixups.
20807 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20808 insn flags on this non-insn.
20809 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20810 as scheduling barriers, for pre-change compatibility.
20811 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20812 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20814 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20815 redundant JUMP_TABLE_DATA_P test.
20816 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20817 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20818 (frv_for_each_packet): Likewise.
20819 * config/i386/i386.c (min_insn_size): Likewise.
20820 (ix86_avoid_jump_mispredicts): Likewise.
20821 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20822 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20823 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20824 (mips16_insn_length): Robustify.
20825 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20826 (mips16_split_long_branches): Likewise.
20827 * config/pa/pa.c (pa_combine_instructions): Likewise.
20828 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20829 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20830 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20831 as contributing to pool range lengths.
20832 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20833 Remove redundant JUMP_TABLE_DATA_P test.
20834 (sh_loop_align): Likewise.
20835 (split_branches): Likewise.
20836 (sh_insn_length_adjustment): Likewise.
20837 * config/spu/spu.c (get_branch_target): Likewise.
20839 2013-03-29 Jan Hubicka <jh@suse.cz>
20841 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20842 gcov streaming; stream hot bb threshold to ltrans.
20843 * predict.c (get_hot_bb_threshold): Break out from ....
20844 (maybe_hot_count_p): ... here.
20845 (set_hot_bb_threshold): New function.
20846 * lto-section-in.c (lto_section_name): Add profile.
20847 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20848 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20849 and data-streamer.h
20850 (histogram_entry): New structure.
20851 (histogram, histogram_pool): New global vars.
20852 (histogram_hash): New structure.
20853 (histogram_hash::hash): New method.
20854 (histogram_hash::equal): Likewise.
20855 (account_time_size): New function.
20856 (cmp_counts): New function.
20857 (dump_histogram): New function.
20858 (ipa_profile_generate_summary): New function.
20859 (ipa_profile_write_summary): New function.
20860 (ipa_profile_read_summary): New function.
20861 (ipa_profile): Decide on threshold.
20862 (pass_ipa_profile): Add ipa_profile_write_summary and
20863 ipa_profile_read_summary.
20864 * Makefile.in (ipa.o): Update dependencies.
20865 * lto-streamer.h (LTO_section_ipa_profile): New section.
20867 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20869 * tree.h (VAR_P): New.
20871 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20874 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20876 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20878 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20879 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20880 (control_flow_insn_p): Likewise.
20881 * cfgrtl.c (duplicate_insn_chain): Likewise.
20882 * final.c (get_attr_length_1): Likewise.
20883 (shorten_branches): Likewise.
20884 (final_scan_insn): Likewise.
20885 * function.c (instantiate_virtual_regs): Likewise.
20886 * gcse.c (insert_insn_end_basic_block): Likewise.
20887 * ira-costs.c (scan_one_insn): Likewise.
20888 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20889 * lra.c (check_rtl): Likewise.
20890 * reload1.c (elimination_costs_in_insn): Likewise.
20891 * reorg.c (follow_jumps): Likewise.
20893 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20894 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20895 (thumb_far_jump_used_p): Likewise.
20896 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20897 (workaround_speculation): Likewise.
20898 (add_sched_insns_for_speculation): Likewise.
20899 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20900 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20901 (frv_for_each_packet): Likewise.
20902 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20903 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20904 (final_emit_insn_group_barriers): Likewise.
20905 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20906 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20907 (mips16_insn_length): Likewise.
20908 * config/pa/pa.c (pa_reorg): Likewise.
20909 (pa_combine_instructions): Likewise.
20910 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20911 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20912 (sh_reorg): Likewise.
20913 (split_branches): Likewise.
20914 * config/spu/spu.c (get_branch_target): Likewise.
20916 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20919 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20921 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
20922 Fix declaration name.
20924 2013-03-28 Lawrence Crowl <crowl@google.com>
20926 * graphds.h (struct graph.indicies): Remove unused.
20927 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20928 (SCOP_ORIGINAL_PDDRS): Remove unused.
20929 * sese.h (extern insert_loop_close_phis): Removed unused.
20930 (extern insert_guard_phis): Removed unused.
20931 (extern ivtype_map_elt_info): Removed unused.
20932 (new_ivtype_map_elt): Removed unused.
20933 * sese.c (ivtype_map_elt_info): Removed unused.
20935 2013-03-28 Lawrence Crowl <crowl@google.com>
20937 * Makefile.in: Add several missing include dependences.
20939 (test-dump.o): New. This object is not added to any executable,
20940 but is present for ad-hoc testing.
20942 (debug (const bitmap_head_def &)): New.
20943 (debug (const bitmap_head_def *)): New.
20945 (extern debug (const bitmap_head_def &)): New.
20946 (extern debug (const bitmap_head_def *)): New.
20948 (debug (edge_def &)): New.
20949 (debug (edge_def *)): New.
20951 (debug (basic_block_def &)): New.
20952 (debug (basic_block_def *)): New.
20954 (dump_node (const_tree, int, FILE *)): Correct source file.
20956 (debug (die_struct &)): New.
20957 (debug (die_struct *)): New.
20959 (extern debug (die_struct &)): New.
20960 (extern debug (die_struct *)): New.
20961 * gimple-pretty-print.c
20962 (debug (gimple_statement_d &)): New.
20963 (debug (gimple_statement_d *)): New.
20964 * gimple-pretty-print.h
20965 (extern debug (gimple_statement_d &)): New.
20966 (extern debug (gimple_statement_d *)): New.
20968 (debug (ira_allocno_copy &)): New.
20969 (debug (ira_allocno_copy *)): New.
20970 (debug (ira_allocno &)): New.
20971 (debug (ira_allocno *)): New.
20973 (extern debug (ira_allocno_copy &)): New.
20974 (extern debug (ira_allocno_copy *)): New.
20975 (extern debug (ira_allocno &)): New.
20976 (extern debug (ira_allocno *)): New.
20978 (debug (live_range &)): New.
20979 (debug (live_range *)): New.
20981 (debug (lra_live_range &)): New.
20982 (debug (lra_live_range *)): New.
20984 (debug (lra_live_range &)): New.
20985 (debug (lra_live_range *)): New.
20987 (debug (omega_pb_d &)): New.
20988 (debug (omega_pb_d *)): New.
20990 (extern debug (omega_pb_d &)): New.
20991 (extern debug (omega_pb_d *)): New.
20993 (debug (const rtx_def &)): New.
20994 (debug (const rtx_def *)): New.
20996 (debug_tree (tree): Move within file.
20997 (debug_raw (const tree_node &)): New.
20998 (debug_raw (const tree_node *)): New.
20999 (dump_tree_via_hooks (const tree_node *, int)): New.
21000 (debug (const tree_node &)): New.
21001 (debug (const tree_node *)): New.
21002 (debug_verbose (const tree_node &)): New.
21003 (debug_verbose (const tree_node *)): New.
21004 (debug_head (const tree_node &)): New.
21005 (debug_head (const tree_node *)): New.
21006 (debug_body (const tree_node &)): New.
21007 (debug_body (const tree_node *)): New.
21008 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21009 (debug (vec<tree, va_gc> &)): New.
21010 (debug (vec<tree, va_gc> *)): New.
21012 (extern debug (const rtx_def &)): New.
21013 (extern debug (const rtx_def *)): New.
21015 (debug_raw (simple_bitmap_def &)): New.
21016 (debug_raw (simple_bitmap_def *)): New.
21017 (debug (simple_bitmap_def &)): New.
21018 (debug (simple_bitmap_def *)): New.
21020 (extern debug (simple_bitmap_def &)): New.
21021 (extern debug (simple_bitmap_def *)): New.
21022 (extern debug_raw (simple_bitmap_def &)): New.
21023 (extern debug_raw (simple_bitmap_def *)): New.
21025 (debug (vinsn_def &)): New.
21026 (debug (vinsn_def *)): New.
21027 (debug_verbose (vinsn_def &)): New.
21028 (debug_verbose (vinsn_def *)): New.
21029 (debug (expr_def &)): New.
21030 (debug (expr_def *)): New.
21031 (debug_verbose (expr_def &)): New.
21032 (debug_verbose (expr_def *)): New.
21033 (debug (vec<rtx> &)): New.
21034 (debug (vec<rtx> *)): New.
21036 (extern debug (vinsn_def &)): New.
21037 (extern debug (vinsn_def *)): New.
21038 (extern debug_verbose (vinsn_def &)): New.
21039 (extern debug_verbose (vinsn_def *)): New.
21040 (extern debug (expr_def &)): New.
21041 (extern debug (expr_def *)): New.
21042 (extern debug_verbose (expr_def &)): New.
21043 (extern debug_verbose (expr_def *)): New.
21044 (extern debug (vec<rtx> &)): New.
21045 (extern debug (vec<rtx> *)): New.
21047 (_list_iter_cond_expr): Make inline instead of static.
21049 (debug (sreal &)): New.
21050 (debug (sreal *)): New.
21052 (extern debug (sreal &)): New.
21053 (extern debug (sreal *)): New.
21055 (extern debug_raw (const tree_node &)): New.
21056 (extern debug_raw (const tree_node *)): New.
21057 (extern debug (const tree_node &)): New.
21058 (extern debug (const tree_node *)): New.
21059 (extern debug_verbose (const tree_node &)): New.
21060 (extern debug_verbose (const tree_node *)): New.
21061 (extern debug_head (const tree_node &)): New.
21062 (extern debug_head (const tree_node *)): New.
21063 (extern debug_body (const tree_node &)): New.
21064 (extern debug_body (const tree_node *)): New.
21065 (extern debug (vec<tree, va_gc> &)): New.
21066 (extern debug (vec<tree, va_gc> *)): New.
21068 (debug (struct loop &)): New.
21069 (debug (struct loop *)): New.
21070 (debug_verbose (struct loop &)): New.
21071 (debug_verbose (struct loop *)): New.
21072 * tree-dump.c: Add header dependence.
21074 (extern debug (struct loop &)): New.
21075 (extern debug (struct loop *)): New.
21076 (extern debug_verbose (struct loop &)): New.
21077 (extern debug_verbose (struct loop *)): New.
21079 (debug (data_reference &)): New.
21080 (debug (data_reference *)): New.
21081 (debug (vec<data_reference_p> &)): New.
21082 (debug (vec<data_reference_p> *)): New.
21083 (debug (vec<ddr_p> &)): New.
21084 (debug (vec<ddr_p> *)): New.
21086 (extern debug (data_reference &)): New.
21087 (extern debug (data_reference *)): New.
21088 (extern debug (vec<data_reference_p> &)): New.
21089 (extern debug (vec<data_reference_p> *)): New.
21090 (extern debug (vec<ddr_p> &)): New.
21091 (extern debug (vec<ddr_p> *)): New.
21093 (debug (pt_solution &)): New.
21094 (debug (pt_solution *)): New.
21096 (extern debug (pt_solution &)): New.
21097 (extern debug (pt_solution *)): New.
21099 (debug (_var_map &)): New.
21100 (debug (_var_map *)): New.
21101 (debug (tree_live_info_d &)): New.
21102 (debug (tree_live_info_d *)): New.
21104 (extern debug (_var_map &)): New.
21105 (extern debug (_var_map *)): New.
21106 (extern debug (tree_live_info_d &)): New.
21107 (extern debug (tree_live_info_d *)): New.
21109 2013-03-28 Jan Hubicka <jh@suse.cz>
21111 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21113 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21115 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21116 record only when desired or required.
21118 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21120 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21121 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21123 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21125 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21126 (*andsi3_compare0_uxtw): New pattern.
21127 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21128 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21130 2013-03-28 Jan Hubicka <jh@suse.cz>
21132 * data-streamer-in.c (streamer_read_gcov_count): New function.
21133 * gimple-streamer-out.c: Include value-prof.h.
21134 (output_gimple_stmt): Output histogram.
21135 (output_bb): Use streamer_write_gcov_count.
21136 * value-prof.c: Include data-streamer.h
21137 (dump_histogram_value): Add HIST_TYPE_MAX.
21138 (stream_out_histogram_value): New function.
21139 (stream_in_histogram_value): New function.
21140 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21141 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21142 * data-streamer-out.c (streamer_write_gcov_count): New function.
21143 (streamer_write_gcov_count_stream): New function.
21144 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21145 (lto_output_node): Likewise.
21146 (input_node, input_edge): Likewise.
21147 * lto-streamer-out.c (output_cfg): Update streaming.
21148 * lto-streamer-in.c (input_cfg): Likewise.
21149 * data-streamer.h (streamer_write_gcov_count,
21150 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21151 * gimple-streamer-in.c: Include value-prof.h
21152 (input_gimple_stmt): Input histograms.
21153 (input_bb): Update profile streaming.
21155 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21157 * genmodes.c (emit_max_int): New function.
21158 (emit_insn_modes_h): Added call to emit_max_function.
21159 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21161 * machmode.def: Fixed comment.
21163 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21165 * combine.c (try_combine): Removed useless assert.
21166 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21168 2013-03-28 Marek Polacek <polacek@redhat.com>
21169 Richard Biener <rguenther@suse.de>
21171 PR tree-optimization/56695
21172 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21173 build signed result of a vector comparison.
21174 * tree-cfg.c (verify_gimple_comparison): Check that a result
21175 of a vector comparison has signed type.
21177 2013-03-28 Richard Biener <rguenther@suse.de>
21179 PR tree-optimization/37021
21180 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21181 do not restrict gaps between groups.
21182 * tree-vect-stmts.c (vectorizable_load): Properly account for
21183 a gap between groups.
21185 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21187 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21188 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21191 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21193 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21194 * doc/extend.texi (Named Address Spaces): Ditto.
21195 (Variable Attributes): Ditto.
21197 2013-03-27 Kai Tietz <ktietz@redhat.com>
21199 * config.build: Add support for cygwin x64 target.
21200 * config.gcc: Likewise.
21201 * config.host: Likewise.
21202 * configure.ac: Likewise
21203 * configure: Regenerated.
21205 2013-03-27 Kai Tietz <ktietz@redhat.com>
21207 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21208 * config/i386/t-cygwin-w64: New file.
21209 * config/i386/cygwin-w64.h: New file.
21210 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21211 and add support for x64-cygwin target.
21212 (CPP_SPEC): Likewise.
21213 (CXX_WRAP_SPEC_LIST): Undefine before define.
21214 (LIBGCJ_SONAME): Use 15 as version.
21216 2013-03-27 Richard Biener <rguenther@suse.de>
21218 PR tree-optimization/56716
21219 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21220 dumping for ref nodes.
21222 2013-03-27 Martin Jambor <mjambor@suse.cz>
21224 PR tree-optimization/55334
21225 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21226 restricted pointers to arrays.
21228 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21230 * Makefile.in (.SUFFIXES): Add .cc.
21231 (.c.o): Apply same recipe for implicit rule .cc.o.
21233 2013-03-27 Richard Biener <rguenther@suse.de>
21235 PR tree-optimization/37021
21236 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21237 REALPART/IMAGPART_EXPRs around the supported refs.
21238 * tree-ssa-structalias.c (find_func_aliases): Assume that
21239 floating-point values are not used to transfer pointers.
21241 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21243 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21244 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21245 * doc/tm.texi: Regenerate.
21246 * targhooks.h (default_has_ifunc_p): New.
21247 * targhooks.c (default_has_ifunc_p): Ditto.
21248 * config/linux-protos.h: New file.
21249 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21250 hook for linux which disables support of indirect functions in android.
21251 * config/linux-android.c: New file.
21252 * config/t-linux-android.c: Ditto.
21253 * config.gcc: Added new object file linux-android.o.
21254 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21255 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21256 * varasm.c (do_assemble_alias): Likewise.
21257 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21258 doesn't support indirect functions.
21259 * configure: Regenerate.
21261 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21264 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21265 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21268 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21270 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21272 2013-03-27 Terry Guo <terry.guo@arm.com>
21274 * config/arm/arm-cores.def: Added core cortex-r7.
21275 * config/arm/arm-tune.md: Regenerated.
21276 * config/arm/arm-tables.opt: Regenerated.
21277 * doc/invoke.texi: Added entry for core cortex-r7.
21279 2013-03-27 Walter Lee <walt@tilera.com>
21281 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21282 double-decrement of next_scratch_regno.
21284 2013-03-27 Walter Lee <walt@tilera.com>
21286 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21288 (insn_v1mulus): Ditto.
21289 (insn_v2muls): Ditto.
21291 2013-03-27 Walter Lee <walt@tilera.com>
21293 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21294 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21296 2013-03-27 Walter Lee <walt@tilera.com>
21298 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21299 (*sibcall_value): Ditto.
21301 2013-03-27 Walter Lee <walt@tilera.com>
21303 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21304 (insn_mnz_v8qi): ... this ...
21305 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21307 (insn_v<n>mnz): Replaced by ...
21308 (insn_v1mnz): ... this ...
21309 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21311 (insn_mz_<mode>): Replaced by ...
21312 (insn_mz_v8qi): ... this ...
21313 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21315 (insn_v<n>mz): Replaced by ...
21316 (insn_v1mz): ... this ...
21317 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21320 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21322 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21324 2013-03-26 Roland McGrath <mcgrathr@google.com>
21326 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21327 than fprintf with a non-constant, non-format string.
21329 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21331 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21332 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21333 operand 0 predicate.
21334 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21335 attribute. Use general_x64nomem_operand as operand 1 predicate.
21336 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21337 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21338 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21339 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21340 (mov<mode>_insv_1): Remove expander. Merge insn with
21341 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21342 Use general_x64nomem_operand as operand 1 predicate.
21343 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21344 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21345 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21346 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21347 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21348 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21349 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21350 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21351 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21352 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21353 (general_x64nomem_operand): Ditto.
21355 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21357 * config/rtems.opt: Add -pthread option.
21359 2013-03-26 Richard Biener <rguenther@suse.de>
21361 * alias.c (find_base_term): Avoid redundant and not used recursion.
21362 (base_alias_check): Get the initial base term from the caller.
21363 (true_dependence_1): Compute and pass base terms to base_alias_check.
21364 (write_dependence_p): Likewise.
21365 (may_alias_p): Likewise.
21367 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21369 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21370 PC-relative load in SI modes and above only.
21372 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21374 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21375 * config/arm/iwmmxt.md (WCGR0): Update.
21376 (WCGR1, WCGR2, WCGR3): Likewise.
21378 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21380 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21381 Use x64 and nox64 isa attributes.
21383 2013-03-26 Richard Biener <rguenther@suse.de>
21385 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21386 alignment computations and rely on get_object_alignment_1
21387 for the !TYPE_P case.
21388 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21390 2013-03-26 Walter Lee <walt@tilera.com>
21392 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21393 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21395 2013-03-25 Jeff Law <law@redhat.com>
21397 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21398 check for INTEGRAL_TYPE_P that was missing due to checking in
21399 wrong version of prior patch.
21401 2013-03-25 Walter Lee <walt@tilera.com>
21403 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21404 TILEGX_INSN_SHUFFLEBYTES1.
21405 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21407 (tilegx_builtins): Ditto.
21408 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21410 2013-03-25 Walter Lee <walt@tilera.com>
21412 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21413 (floatunssisf2): New pattern.
21414 (floatsidf2): New pattern.
21415 (floatunssidf2): New pattern.
21417 2013-03-25 Walter Lee <walt@tilera.com>
21419 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21420 tests for constraint J, K, N, P.
21422 2013-03-25 Walter Lee <walt@tilera.com>
21424 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21425 Use indirect/pcrel encoding.
21426 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21429 2013-03-25 Steve Ellcey <sellcey@mips.com>
21431 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21432 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21433 * config/mips/mips.c (mips_option_override): Set IMADD default.
21434 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21435 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21436 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21437 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21438 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21440 2013-03-25 Jeff Law <law@redhat.com>
21442 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21443 slightly to avoid creating and folding useless trees. Simplify
21444 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21446 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21448 * config/i386/i386.md (*zero_extendsidi2): Merge with
21449 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21450 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21451 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21452 targets. Clarify comment.
21454 2013-03-25 Martin Jambor <mjambor@suse.cz>
21456 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21457 pass-through jump functions differently.
21458 (ipa_read_jump_function): Likewise. Also use setter functions to set
21461 2013-03-25 Martin Jambor <mjambor@suse.cz>
21463 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21464 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21466 (ipa_get_indirect_edge_target): New function.
21467 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21468 ipa_get_indirect_edge_target. Update all callers.
21469 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21470 ipa_get_indirect_edge_target_1 instead of calling
21471 ipa_get_indirect_edge_target.
21472 (create_specialized_node): Pass aggvlas to
21473 ipcp_discover_new_direct_edges.
21475 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21477 * config/arm/arm.md (f_sels, f_seld): New types.
21478 (*cmov<mode>): New pattern.
21479 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21482 2013-03-25 Kai Tietz <ktietz@redhat.com>
21484 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21485 POSIX-printf for mingw-hosted builds.
21487 2013-03-25 Richard Biener <rguenther@suse.de>
21489 PR middle-end/56694
21490 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21491 must-not-throw stmt location.
21493 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21495 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21496 Emit load-acquire versions when acq is true.
21497 (arm_emit_store_exclusive): Add rel parameter.
21498 Emit store-release versions when rel is true.
21499 (arm_split_compare_and_swap): Use acquire-release instructions
21501 of barriers when appropriate.
21502 (arm_split_atomic_op): Likewise.
21503 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21504 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21505 (VUNSPEC_SLX): Likewise.
21506 (VUNSPEC_LDA): Likewise.
21507 (VUNSPEC_STL): Likewise.
21508 * config/arm/sync.md (atomic_load<mode>): New pattern.
21509 (atomic_store<mode>): Likewise.
21510 (arm_load_acquire_exclusive<mode>): Likewise.
21511 (arm_load_acquire_exclusivesi): Likewise.
21512 (arm_load_acquire_exclusivedi): Likewise.
21513 (arm_store_release_exclusive<mode>): Likewise.
21515 2013-03-25 Catherine Moore <clm@codesourcery.com>
21517 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21518 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21519 * config/mip/predicates.md (lwsp_swsp_operand,
21520 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21521 sb16_operand, db4_operand, db7_operand, ib3_operand,
21522 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21523 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21524 andi16_operand): New predicates.
21525 * config/mips/mips.md (compression): New attribute.
21526 (enabled): New attribute.
21527 (length): Consider compression in computing length.
21528 (shift_compression): New code attribute.
21529 (*add<mode>3): New operands. Record compression.
21530 (sub<mode>3): Likewise.
21531 (one_cmpl<mode>2): Likewise.
21532 (*and<mode>3): Likewise.
21533 (*ior<mode>3): Likewise.
21534 (unnamed pattern for xor): Likewise.
21535 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21536 (*<optab><mode>3): Likewise.
21537 (*mov<mode>_internal: Likewise.
21538 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21539 (mips_unsigned_immediate_p): New.
21540 (umips_lwsp_swsp_address_p): New.
21541 (m16_based_address_p): New.
21542 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21543 (mips_unsigned_immediate_p): New prototype.
21544 (lwsp_swsp_address_p): New prototype.
21545 (m16_based_address_p): New prototype.
21546 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21547 (mips_signed_immediate_p): New function.
21548 (m16_based_address_p): New function.
21549 (lwsp_swsp_address_p): New function.
21550 (mips_print_operand_punctuation): Recognize short delay slot insns
21551 for microMIPS.add<mode>3"
21553 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21556 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21557 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21559 2013-03-25 Richard Biener <rguenther@suse.de>
21561 PR tree-optimization/56689
21562 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21565 2013-03-25 Richard Biener <rguenther@suse.de>
21567 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21569 (memory_references): Likewise.
21570 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21571 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21572 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21573 (gather_mem_refs_in_loops): Fold into ...
21574 (analyze_memory_references): ... this. Move initialization
21575 to tree_ssa_lim_initialize.
21576 (fill_always_executed_in): Rename to ...
21577 (fill_always_executed_in_1): ... this.
21578 (fill_always_executed_in): Move contains_call computation to
21579 this new function from ...
21580 (tree_ssa_lim_initialize): ... here.
21581 (tree_ssa_lim): Call fill_always_executed_in.
21583 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21585 * postreload.c (reload_combine): Fix code detecting returns.
21587 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21589 * function.c (emit_use_return_register_into_block): On cc0 targets,
21590 do not emit the sequence between cc0 setter and user.
21592 2013-03-25 Kai Tietz <ktietz@redhat.com>
21594 * config/i386/predicates.md (local_symbolic_operand): Interpret
21595 dll-imported symbols as none-local.
21597 2013-03-25 Richard Biener <rguenther@suse.de>
21599 * tree-ssa-loop-im.c (struct depend): Remove.
21600 (struct lim_aux_data): Make depends a vec of gimples.
21601 (free_lim_aux_data): Adjust.
21602 (add_dependency): Likewise.
21603 (set_level): Likewise.
21605 2013-03-25 Richard Biener <rguenther@suse.de>
21607 PR middle-end/56434
21608 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21609 the pointer returned by calls with ECF_MALLOC set.
21611 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21613 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21615 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21617 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21618 using MMXMODE mode iterator.
21619 (*move<mode>_internal): Merge with *movv2sf_internal and
21620 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21622 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21624 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21625 (record_last_mem_set_info): Likewise.
21627 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21628 of XNEWVEC followed by memset.
21629 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21631 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21633 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21634 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21635 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21636 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21637 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21638 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21639 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21640 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21641 BARRIER_P instead of GET_CODE.
21643 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21645 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21646 inaccuracy in the probing code.
21648 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21649 (ctrapdi4): Likewise.
21651 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21653 * calls.c (expand_call): Add missing guard to code handling return
21654 of non-BLKmode structures in MSB.
21655 * function.c (expand_function_end): Likewise.
21657 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21659 * combine.c (try_combine): Adjust comment. Do not add the set of
21660 insn #0 if the destination indirectly is set or dies in insn #2.
21661 Tidy up code to distribute a new note.
21663 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21665 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21666 also for alternatives 16 and 17.
21668 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21670 * config/i386/sse.md (*mov<mode>_internal): Merge with
21671 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21672 Emit insn template depending on type attribute. Use
21673 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21674 movd instead of movq mnemonic for interunit moves. Rewrite mode
21675 attribute calculation. Remove unit attribute calculation.
21676 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21677 Set prefix_data16 attribute for DImode ssemov types.
21678 Use Ym instead of y for SSE-MMX conversion alternatives.
21679 Reorder operand constraints.
21681 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21683 * df.h (df_insn_delete): Adjust prototype.
21684 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21685 and let it decide whether mark the basic block dirty.
21686 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21687 * df-scan.c (df_insn_info_delete): New helper function, split
21688 off from df_insn_delete.
21689 (df_scan_free_bb_info): Use it.
21690 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21692 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21693 that the insn is actually an insn and it has a non-NULL basic block.
21694 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21696 2013-03-22 Richard Biener <rguenther@suse.de>
21698 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21700 (mem_ref_alloc): Do not allocate them.
21701 (refs_independent_p): Do not query or maintain a cache.
21703 2013-03-22 Richard Biener <rguenther@suse.de>
21705 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21706 (gather_mem_refs_in_loops): Do not compute it.
21707 (analyze_memory_references): Do not allocate it.
21708 (tree_ssa_lim_finalize): Do not free it.
21709 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21711 2013-03-22 Richard Biener <rguenther@suse.de>
21713 * is-a.h (as_a): Use gcc_checking_assert.
21715 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21717 * config/aarch64/aarch64.c (aarch64_print_operand): New
21718 format specifier for printing a constant in hex.
21719 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21720 format specifier for printing second operand.
21722 2013-03-22 Richard Biener <rguenther@suse.de>
21724 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21726 (gather_mem_refs_in_loops): Perform store accumulation here.
21727 (create_vop_ref_mapping_loop): Remove.
21728 (create_vop_ref_mapping): Likewise.
21729 (analyze_memory_references): Initialize refs_stored_in_loop.
21730 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21731 (record_indep_loop): Remove.
21732 (record_dep_loop): New function.
21733 (ref_indep_loop_p_1): Adjust to only walk over references
21734 in the loop, not its subloops.
21735 (ref_indep_loop_p): Rename to ...
21736 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21737 maintaining a more fine-grained cache.
21738 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21739 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21741 2013-03-22 Richard Biener <rguenther@suse.de>
21743 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21744 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21745 aggregate mem_ref_loc.
21746 (free_mem_ref_locs): Inline into ...
21747 (memref_free): ... this and adjust.
21748 (mem_ref_alloc): Adjust.
21749 (mem_ref_locs_alloc): Remove.
21750 (record_mem_ref_loc): Adjust.
21751 (get_all_locs_in_loop): Rewrite into ...
21752 (for_all_locs_in_loop): ... this iterator.
21753 (rewrite_mem_ref_loc): New functor.
21754 (rewrite_mem_refs): Use for_all_locs_in_loop.
21755 (sm_set_flag_if_changed): New functor.
21756 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21757 (ref_always_accessed): New functor.
21758 (ref_always_accessed_p): Use for_all_locs_in_loop.
21760 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21762 * tree-pass.h (PROP_gimple_lvec): New.
21763 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21764 (init_optimization_passes): Move pass_lower_vector.
21765 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21767 (pass_lower_vector): Provide PROP_gimple_lvec.
21768 (pass_lower_vector_ssa): Likewise.
21769 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21771 2013-03-21 Mark Wielaard <mjw@redhat.com>
21773 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21775 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21777 * config/i386/i386.md (*movdi_internal): Disparage slightly
21778 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21779 conversion alternatives.
21781 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21783 PR middle-end/48087
21784 * diagnostic.def (DK_WERROR): New kind.
21785 * diagnostic.h (werrorcount): Define.
21786 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21787 promoted to DK_ERROR, increment DK_WERROR counter instead of
21789 * toplev.c (toplev_main): Call print_ignored_options even if
21790 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21791 even if just werrorcount is non-zero.
21794 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21796 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21797 (string_cst_pool_decl): New function.
21798 (optimize_one_addr_into_implicit_ptr): New function.
21799 (resolve_addr_in_expr): Optimize DWARF location expression
21800 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21801 which doesn't live in memory, but has DW_AT_location or
21802 DW_AT_const_value, or refers to a string literal, into
21803 DW_OP_GNU_implicit_pointer.
21804 (optimize_location_into_implicit_ptr): New function.
21805 (resolve_addr): If removing DW_AT_location of a variable because
21806 it was DW_OP_addr of address of the variable, but the variable doesn't
21807 live in memory, try to emit const value attribute for the initializer.
21809 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21811 * tree.h (VECTOR_TYPE_P): New macro.
21812 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21813 TYPE_MODE): Use it.
21814 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21815 VEC_COND_EXPR cannot be lvalues.
21816 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21818 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21820 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21821 Restrict the transformation to equal modes.
21823 2013-03-21 Richard Biener <rguenther@suse.de>
21825 PR tree-optimization/39326
21826 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21827 (MEM_ANALYZABLE): Adjust.
21828 (record_mem_ref_loc): Move bitmap ops ...
21829 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21830 unanalyzable refs, do not record locations for it.
21831 (analyze_memory_references): Allocate ref zero as shared
21833 (refs_independent_p): Do not test for unanalyzed mems here.
21834 (ref_indep_loop_p_1): Special-case disambiguation against
21835 the unanalyzed ref.
21836 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
21838 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21840 * config/arm/arm-protos.h (tune_params): Add
21841 prefer_neon_for_64bits field.
21842 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21843 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21844 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21845 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21846 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21847 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21848 (arm_option_override): Handle -mneon-for-64bits new option.
21849 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21850 (prefer_neon_for_64bits): Declare new variable.
21851 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
21852 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
21853 (arch_enabled): Handle new arch types. Remove support for onlya8
21855 (one_cmpldi2): Use new arch names.
21856 (zero_extend<mode>di2, extend<mode>di2): Ditto.
21857 * config/arm/arm.opt (mneon-for-64bits): Add option.
21858 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21859 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21860 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21862 * doc/invoke.texi (-mneon-for-64bits): Document.
21864 2013-03-21 Richard Biener <rguenther@suse.de>
21866 PR tree-optimization/39326
21867 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21868 (sort_bbs_in_loop_postorder_cmp): New function.
21869 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
21871 2013-03-21 Richard Biener <rguenther@suse.de>
21873 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21874 (vect_insert_into_interleaving_chain): Likewise.
21875 (vect_drs_dependent_in_basic_block): Inline ...
21876 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21878 (vect_analyze_data_ref_dependence): ... here. Simplify.
21879 (vect_check_interleaving): Simplify.
21880 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21881 (vect_slp_analyze_data_ref_dependences): ... this new function.
21882 (dr_group_sort_cmp): New function.
21883 (vect_analyze_data_ref_accesses): Compute data-reference groups
21884 here instead of in vect_analyze_data_ref_dependence. Use
21885 a more efficient algorithm.
21886 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21887 vect_slp_analyze_data_ref_dependences. Call
21888 vect_analyze_data_ref_accesses earlier.
21889 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21890 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21891 (vect_slp_analyze_data_ref_dependences): New prototype.
21893 2013-03-21 Richard Biener <rguenther@suse.de>
21895 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21896 ref is stored in the loop.
21897 (find_refs_for_sm): Walk only over all stores.
21898 (store_motion_loop): Allocate from lim_bitmap_obstack.
21899 (store_motion): Likewise.
21901 2013-03-21 Richard Biener <rguenther@suse.de>
21903 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21904 Update virtual SSA form.
21906 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21908 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21909 * configure: Regenerate.
21910 * config.in: Regenerate.
21911 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21912 if !HAVE_LD_EH_FRAME_CIEV3.
21914 2013-03-21 Richard Biener <rguenther@suse.de>
21916 * tree-cfg.c (verify_expr_no_block): New function.
21917 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21918 nor DECL_VALUE_EXPR have locations with associated blocks.
21919 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21920 (clear_unused_block_pointer): Remove code dealing with
21921 blocks in DECL_DEBUG_EXPR locations.
21923 2013-03-21 Richard Biener <rguenther@suse.de>
21925 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21926 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21927 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21928 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21929 instead of DECL_DEBUG_EXPR_IS_FROM.
21930 * gimplify.c (gimplify_modify_expr): Likewise.
21931 * tree-cfg.c (verify_expr_location_1): Likewise.
21932 * tree-complex.c (create_one_component_var): Likewise.
21933 * tree-sra.c (create_access_replacement): Likewise.
21934 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21935 (clear_unused_block_pointer): Likewise.
21936 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21937 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
21938 * var-tracking.c (var_debug_decl): Likewise.
21939 (track_expr_p): Likewise.
21940 * tree-inline.c (add_local_variables): Likewise. Set
21941 DECL_HAS_DEBUG_EXPR_P after copying it.
21942 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
21943 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
21945 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21948 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
21949 * configure: Regenerate.
21950 * config.in: Regenerate.
21951 * config/i386/i386.md (*movdf_internal): Use
21952 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21953 movd instead of movq mnemonic for interunit moves.
21954 (*movdi_internal): Ditto.
21956 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21958 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
21959 (abd<mode>_3): New pattern.
21960 (aba<mode>_3): New pattern.
21961 (fabd<mode>_3): New pattern.
21963 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21965 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
21966 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
21967 occurrence of REGISTER_PREFIX as its empty string.
21969 2013-03-20 Jeff Law <law@redhat.com>
21971 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
21972 addititional equivalences for equality comparisons between an SSA_NAME
21973 and a constant where the SSA_NAME was set from a widening conversion.
21975 2013-03-20 Walter Lee <walt@tilera.com>
21977 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
21979 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21981 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
21982 depending on type attribute.
21983 (*movti_internal): Ditto.
21984 (*movtf_internal): Ditto.
21985 (*movxf_internal): Ditto.
21986 (*movdf_internal): Ditto.
21987 (*movsf_internal): Ditto.
21989 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21991 * config/i386/i386.md (*movti_internal): Set prefix attribute to
21992 maybe_vex for sselog1 and ssemov types.
21993 (*movdi_internal): Reorder operand constraints.
21994 (*movsi_internal): Ditto. Set prefix attribute to
21995 maybe_vex for sselog1 and ssemov types.
21996 (*movtf_internal): Set prefix attribute to maybe_vex
21997 for sselog1 and ssemov types.
21998 (*movdf_internal): Ditto. Set prefix_data16 attribute for
21999 DImode ssemov types. Reorder operand constraints.
22000 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22001 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22002 attribute for SImode ssemov types. Reorder operand constraints.
22004 2013-03-20 Martin Jambor <mjambor@suse.cz>
22006 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22007 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22009 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22011 * config/rs6000/predicates.md (indexed_address, update_address_mem
22012 update_indexed_address_mem): New predicates.
22013 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22014 attribute for load/store instructions.
22015 * config/rs6000/dfp.md (movsd_store): Likewise.
22016 (movsd_load): Likewise.
22017 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22018 (unnamed HI->DI extend define_insn): Likewise.
22019 (unnamed SI->DI extend define_insn): Likewise.
22020 (unnamed QI->SI extend define_insn): Likewise.
22021 (unnamed QI->HI extend define_insn): Likewise.
22022 (unnamed HI->SI extend define_insn): Likewise.
22023 (unnamed HI->SI extend define_insn): Likewise.
22024 (extendsfdf2_fpr): Likewise.
22025 (movsi_internal1): Likewise.
22026 (movsi_internal1_single): Likewise.
22027 (movhi_internal): Likewise.
22028 (movqi_internal): Likewise.
22029 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22030 attribute for load/store instructions.
22031 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22033 (mov<mode>_softfloat): Likewise.
22034 (mov<mode>_hardfloat32): Likewise.
22035 (mov<mode>_hardfloat64): Likewise.
22036 (mov<mode>_softfloat64): Likewise.
22037 (movdi_internal32): Likewise.
22038 (movdi_internal64): Likewise.
22039 (probe_stack_<mode>): Likewise.
22041 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22043 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22044 floating point, and decimal floating point to reload iterator.
22046 * config/rs6000/constraints.md (wl constraint): New constraints to
22047 return FLOAT_REGS if certain options are used to reduce the number
22048 of separate patterns that exist in the file.
22049 (wx constraint): Likewise.
22050 (wz constraint): Likewise.
22052 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22053 -mdebug=reg, print wg, wl, wx, and wz constraints.
22054 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22055 Initialize the reload functions for 64-bit binary/decimal floating
22057 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22058 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22059 create the buffer on the stack to overcome not having a 32-bit
22061 (rs6000_emit_move): Likewise.
22062 (rs6000_secondary_memory_needed_rtx): Likewise.
22063 (rs6000_alloc_sdmode_stack_slot): Likewise.
22064 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22065 via xxlxor, just like DFmode 0.0.
22067 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22068 define as 1 if we are running on a power7 or newer.
22069 (enum r6000_reg_class_enum): Add new constraints.
22071 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22072 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22073 with other moves by using conditional constraits (wg). Use LFIWZX
22074 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22075 (movsd splitter): Likewise.
22076 (movsd_hardfloat): Likewise.
22077 (movsd_softfloat): Likewise.
22079 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22080 binary and decimal floating point moves.
22081 (fmove_ok): New attributes to combine binary and decimal floating
22082 point moves, and to combine power6x (mfpgpr) moves along normal
22084 (real_value_to_target): Likewise.
22085 (f32_lr): Likewise.
22086 (f32_lm): Likewise.
22087 (f32_li): Likewise.
22088 (f32_sr): Likewise.
22089 (f32_sm): Likewise.
22090 (f32_si): Likewise.
22091 (movsf): Combine binary and decimal floating point moves. Combine
22092 power6x (mfpgpr) moves with other moves by using conditional
22093 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22094 (mov<mode> for SFmode/SDmode); Likewise.
22095 (SFmode/SDmode splitters): Likewise.
22096 (movsf_hardfloat): Likewise.
22097 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22098 (movsf_softfloat): Likewise.
22099 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22101 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22102 wx and wz constraints.
22104 * config/rs6000/constraints.md (wg constraint): New constraint to
22105 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22107 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22110 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22111 -mdebug=reg, print wg, wl, wx, and wz constraints.
22112 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22113 Initialize the reload functions for 64-bit binary/decimal floating
22115 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22116 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22117 create the buffer on the stack to overcome not having a 32-bit
22119 (rs6000_emit_move): Likewise.
22120 (rs6000_secondary_memory_needed_rtx): Likewise.
22121 (rs6000_alloc_sdmode_stack_slot): Likewise.
22122 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22123 via xxlxor, just like DFmode 0.0.
22125 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22126 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22127 with other moves by using conditional constraits (wg). Use LFIWZX
22128 and STFIWX for loading SDmode on power7.
22129 (movdd splitters): Likewise.
22130 (movdd_hardfloat32): Likewise.
22131 (movdd_softfloat32): Likewise.
22132 (movdd_hardfloat64_mfpgpr): Likewise.
22133 (movdd_hardfloat64): Likewise.
22134 (movdd_softfloat64): Likewise.
22136 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22137 64-bit binary and decimal floating point moves.
22138 (FMOVE64X): Likewise.
22139 (movdf): Combine 64-bit binary and decimal floating point moves.
22140 Combine power6x (mfpgpr) moves with other moves by using
22141 conditional constraits (wg).
22142 (mov<mode> for DFmode/DDmode): Likewise.
22143 (DFmode/DDmode splitters): Likewise.
22144 (movdf_hardfloat32): Likewise.
22145 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22146 (movdf_softfloat32): Likewise.
22147 (movdf_hardfloat64_mfpgpr): Likewise.
22148 (movdf_hardfloat64): Likewise.
22149 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22150 (movdf_softfloat64): Likewise.
22151 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22152 (reload_<mode>_load): Move to later in the file so they aren't in
22153 the middle of the floating point move insns.
22154 (reload_<mode>_store): Likewise.
22156 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22159 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22160 constraint if -mdebug=reg.
22161 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22162 Enable using dd reload support if needed.
22164 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22165 binary and decimal floating point moves in rs6000.md.
22166 (movtd_internal): Likewise.
22168 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22169 decimal floating point moves.
22171 (movtf_internal): Likewise.
22172 (mov<mode>_internal, TDmode/TFmode): Likewise.
22173 (movtf_softfloat): Likewise.
22174 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22176 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22177 movdi_internal64, using wg constraint for move direct operations.
22178 (movdi_internal64): Likewise.
22180 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22181 MODES_TIEABLE_P for selected modes. Print the numerical value of
22182 the various virtual registers. Use GPR/FPR first/last values,
22183 instead of hard coding the register numbers. Print which modes
22184 have reload functions registered.
22185 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22186 settings before/after setting cpu, target and subtarget settings.
22187 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22188 and for secondary reload failures in rs6000_secondary_reload_inner.
22189 (rs6000_secondary_reload_fail): Likewise.
22190 (rs6000_secondary_reload_inner): Likewise.
22192 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22193 macros for first/last GPR and FPR registers.
22194 (LAST_GPR_REGNO): Likewise.
22195 (FIRST_FPR_REGNO): Likewise.
22196 (LAST_FPR_REGNO): Likewise.
22198 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22199 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22200 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22201 (vcond<mode><mode>): Likewise.
22202 (vcondu<mode><mode>): Likewise.
22203 (vector_gtu<mode>): Likewise.
22204 (vector_gte<mode>): Likewise.
22205 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22206 to prevent the compiler from converting DImode operations to TImode.
22207 (ior<mode>3): Likewise.
22208 (and<mode>3): Likewise.
22209 (one_cmpl<mode>2): Likewise.
22210 (nor<mode>3): Likewise.
22211 (andc<mode>3): Likewise.
22213 * config/rs6000/constraints.md (wt constraint): New constraint
22214 that returns VSX_REGS if TImode is allowed in VSX registers.
22216 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22217 constant under VSX.
22219 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22220 similar to TImode, but it is restricted to being in the GPRs.
22222 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22223 TImode to occupy a single VSX register.
22225 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22226 -mvsx-timode for power7/power8.
22227 (power7 cpu): Likewise.
22228 (power8 cpu): Likewise.
22230 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22231 sure that TFmode/TDmode take up two registers if they are ever
22232 allowed in the upper VSX registers.
22233 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22235 (rs6000_init_hard_regno_mode_ok): Likewise.
22236 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22237 constraint. Print if LRA is turned on.
22238 (rs6000_option_override_internal): Give an error if -mvsx-timode
22239 and VSX is not enabled.
22240 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22241 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22242 to reg+offset addressing. Use PTImode when checking offset
22243 addresses for validity.
22244 (reg_offset_addressing_ok_p): Likewise.
22245 (rs6000_legitimate_offset_address_p): Likewise.
22246 (rs6000_legitimize_address): Likewise.
22247 (rs6000_legitimize_reload_address): Likewise.
22248 (rs6000_legitimate_address_p): Likewise.
22249 (rs6000_eliminate_indexed_memrefs): Likewise.
22250 (rs6000_emit_move): Likewise.
22251 (rs6000_secondary_reload): Likewise.
22252 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22253 reloads to fpr registers to continue to use reg+offset addressing,
22254 but 64-bit reloads to altivec registers need reg+reg addressing.
22255 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22256 it. Treat LO_SUM like a PLUS operation.
22257 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22258 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22259 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22260 registers to share a register with a smaller sized type, since VSX
22261 puts scalars in the upper 64-bits.
22262 (print_operand): Add support for PTImode.
22263 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22264 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22265 registers, but don't have arithmetic support.
22266 (rs6000_memory_move_cost): Add test for VSX.
22267 (rs6000_opt_masks): Add -mvsx-timode.
22269 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22272 (VSr): Use wt constraint for TImode.
22273 (VSv): Drop TImode support.
22274 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22275 (vsx_movti_64bit): Likewise.
22276 (vsx_movti_32bit): Likewise.
22277 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22278 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22279 one '?' on the appropriate output constraint. Do not allow TImode
22280 logical operations on 32-bit systems.
22281 (vsx_ior<mode>3): Likewise.
22282 (vsx_xor<mode>3): Likewise.
22283 (vsx_one_cmpl<mode>2): Likewise.
22284 (vsx_nor<mode>3): Likewise.
22285 (vsx_andc<mode>3): Likewise.
22286 (vsx_concat_<mode>): Likewise.
22287 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22289 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22290 OPTION_MASK_VSX_TIMODE.
22291 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22292 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22294 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22295 (TI2 iterator): New iterator for TImode, PTImode.
22296 (wd mode attribute): Add values for vector types.
22297 (movti_string): Replace TI move operations with operations for TImode
22298 and PTImode. Add support for TImode being allowed in VSX registers.
22299 (mov<mode>_string, TImode/PTImode): Likewise.
22300 (movti_ppc64): Likewise.
22301 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22302 (TI mode splitters): Likewise.
22304 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22307 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22309 PR tree-optimization/56355
22310 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22311 Also handle integers with undefined overflow.
22313 2013-03-20 Catherine Moore <clm@codesourcery.com>
22314 Maciej W. Rozycki <macro@codesourcery.com>
22315 Tom de Vries <tom@codesourcery.com>
22316 Nathan Sidwell <nathan@codesourcery.com>
22317 Iain Sandoe <iain@codesourcery.com>
22318 Nathan Froyd <froydnj@codesourcery.com>
22319 Chao-ying Fu <fu@mips.com>
22321 * doc/extend.texi: (micromips, nomicromips, nocompression):
22322 Document new function attributes.
22323 * doc/invoke.texi (minterlink-compressed, mmicromips,
22324 m14k, m14ke, m14kec): Document new options.
22325 (minterlink-mips16): Update documentation.
22326 * doc/md.texi (ZC, ZD): Document new constraints.
22327 * configure.ac (gcc_cv_as_micromips): Check if linker
22328 supports the .set micromips directive.
22329 * configure: Regenerate.
22330 * config.in: Regenerate.
22331 * config/mips/mips-tables.opt: Regenerate.
22332 * config/mips/micromips.md: New file.
22333 * constraints.md (ZC, ZD): New constraints.
22334 * config/mips/predicates.md (movep_src_register): New predicate.
22335 (movep_src_operand): New predicate.
22336 (non_volatile_mem_operand): New predicate.
22337 * config/mips/mips.md (multimem): New type.
22338 (length): Differentiate between 17-bit and 18-bit branch offsets.
22339 (MOVEP1, MOVEP2): New mode iterator.
22340 (mov_<load>l): Use ZC constraint.
22341 (mov_<load>r): Likewise.
22342 (mov_<store>l): Likewise.
22343 (mov_<store>r): Likewise.
22344 (*branch_equality<mode>_inverted): Add microMIPS support.
22345 (*branch_equality<mode>): Likewise.
22346 (*jump_absolute): Likewise.
22347 (indirect_jump_<mode>): Likewise.
22348 (tablejump_<mode>): Likewise.
22349 (<optab>_internal): Likewise.
22350 (sibcall_internal): Likewise.
22351 (sibcall_value_internal): Likewise.
22352 (prefetch): Use constraint ZD.
22353 * config/mips/mips.opt (minterlink-compressed): New option.
22354 (minterlink-mips16): Now an alias for minterlink-compressed.
22355 (mmicromips): New option.
22356 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22357 (compare_and_swap_12): Likewise.
22358 (sync_add<mode>): Likewise.
22359 (sync_<optab>_12): Likewise.
22360 (sync_old_<optab>_12): Likewise.
22361 (sync_new_<optab>_12): Likewise.
22362 (sync_nand_12): Likewise.
22363 (sync_old_nand_12): Likewise.
22364 (sync_new_nand_12): Likewise.
22365 (sync_sub<mode>): Likewise.
22366 (sync_old_add<mode>): Likewise.
22367 (sync_old_sub<mode>): Likewise.
22368 (sync_new_add<mode>): Likewise.
22369 (sync_new_sub<mode>): Likewise.
22370 (sync_<optab><mode>): Likewise.
22371 (sync_old_<optab><mode>): Likewise.
22372 (sync_new_<optab><mode>): Likewise.
22373 (sync_nand<mode>): Likewise.
22374 (sync_old_nand<mode>): Likewise.
22375 (sync_new_nand<mode>): Likewise.
22376 (sync_lock_test_and_set<mode>): Likewise.
22377 (test_and_set_12): Likewise.
22378 (atomic_compare_and_swap<mode>): Likewise.
22379 (atomic_exchange<mode>_llsc): Likewise.
22380 (atomic_fetch_add<mode>_llsc): Likewise.
22381 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22382 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22383 (umips_save_restore_pattern_p): Likewise.
22384 (umips_load_store_pair_p): Likewise.
22385 (umips_output_load_store_pair): Likewise.
22386 (umips_movep_target_p): Likewise.
22387 (umips_12bit_offset_address_p): Likewise.
22388 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22389 (mips_base_mips16): Rename this...
22390 (mips_base_compression_flags): ...to this. Update all uses.
22391 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22392 (mips_mips16_decl_p): Delete.
22393 (mips_nomips16_decl_p): Delete.
22394 (mips_get_compress_on_flags): New function.
22395 (mips_get_compress_off_flags): New function.
22396 (mips_get_compress_mode): New function.
22397 (mips_get_compress_on_name): New function.
22398 (mips_get_compress_off_name): New function.
22399 (mips_insert_attributes): Support multiple compression types.
22400 (mips_merge_decl_attributes): Likewise.
22401 (umips_12bit_offset_address_p): New function.
22402 (mips_start_function_definition): Emit .set micromips directive.
22403 (mips_call_may_need_jalx_p): New function.
22404 (mips_function_ok_for_sibcall): Add microMIPS support.
22405 (mips_print_operand_punctuation): Support short delay slots and
22407 (umips_swm_mask, umips_swm_encoding): New.
22408 (umips_build_save_restore): New function.
22409 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22410 (was_mips16_p): Remove.
22411 (old_compression_mode): New.
22412 (mips_set_compression_mode): New function.
22413 (mips_set_current_function): Add microMIPS support.
22414 (mips_option_override): Likewise.
22415 (umips_save_restore_pattern_p): New function.
22416 (umips_output_save_restore): New function.
22417 (umips_load_store_pair_p_1): New function.
22418 (umips_load_store_pair_p): New function.
22419 (umips_output_load_store_pair_1): New function.
22420 (umips_output_load_store_pair): New function.
22421 (umips_movep_target_p) New function.
22422 (mips_prepare_pch_save): Add microMIPS support.
22423 * config/mips/mips.h (TARGET_COMPRESSION): New.
22424 (TARGET_CPU_CPP_BUILTINS): Update macro
22425 to use new compression flags and to support microMIPS.
22426 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22427 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22428 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22429 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22430 (ASM_SPEC): Support mmicromips and mno-micromips.
22431 (M16STORE_REG_P): New macro.
22432 (MIPS_CALL): Support TARGET_MICROMIPS.
22433 (MICROMIPS_J): New macro.
22434 (mips_base_mips16): Rename this...
22435 (mips_base_compression_flags): ...to this.
22436 (UMIPS_12BIT_OFFSET_P): New macro.
22437 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22438 (MULTILIB_DIRNAMES): Likewise.
22439 2013-03-20 Richard Biener <rguenther@suse.de>
22441 PR tree-optimization/56661
22442 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22443 the result does not have to be distinct.
22445 2013-03-20 Richard Biener <rguenther@suse.de>
22447 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22450 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22451 Steven Bosscher <steven@gcc.gnu.org>
22453 PR rtl-optimization/56605
22454 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22456 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22459 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22460 that require movd instead of movq.
22462 2013-03-20 Richard Biener <rguenther@suse.de>
22464 * tree-ssa-structalias.c (struct variable_info): Add pointer
22465 to the first field of an aggregate with sub-vars. Make
22466 this and the pointer to the next subfield its ID.
22467 (vi_next): New function.
22468 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22469 storedanything_id, integer_id): Increment by one.
22470 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22471 get_call_clobber_vi): Adjust.
22472 (solution_set_expand): Simplify and speedup.
22473 (solution_set_add): Inline into ...
22474 (set_union_with_increment): ... this. Adjust accordingly.
22475 (do_sd_constraint): Likewise.
22476 (do_ds_constraint): Likewise.
22477 (do_complex_constraint): Simplify.
22478 (build_pred_graph): Adjust.
22479 (solve_graph): Likewise. Simplify and speedup.
22480 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22481 get_constraint_for_component_ref, get_constraint_for_1,
22482 first_vi_for_offset, first_or_preceding_vi_for_offset,
22483 create_function_info_for, create_variable_info_for_1,
22484 create_variable_info_for, intra_create_variable_infos): Adjust.
22485 (init_base_vars): Push NULL for ID zero.
22486 (compute_points_to_sets): Adjust.
22488 2013-03-20 Richard Biener <rguenther@suse.de>
22490 * cfgloop.c (verify_loop_structure): Streamline and avoid
22491 ICEing on corrupt loop tree.
22492 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22495 2013-03-20 Richard Biener <rguenther@suse.de>
22497 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22498 check whether an SSA update is needed.
22500 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22502 * config/mips/constraints.md (T): Rename to...
22506 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22507 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22509 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22511 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22512 (*subsi3_carryin_uxtw): Likewise.
22514 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22516 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22517 (*rorsi3_insn_uxtw): Likewise.
22519 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22521 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22522 (*extrsi5_insn_uxtw): Likewise.
22524 2013-03-19 Richard Biener <rguenther@suse.de>
22526 PR tree-optimization/56273
22527 * passes.c (init_optimization_passes): Move second VRP after DOM.
22529 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22531 * config/i386/i386.md (*movti_internal): Merge from
22532 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22533 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22534 nox64 isa attributes.
22536 2013-03-18 Richard Biener <rguenther@suse.de>
22538 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22540 (merge_node_constraints): Likewise.
22541 (build_succ_graph): Likewise.
22542 (valid_graph_edge): Inline into single caller.
22543 (unify_nodes): Likewise. Use bitmap_set_bit return value
22545 (scc_visit): Fix formatting and variable use.
22546 (do_sd_constraint): Use gcc_checking_assert.
22547 (do_ds_constraint): Likewise.
22548 (do_complex_constraint): Likewise.
22549 (condense_visit): Likewise. Cleanup.
22550 (dump_pred_graph): New function.
22551 (perform_var_substitution): Dump the pred-graph before
22552 variable substitution.
22553 (find_equivalent_node): Use gcc_checking_assert.
22554 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22556 2013-03-18 Richard Biener <rguenther@suse.de>
22558 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22559 Remove cond_expr_stmt_list argument and do not gimplify the
22561 (vect_loop_versioning): Adjust.
22562 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22563 Cleanup to use less temporaries.
22564 (vect_create_data_ref_ptr): Cleanup.
22566 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22568 PR tree-optimization/56635
22569 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22570 require types_compatible_p types.
22572 2013-03-18 Nick Clifton <nickc@redhat.com>
22574 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22575 spurious backslash.
22577 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22578 Add missing line to comment describing function.
22580 2013-03-18 Richard Biener <rguenther@suse.de>
22582 PR tree-optimization/56210
22583 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22584 Handle string / character search functions.
22585 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22587 2013-03-18 Richard Biener <rguenther@suse.de>
22589 PR middle-end/56483
22590 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22591 and implement properly.
22592 * gimple.h (gimple_cond_single_var_p): Remove.
22594 2013-03-18 Richard Biener <rguenther@suse.de>
22596 * tree-data-ref.h (find_data_references_in_loop): Declare.
22597 * tree-data-ref.c (get_references_in_stmt): Use a stack
22598 vector pre-allocated in the callers.
22599 (find_data_references_in_stmt): Adjust.
22600 (graphite_find_data_references_in_stmt): Likewise.
22601 (create_rdg_vertices): Likewise.
22602 (find_data_references_in_loop): Export.
22603 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22604 Compute dependences here...
22605 (vect_analyze_data_refs): ...not here. When we encounter
22606 a non-vectorizable data reference in basic-block vectorization
22607 truncate the data reference vector. Do not bother to
22608 fixup data-dependence information for gather loads.
22609 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22610 of data references, as reported.
22612 2013-03-18 Richard Biener <rguenther@suse.de>
22614 PR tree-optimization/3713
22615 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22616 has_constants and expr.
22617 (stmt_has_constants): Properly valueize SSA names when deciding
22618 whether the stmt has constants.
22620 2013-03-18 Richard Biener <rguenther@suse.de>
22622 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22623 whole function when there is nothing to do.
22624 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22625 * tree-vectorizer.c (vectorize_loops): Update virtual and
22626 loop-closed SSA once.
22627 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22629 2013-03-18 Richard Biener <rguenther@suse.de>
22631 PR middle-end/56113
22632 * domwalk.c (bb_postorder): New global static.
22633 (cmp_bb_postorder): New function.
22634 (walk_dominator_tree): Replace scheme imposing an order for
22635 visiting dominator sons by one sorting them at the time they
22636 are pushed on the stack.
22638 2013-03-18 Richard Biener <rguenther@suse.de>
22640 PR tree-optimization/39326
22641 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22642 (struct mem_ref): Replace mem member with ao_ref typed member.
22643 (MEM_ANALYZABLE): Adjust.
22644 (memref_eq): Likewise.
22645 (mem_ref_alloc): Likewise.
22646 (gather_mem_refs_stmt): Likewise.
22647 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22648 (execute_sm_if_changed_flag_set): Adjust.
22649 (execute_sm): Likewise.
22650 (ref_always_accessed_p): Likewise.
22651 (refs_independent_p): Likewise.
22652 (can_sm_ref_p): Likewise.
22654 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22657 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22658 return 1 even for !unsignedp.
22660 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22662 * config/i386/i386.md (isa): Add x64 and nox64.
22663 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22664 (*pushtf): Enable *roF alternative for x64 isa only.
22665 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22666 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22667 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22668 (*movtf_internal): Merge from *movtf_internal_rex64 and
22669 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22670 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22671 nox64 isa attributes.
22672 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22673 nox64 isa attributes.
22674 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22676 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22678 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22680 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22683 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22684 class names. Remove trailing comma after "ALL_REGS".
22686 2013-03-16 Jan Hubicka <jh@suse.cz>
22688 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22689 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22690 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22691 of cgraph_get_create_node.
22692 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22694 2013-03-16 Jason Merrill <jason@redhat.com>
22697 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22698 with DW_AT_default_value.
22700 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22702 * BASE-VER: Set to 4.9.0.
22704 2013-03-14 Andi Kleen <ak@linux.intel.com>
22707 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22708 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22709 Document _x* TSX intrinsics.
22711 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22712 David Holsgrove <david.holsgrove@xilinx.com>
22714 * configure.ac: Add MicroBlaze TLS support detection.
22715 * configure: Regenerate.
22716 * config/microblaze/microblaze-protos.h
22717 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22718 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22719 * config/microblaze/microblaze.c (microblaze_address_type): Add
22720 ADDRESS_TLS and tls_reloc address types.
22721 (microblaze_address_info): Add tls_reloc.
22722 (TARGET_HAVE_TLS): Define.
22723 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22724 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22725 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22726 load_tls_operand, microblaze_call_tls_get_addr,
22727 microblaze_legitimize_tls_address): New functions.
22728 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22729 (get_base_reg): Use microblaze_tls_symbol_p.
22730 (microblaze_classify_address): Handle TLS.
22731 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22732 label_mentioned_p and microblaze_tls_referenced_p.
22733 (microblaze_legitimize_address): Handle TLS.
22734 (microblaze_address_insns): Handle ADDRESS_TLS.
22735 (pic_address_needs_scratch): Handle TLS.
22736 (print_operand_address): Handle TLS.
22737 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22738 (microblaze_expand_move): Handle TLS.
22739 (microblaze_legitimate_constant_p): Check
22740 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22741 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22742 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22743 (PIC_OFFSET_TABLE_REGNUM): Set.
22744 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22745 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22746 (addsi3, movsi_internal2, movdf_internal): Update constraints
22747 * config/microblaze/predicates.md (arith_plus_operand): Define
22748 (move_operand): Redefine as move_src_operand,
22749 check microblaze_tls_referenced_p.
22751 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22753 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22754 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22756 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22758 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22761 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22763 PR tree-optimization/53265
22764 * common.opt (Waggressive-loop-optimizations): New option.
22765 * tree-ssa-loop-niter.c: Include tree-pass.h.
22766 (do_warn_aggressive_loop_optimizations): New function.
22767 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22768 if number_of_latch_executions returned constant.
22769 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22770 early. If number_of_latch_executions returned constant, set
22771 nb_iterations_upper_bound back to it.
22772 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22774 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22775 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22777 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22778 (MULTILIB_OSDIRNAMES): Set.
22779 * genmultilib: If defaultosdirname doesn't start with :: , set
22780 defaultosdirname2 instead, clear it and emit two . multilib_raw
22781 entries instead of just one.
22783 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22785 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22786 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22787 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22788 (SUBTARGET_OVERRIDE_OPTIONS): New.
22790 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22793 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22794 (musermode): Convert to Var(TARGET_USERMODE).
22795 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22796 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22797 * config/sh/sh.c (sh_option_override): Use
22798 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22799 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22801 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22803 (udivsi3_i4_single, divsi3_i4_single): Use
22804 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22806 2013-03-13 Dave Korn <dave.korn.cygwin@....>
22808 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22811 2013-03-13 Richard Biener <rguenther@suse.de>
22813 PR tree-optimization/56608
22814 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22815 calls when vectorizing basic-blocks.
22817 2013-03-13 Jakub Jelinek <jakub@redhat.com>
22820 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22823 2013-03-12 Jakub Jelinek <jakub@redhat.com>
22825 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22827 2013-03-11 Jan Hubicka <jh@suse.cz>
22830 * lto-streamer-out.c (output_symbol_p): Skip references from
22831 constructors of external variables.
22833 2013-03-11 Jan Hubicka <jh@suse.cz>
22835 PR middle-end/56571
22836 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22838 * emit-rtl.c (verify_rtx_sharing): Likewise.
22839 (copy_insn_1): Likewise.
22840 * rtl.c (copy_rtx): Likewise.
22842 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
22845 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22846 output_operand_lossage message.
22848 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
22851 * arm.c (shift_op): Validate RTL pattern on the fly.
22852 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22855 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22858 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22859 new call patterns that clobber %fr12.
22860 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22861 split and postreload patterns.
22862 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22863 registers %fr12 and %fr12R as call used.
22865 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22867 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22868 (canon_address, record_store, replace_read, check_mem_read_rtx,
22869 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22870 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22871 rest_of_handle_dse): Likewise.
22873 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22875 PR middle-end/56524
22876 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22878 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22879 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22880 (save_optabs_if_changed): Replace with...
22881 (init_tree_optimization_optabs): ...this.
22882 * optabs.c (save_optabs_if_changed): Rename to...
22883 (init_tree_optimization_optabs): ...this. Take the optimization node
22884 as argument. Do nothing if the base optabs are already correct.
22885 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22886 to recompute optabs.
22887 * function.h (function): Remove optabs field.
22888 * function.c (invoke_set_current_function_hook): Call
22889 init_tree_optimization_optabs. Use the result to initialize
22892 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
22894 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22895 if GTMA_HAS_NO_INSTRUMENTATION.
22896 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22897 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22898 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22899 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22900 GTMA_HAS_NO_INSTRUMENTATION.
22902 2013-03-08 Jakub Jelinek <jakub@redhat.com>
22904 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22907 2013-03-08 Marek Polacek <polacek@redhat.com>
22908 Jakub Jelinek <jakub@redhat.com>
22910 PR tree-optimization/56478
22911 * predict.c (is_comparison_with_loop_invariant_p): Change the
22912 type of loop_step to tree.
22913 (predict_loops): Adjust.
22914 (predict_iv_comparison): Perform the computations on double_ints.
22916 2013-03-08 Richard Biener <rguenther@suse.de>
22918 PR tree-optimization/56570
22919 * tree-cfg.c (verify_expr_location_1): Verify locations for
22921 * tree-sra.c (create_access_replacement): Strip locations
22922 from DECL_DEBUG_EXPRs.
22924 2013-03-08 Richard Biener <rguenther@suse.de>
22926 * tree-inline.c (expand_call_inline): Do not associate
22927 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22928 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22930 2013-03-08 Richard Biener <rguenther@suse.de>
22932 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22933 or block changes with -Og. Fix for location / block encoding
22934 changes and PHI arguments with locations.
22936 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
22938 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
22940 (struct output_info): Likewise.
22941 (register_overhead): Remove bad gcc_assert.
22942 (bitmap_find_bit): If there is only a single bitmap element, do not
22943 count a miss as a search.
22944 (print_statistics): Update for counter type changes.
22945 (dump_bitmap_statistics): Likewise. Print headers such that they
22946 are properly lined up with the printed counters.
22948 2013-03-07 Jakub Jelinek <jakub@redhat.com>
22950 PR tree-optimization/56559
22951 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
22952 check that it has only a single use.
22954 2013-03-07 Richard Biener <rguenther@suse.de>
22956 * doc/invoke.texi (fwhole-program): Discourage use in combination
22959 2013-03-06 Jakub Jelinek <jakub@redhat.com>
22961 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
22963 PR tree-optimization/56539
22964 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
22965 instead of GSI_CONTINUE_LINKING as last argument to
22966 force_gimple_operand_gsi. Adjust function comment.
22968 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
22971 PR middle-end/56548
22972 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
22973 promoted mode, convert the result back to the original mode.
22975 2013-03-06 Richard Biener <rguenther@suse.de>
22977 PR middle-end/56294
22978 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
22979 (insert_updated_phi_nodes_compare_uids): New function.
22980 (update_ssa): Sort symbols_to_rename after UID before
22981 traversing it to insert PHI nodes.
22983 2013-03-06 Richard Biener <rguenther@suse.de>
22985 PR middle-end/50494
22986 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
22987 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
22990 2013-02-13 Richard Biener <rguenther@suse.de>
22993 * varasm.c (output_constant_def_1): Get the decl representing
22994 the constant as argument.
22995 (output_constant_def): Wrap output_constant_def_1.
22996 (make_decl_rtl): Use output_constant_def_1 with the decl
22997 representing the constant.
22998 (build_constant_desc): Optionally re-use a decl already
22999 representing the constant.
23000 (tree_output_constant_def): Adjust.
23002 2013-03-06 Joey Ye <joey.ye@arm.com>
23005 * gcc.c (convert_white_space): New function.
23006 (main): Handles white space in function name.
23008 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23011 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23012 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23013 to SH_DIV_CALL_TABLE for TARGET_SH2.
23014 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23016 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23017 call-table options.
23019 2013-03-05 Sterling Augustine <saugustine@google.com>
23020 Cary Coutant <ccoutant@google.com>
23023 * dwarf2out.c (resolve_addr): Don't call
23024 remove_loc_list_addr_table_entries a second time for the same
23027 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23030 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23031 (avoid_complex_debug_insns): New function.
23032 (expand_debug_locations): Call it.
23034 PR rtl-optimization/56484
23035 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23036 lifetimes of hard registers on small register class machines.
23038 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23040 * config/microblaze/microblaze-protos.h: Rename
23041 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23042 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23044 (microblaze_fast_interrupt_function_p): New function.
23045 (microblaze_is_interrupt_handler): Rename to
23046 microblaze_is_interrupt_variant and add fast_interrupt check.
23047 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23048 (save_restore_insns): Likewise.
23049 (compute_frame_size): Likewise.
23050 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23051 (microblaze_globalize_label): Likewise.
23052 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23053 * config/microblaze/microblaze.md: Use wrapper
23054 microblaze_is_interrupt_variant.
23056 2013-03-05 Kai Tietz <ktietz@redhat.com>
23058 * sdbout.c (sdbout_one_type): Switch to current function's section
23059 supporting cold/hot.
23061 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23063 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23066 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23068 PR middle-end/56461
23069 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23070 if VALGRIND_GET_VBITS is defined, temporarily make object
23071 memory all defined, and restore previous valgrind addressability
23072 and definability afterwards. Free this_object at the end.
23074 PR middle-end/56461
23075 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23076 right before calling lra_create_live_ranges, also call it
23077 when clearing live_p. Only call lra_clear_live_ranges
23078 at the end if live_p.
23080 PR middle-end/56461
23081 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23083 2013-03-05 Richard Biener <rguenther@suse.de>
23085 PR tree-optimization/56521
23086 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23089 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23092 * except.h (remove_unreachable_eh_regions): New prototype.
23093 * except.c (remove_eh_handler_splicer): New function, split out
23094 of remove_eh_handler.
23095 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23096 warning about running it on many EH regions one at a time.
23097 (remove_unreachable_eh_regions_worker): New function, walk the
23098 EH tree in depth-first order and remove non-marked regions.
23099 (remove_unreachable_eh_regions): New function.
23100 * tree-eh.c (mark_reachable_handlers): New function, split out
23101 from remove_unreachable_handlers.
23102 (remove_unreachable_handlers): Use mark_reachable_handlers and
23103 remove_unreachable_eh_regions.
23104 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23105 and remove_unreachable_eh_regions.
23107 2013-03-05 Richard Biener <rguenther@suse.de>
23109 PR middle-end/56525
23110 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23111 not freeing them until the end.
23113 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23115 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23117 2013-03-05 Richard Biener <rguenther@suse.de>
23119 PR tree-optimization/56270
23120 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23121 of loads after scheduling an SLP instance.
23123 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23125 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23127 (check_gcc_parallelize): Run guality.exp as a separate job from
23128 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23129 struct-layout-1.exp with stackalign.exp.
23131 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23133 PR middle-end/56461
23134 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23135 load_index sbitmap even if some bit in it isn't set.
23137 PR middle-end/56461
23138 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23139 (discover_iteration_bound_by_body_walk): Change queues to
23140 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23141 spelling in comment. Call safe_push on queues[bound_index] directly.
23142 Release queues[queue_index] in every iteration unconditionally.
23143 Release bounds vector.
23145 PR middle-end/56461
23146 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23147 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23148 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23151 2013-03-05 Richard Biener <rguenther@suse.de>
23154 * tree-inline.c (remap_blocks_to_null): New function.
23155 (expand_call_inline): When expanding a call stmt without
23156 an associated block inline remap all callee blocks to NULL.
23158 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23160 PR rtl-optimization/56494
23161 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23162 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23163 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23165 PR middle-end/56461
23166 * sel-sched-ir.c (free_sched_pools): Release
23167 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23168 if succs_info_pool.max_top isn't -1.
23171 * opts.c (opts_obstack, opts_concat): Moved to...
23172 * opts-common.c (opts_obstack, opts_concat): ... here.
23174 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23176 PR middle-end/56461
23177 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23179 2013-03-04 Martin Jambor <mjambor@suse.cz>
23181 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23182 all appropriate places.
23184 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23186 PR tree-optimization/56424
23187 * ipa-split.c (split_function): Do not set the RSO flag if result is
23188 not by reference and its type is a register type.
23190 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23192 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23193 (microblaze_legitimate_pic_operand): Likewise
23194 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23195 new function microblaze_legitimate_pic_operand
23196 * config/microblaze/microblaze-protos.h
23197 (microblaze_legitimate_pic_operand): Declare.
23199 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23201 * config/microblaze/predicates.md (call_insn_simple_operand):
23202 New predicate for supported rtx code types.
23203 * config/microblaze/microblaze.md (call_internal1): Use
23204 call_insn_simple_operand predicate.
23206 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23208 PR middle-end/56461
23209 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23210 partitions.ordered_remove.
23212 PR middle-end/56461
23213 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23214 vec_oprnds0.create (1) for modifier == NONE.
23216 PR middle-end/56461
23217 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23218 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23219 vec_oprnds1 right before pushing anything to it for
23222 PR middle-end/56461
23223 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23224 set nbbs to 0 instead of having separate code path.
23225 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23226 instead of false as last argument if returning NULL.
23228 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23230 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23231 the attribute is now called "target" instead of "option".
23232 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23233 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23234 attribute/pragma name for TARGET_OPTION_VALID_P and
23235 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23236 * doc/tm.texi: Regenerated.
23238 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23240 * config/microblaze/microblaze.c:
23241 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23242 * config/microblaze/microblaze.h: Add -mxl-reorder to
23244 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23245 instructions emitted if TARGET_REORDER.
23246 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23247 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23250 2013-03-01 Xinliang David Li <davidxl@google.com>
23252 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23255 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23257 PR middle-end/56461
23258 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23259 vector even when returning true. Fix up function comment formatting.
23261 PR middle-end/56461
23262 * ira-build.c (ira_loop_nodes_count): New variable.
23263 (create_loop_tree_nodes): Initialize it.
23264 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23266 PR middle-end/56461
23267 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23268 method on dr_chain and result_chain.
23269 * tree-vect-stmts.c (vectorizable_store): Only call
23270 result_chain.create if j == 0.
23272 PR middle-end/56461
23273 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23274 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23275 before overwriting it.
23277 2013-03-01 Tobias Burnus <burnus@net-b.de>
23279 * doc/extended.texi (C Extensions): Change order in @menu
23281 (Other MIPS Built-in Functions): Move last MIPS entry before
23282 "picoChip Built-in Functions".
23283 (SH Built-in Functions): Move after RX Built-in Functions.
23284 * doc/gcc.texi (Introduction): Change order in @menu
23286 * doc/md.texi (Constraints): Ditto.
23287 * gty.texi (Type Information): Ditto.
23288 (User-provided marking routines for template types): Make
23290 * doc/invoke.texi (AArch64 Options): Move before
23291 "Adapteva Epiphany Options".
23293 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23294 Jakub Jelinek <jakub@redhat.com>
23297 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23298 no_address_safety_analysis attribute.
23299 * doc/extend.texi (no_address_safety_attribute): Rename to
23300 no_sanitize_address attribute, mention no_address_safety_analysis
23301 attribute as deprecated alias.
23303 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23305 PR middle-end/56461
23306 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23307 type to vec<vec<tree> > *.
23308 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23309 to be vec<tree> instead of vec<tree> *, set vec_defs
23310 to vNULL and call vec_defs.create (number_of_vects), adjust other
23312 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23313 vectorizable_condition): Adjust vect_get_slp_defs callers.
23315 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23317 * config/aarch64/aarch64.c
23318 (aarch64_float_const_representable): Remove unused variable.
23320 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23322 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23324 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23326 * config/aarch64/aarch64-builtins.c
23327 (aarch64_init_simd_builtins): Make static.
23329 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23331 * config/aarch64/aarch64.c
23332 (aarch64_simd_make_constant): Make static.
23334 2013-02-28 Martin Jambor <mjambor@suse.cz>
23336 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23337 with no initialization to the RHS of debug statements.
23339 2013-02-28 Martin Jambor <mjambor@suse.cz>
23341 PR tree-optimization/56294
23342 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23344 (get_access_replacement): Do not call create_access_replacement.
23345 Assert a replacement exists.
23346 (get_repl_default_def_ssa_name): Create the replacement declaration
23349 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23351 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23352 final_end_function.
23354 2013-02-28 Marek Polacek <polacek@redhat.com>
23356 PR rtl-optimization/56466
23357 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23358 if we're changing a loop.
23359 (peel_loops_completely): Likewise.
23361 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23364 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23366 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23369 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23370 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23371 INTX_FTYPE_FX, FX_FTYPE_INTX.
23372 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23374 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23376 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23377 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23378 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23379 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23380 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23381 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23382 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23383 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23384 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23385 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23386 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23387 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23388 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23389 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23390 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23391 (avrxmega6): Increase max flash segments from 5 to 6.
23392 * config/avr/t-multilib: Regenerate.
23393 * config/avr/avr-tables.opt: Regenerate.
23394 * doc/avr-mmcu.texi: Regenerate.
23396 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23398 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23399 (avr_device_to_arch): Rename to avr_device_to_ld.
23400 (avr_device_to_as): New prototype.
23401 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23402 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23403 * config/avr/driver-avr.c (avr_device_to_as): New.
23404 (avr_device_to_arch): Rename to avr_device_to_ld.
23406 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23408 PR middle-end/56461
23409 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23410 method on dr_chain and result_chain.
23412 PR middle-end/56461
23413 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23414 pointer_set_destroy on not_executed_last_iteration.
23416 PR middle-end/56461
23417 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23419 PR middle-end/56461
23420 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23421 FOR_EACH_DEFINED_FUNCTION when freeing state.
23423 PR middle-end/56461
23424 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23426 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23429 PR middle-end/56461
23430 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23431 known_aggs[i].items and release known_aggs vector.
23433 PR middle-end/56461
23434 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23436 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23438 * config/microblaze/microblaze.c (microblaze_emit_compare):
23439 Use xor for EQ/NE comparisions.
23440 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23441 (cbranchsf4): Adjust operator to comparison_operator.
23443 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23445 PR middle-end/56461
23446 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23448 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23449 vec_safe_push, always update *slot.
23450 (redirect_edge_var_map_clear): Use vec_free.
23451 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23452 (free_var_map_entry): Use vec_free.
23453 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23454 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23456 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23458 PR middle-end/45472
23459 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23460 when the may_trap_p bit of the exprs being merged differs.
23461 Reorder tests for speculativeness in the logical and operator.
23463 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23465 * incpath.c (add_standard_paths): Use reconcat instead of concat
23466 where appropriate and avoid leaking memory.
23468 * opts.h: Include obstack.h.
23469 (opts_concat): New prototype.
23470 (opts_obstack): New declaration.
23471 * opts.c (opts_concat): New function.
23472 (opts_obstack): New variable.
23473 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23474 (finish_options): Use opts_concat instead of concat
23475 and XOBNEWVEC instead of XNEWVEC.
23476 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23477 generate_option): Likewise.
23478 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23479 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23482 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23483 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23485 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23487 PR middle-end/56461
23488 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23490 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23492 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23493 (arm_block_move_unaligned_straight): Likewise.
23494 (arm_adjust_block_mem): Likewise.
23496 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23499 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23500 temp, cond and label.
23501 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23504 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23505 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23508 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23509 of prologue/epilogue functions.
23512 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23513 Remove unused variable cfa_offset.
23514 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23517 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23518 type promotion to unsigned.
23521 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23522 for HOST_WIDE_INT of 32 bit / same size as int.
23523 (arm_block_move_unaligned_straight): Likewise.
23524 (arm_adjust_block_mem): Likewise.
23527 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23530 2013-02-26 Marek Polacek <polacek@redhat.com>
23532 PR tree-optimization/56426
23533 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23535 2013-02-26 Richard Biener <rguenther@suse.de>
23538 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23539 unused variable loops.
23541 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23543 PR tree-optimization/56448
23544 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23545 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23546 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23547 later operands of the references, or even first operand for
23548 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23550 PR tree-optimization/56443
23551 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23552 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23553 to type_for_mode langhook.
23555 2013-02-25 Matt Turner <mattst88@gmail.com>
23557 * doc/invoke.texi: Document r4700.
23559 2013-02-25 Richard Biener <rguenther@suse.de>
23561 PR tree-optimization/56175
23562 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23564 (simplify_bitwise_binary): ... here. Also guard the conversion
23565 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23567 2013-02-25 Catherine Moore <clm@codesourcery.com>
23570 2013-02-24 Catherine Moore <clm@codesourcery.com>
23571 Maciej W. Rozycki <macro@codesourcery.com>
23572 Tom de Vries <tom@codesourcery.com>
23573 Nathan Sidwell <nathan@codesourcery.com>
23574 Iain Sandoe <iain@codesourcery.com>
23575 Nathan Froyd <froydnj@codesourcery.com>
23576 Chao-ying Fu <fu@mips.com>
23578 * doc/extend.texi: (micromips, nomicromips, nocompression):
23579 Document new function attributes.
23580 * doc/invoke.texi (minterlink-compressed, mmicromips,
23581 m14k, m14ke, m14kec): Document new options.
23582 (minterlink-mips16): Update documentation.
23583 * doc/md.texi (ZC, ZD): Document new constraints.
23584 * configure.ac (gcc_cv_as_micromips): Check if linker
23585 supports the .set micromips directive.
23586 * configure: Regenerate.
23587 * config.in: Regenerate.
23588 * config/mips/mips-tables.opt: Regenerate.
23589 * config/mips/micromips.md: New file.
23590 * constraints.md (ZC, AD): New constraints.
23591 * config/mips/predicates.md (movep_src_register): New predicate.
23592 (movep_src_operand): New predicate.
23593 (non_volatile_mem_operand): New predicate.
23594 * config/mips/mips.md (multimem): New type.
23595 (length): Differentiate between 17-bit and 18-bit branch offsets.
23596 (MOVEP1, MOVEP2): New mode iterator.
23597 (mov_<load>l): Use ZC constraint.
23598 (mov_<load>r): Likewise.
23599 (mov_<store>l): Likewise.
23600 (mov_<store>r): Likewise.
23601 (*branch_equality<mode>_inverted): Add microMIPS support.
23602 (*branch_equality<mode>): Likewise.
23603 (*jump_absolute): Likewise.
23604 (indirect_jump_<mode>): Likewise.
23605 (tablejump_<mode>): Likewise.
23606 (<optab>_internal): Likewise.
23607 (sibcall_internal): Likewise.
23608 (sibcall_value_internal): Likewise.
23609 (prefetch): Use constraint ZD.
23610 * config/mips/mips.opt (minterlink-compressed): New option.
23611 (minterlink-mips16): Now an alias for minterlink-compressed.
23612 (mmicromips): New option.
23613 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23614 (compare_and_swap_12): Likewise.
23615 (sync_add<mode>): Likewise.
23616 (sync_<optab>_12): Likewise.
23617 (sync_old_<optab>_12): Likewise.
23618 (sync_new_<optab>_12): Likewise.
23619 (sync_nand_12): Likewise.
23620 (sync_old_nand_12): Likewise.
23621 (sync_new_nand_12): Likewise.
23622 (sync_sub<mode>): Likewise.
23623 (sync_old_add<mode>): Likewise.
23624 (sync_old_sub<mode>): Likewise.
23625 (sync_new_add<mode>): Likewise.
23626 (sync_new_sub<mode>): Likewise.
23627 (sync_<optab><mode>): Likewise.
23628 (sync_old_<optab><mode>): Likewise.
23629 (sync_new_<optab><mode>): Likewise.
23630 (sync_nand<mode>): Likewise.
23631 (sync_old_nand<mode>): Likewise.
23632 (sync_new_nand<mode>): Likewise.
23633 (sync_lock_test_and_set<mode>): Likewise.
23634 (test_and_set_12): Likewise.
23635 (atomic_compare_and_swap<mode>): Likewise.
23636 (atomic_exchange<mode>_llsc): Likewise.
23637 (atomic_fetch_add<mode>_llsc): Likewise.
23638 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23639 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23640 (umips_save_restore_pattern_p): Likewise.
23641 (umips_load_store_pair_p): Likewise.
23642 (umips_output_load_store_pair): Likewise.
23643 (umips_movep_target_p): Likewise.
23644 (umips_12bit_offset_address_p): Likewise.
23645 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23646 (mips_base_mips16): Rename this...
23647 (mips_base_compression_flags): ...to this. Update all uses.
23648 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23649 (mips_mips16_decl_p): Delete.
23650 (mips_nomips16_decl_p): Delete.
23651 (mips_get_compress_on_flags): New function.
23652 (mips_get_compress_off_flags): New function.
23653 (mips_get_compress_mode): New function.
23654 (mips_get_compress_on_name): New function.
23655 (mips_get_compress_off_name): New function.
23656 (mips_insert_attributes): Support multiple compression types.
23657 (mips_merge_decl_attributes): Likewise.
23658 (umips_12bit_offset_address_p): New function.
23659 (mips_start_function_definition): Emit .set micromips directive.
23660 (mips_call_may_need_jalx_p): New function.
23661 (mips_function_ok_for_sibcall): Add microMIPS support.
23662 (mips_print_operand_punctuation): Support short delay slots and
23664 (umips_swm_mask, umips_swm_encoding): New.
23665 (umips_build_save_restore): New function.
23666 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23667 (was_mips16_p): Remove.
23668 (old_compression_mode): New.
23669 (mips_set_compression_mode): New function.
23670 (mips_set_current_function): Add microMIPS support.
23671 (mips_option_override): Likewise.
23672 (umips_save_restore_pattern_p): New function.
23673 (umips_output_save_restore): New function.
23674 (umips_load_store_pair_p_1): New function.
23675 (umips_load_store_pair_p): New function.
23676 (umips_output_load_store_pair_1): New function.
23677 (umips_output_load_store_pair): New function.
23678 (umips_movep_target_p) New function.
23679 (mips_prepare_pch_save): Add microMIPS support.
23680 * config/mips/mips.h (TARGET_COMPRESSION): New.
23681 (TARGET_CPU_CPP_BUILTINS): Update macro
23682 to use new compression flags and to support microMIPS.
23683 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23684 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23685 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23686 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23687 (ASM_SPEC): Support mmicromips and mno-micromips.
23688 (M16STORE_REG_P): New macro.
23689 (MIPS_CALL): Support TARGET_MICROMIPS.
23690 (MICROMIPS_J): New macro.
23691 (mips_base_mips16): Rename this...
23692 (mips_base_compression_flags): ...to this.
23693 (UMIPS_12BIT_OFFSET_P): New macro.
23694 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23695 (MULTILIB_DIRNAMES): Likewise.
23697 2013-02-25 Tom de Vries <tom@codesourcery.com>
23699 PR rtl-optimization/56131
23700 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23701 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23702 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23704 2013-02-25 Tobias Burnus <burnus@net-b.de>
23706 * doc/invoke.texi (-fsanitize=): Move from optimization
23707 to debugging options.
23709 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23711 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23713 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23714 Alexander Monakov <amonakov@ispras.ru>
23716 PR middle-end/56077
23717 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23718 flush pending lists also on non-jumps. Adjust comment.
23720 2013-02-24 Catherine Moore <clm@codesourcery.com>
23721 Maciej W. Rozycki <macro@codesourcery.com>
23722 Tom de Vries <tom@codesourcery.com>
23723 Nathan Sidwell <nathan@codesourcery.com>
23724 Iain Sandoe <iain@codesourcery.com>
23725 Nathan Froyd <froydnj@codesourcery.com>
23726 Chao-ying Fu <fu@mips.com>
23728 * doc/extend.texi: (micromips, nomicromips, nocompression):
23729 Document new function attributes.
23730 * doc/invoke.texi (minterlink-compressed, mmicromips,
23731 m14k, m14ke, m14kec): Document new options.
23732 (minterlink-mips16): Update documentation.
23733 * doc/md.texi (ZC, ZD): Document new constraints.
23734 * configure.ac (gcc_cv_as_micromips): Check if linker
23735 supports the .set micromips directive.
23736 * configure: Regenerate.
23737 * config.in: Regenerate.
23738 * config/mips/mips-tables.opt: Regenerate.
23739 * config/mips/micromips.md: New file.
23740 * constraints.md (ZC, AD): New constraints.
23741 * config/mips/predicates.md (movep_src_register): New predicate.
23742 (movep_src_operand): New predicate.
23743 (non_volatile_mem_operand): New predicate.
23744 * config/mips/mips.md (multimem): New type.
23745 (length): Differentiate between 17-bit and 18-bit branch offsets.
23746 (MOVEP1, MOVEP2): New mode iterator.
23747 (mov_<load>l): Use ZC constraint.
23748 (mov_<load>r): Likewise.
23749 (mov_<store>l): Likewise.
23750 (mov_<store>r): Likewise.
23751 (*branch_equality<mode>_inverted): Add microMIPS support.
23752 (*branch_equality<mode>): Likewise.
23753 (*jump_absolute): Likewise.
23754 (indirect_jump_<mode>): Likewise.
23755 (tablejump_<mode>): Likewise.
23756 (<optab>_internal): Likewise.
23757 (sibcall_internal): Likewise.
23758 (sibcall_value_internal): Likewise.
23759 (prefetch): Use constraint ZD.
23760 * config/mips/mips.opt (minterlink-compressed): New option.
23761 (minterlink-mips16): Now an alias for minterlink-compressed.
23762 (mmicromips): New option.
23763 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23764 (compare_and_swap_12): Likewise.
23765 (sync_add<mode>): Likewise.
23766 (sync_<optab>_12): Likewise.
23767 (sync_old_<optab>_12): Likewise.
23768 (sync_new_<optab>_12): Likewise.
23769 (sync_nand_12): Likewise.
23770 (sync_old_nand_12): Likewise.
23771 (sync_new_nand_12): Likewise.
23772 (sync_sub<mode>): Likewise.
23773 (sync_old_add<mode>): Likewise.
23774 (sync_old_sub<mode>): Likewise.
23775 (sync_new_add<mode>): Likewise.
23776 (sync_new_sub<mode>): Likewise.
23777 (sync_<optab><mode>): Likewise.
23778 (sync_old_<optab><mode>): Likewise.
23779 (sync_new_<optab><mode>): Likewise.
23780 (sync_nand<mode>): Likewise.
23781 (sync_old_nand<mode>): Likewise.
23782 (sync_new_nand<mode>): Likewise.
23783 (sync_lock_test_and_set<mode>): Likewise.
23784 (test_and_set_12): Likewise.
23785 (atomic_compare_and_swap<mode>): Likewise.
23786 (atomic_exchange<mode>_llsc): Likewise.
23787 (atomic_fetch_add<mode>_llsc): Likewise.
23788 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23789 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23790 (umips_save_restore_pattern_p): Likewise.
23791 (umips_load_store_pair_p): Likewise.
23792 (umips_output_load_store_pair): Likewise.
23793 (umips_movep_target_p): Likewise.
23794 (umips_12bit_offset_address_p): Likewise.
23795 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23796 (mips_base_mips16): Rename this...
23797 (mips_base_compression_flags): ...to this. Update all uses.
23798 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23799 (mips_mips16_decl_p): Delete.
23800 (mips_nomips16_decl_p): Delete.
23801 (mips_get_compress_on_flags): New function.
23802 (mips_get_compress_off_flags): New function.
23803 (mips_get_compress_mode): New function.
23804 (mips_get_compress_on_name): New function.
23805 (mips_get_compress_off_name): New function.
23806 (mips_insert_attributes): Support multiple compression types.
23807 (mips_merge_decl_attributes): Likewise.
23808 (umips_12bit_offset_address_p): New function.
23809 (mips_start_function_definition): Emit .set micromips directive.
23810 (mips_call_may_need_jalx_p): New function.
23811 (mips_function_ok_for_sibcall): Add microMIPS support.
23812 (mips_print_operand_punctuation): Support short delay slots and
23814 (umips_swm_mask, umips_swm_encoding): New.
23815 (umips_build_save_restore): New function.
23816 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23817 (was_mips16_p): Remove.
23818 (old_compression_mode): New.
23819 (mips_set_compression_mode): New function.
23820 (mips_set_current_function): Add microMIPS support.
23821 (mips_option_override): Likewise.
23822 (umips_save_restore_pattern_p): New function.
23823 (umips_output_save_restore): New function.
23824 (umips_load_store_pair_p_1): New function.
23825 (umips_load_store_pair_p): New function.
23826 (umips_output_load_store_pair_1): New function.
23827 (umips_output_load_store_pair): New function.
23828 (umips_movep_target_p) New function.
23829 (mips_prepare_pch_save): Add microMIPS support.
23830 * config/mips/mips.h (TARGET_COMPRESSION): New.
23831 (TARGET_CPU_CPP_BUILTINS): Update macro
23832 to use new compression flags and to support microMIPS.
23833 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23834 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23835 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23836 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23837 (ASM_SPEC): Support mmicromips and mno-micromips.
23838 (M16STORE_REG_P): New macro.
23839 (MIPS_CALL): Support TARGET_MICROMIPS.
23840 (MICROMIPS_J): New macro.
23841 (mips_base_mips16): Rename this...
23842 (mips_base_compression_flags): ...to this.
23843 (UMIPS_12BIT_OFFSET_P): New macro.
23844 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23845 (MULTILIB_DIRNAMES): Likewise.
23847 2013-02-24 Jakub Jelinek <jakub@redhat.com>
23850 * target-globals.c (save_target_globals): For init_reg_sets and
23851 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23853 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23855 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23856 * config/aarch64/t-aarch64
23857 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23859 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23861 PR inline-asm/56148
23862 * lra-constraints.c (process_alt_operands): Reload operand
23863 conflicting with earlier clobber only if no more other conflicting
23866 2013-02-22 Jakub Jelinek <jakub@redhat.com>
23869 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23870 if not linking a shared library.
23872 2013-02-22 Seth LaForge <sethml@google.com>
23874 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23876 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23878 * config/arm/arm.md (split for extendsidi): Update condition.
23879 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23880 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23881 (qhs_zextenddi_cstr): Likewise.
23883 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23885 PR middle-end/56420
23886 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23887 avoid signed wrapping.
23888 (expand_mult): Handle properly multiplication by
23889 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23890 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23891 in the compiler if coeff is HOST_WIDE_INT_MIN.
23892 (expand_divmod): Don't make ext_op1 static, change it's type to
23893 uhwi. Avoid undefined behavior in -INTVAL (op1).
23895 PR rtl-optimization/50339
23896 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23898 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23899 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23900 into splitting_ashiftrt field.
23901 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23903 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23906 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
23908 PR middle-end/56108
23909 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23910 are sure to go irrevocable.
23912 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
23914 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23915 scalars are valid operands.
23917 2013-02-21 Martin Jambor <mjambor@suse.cz>
23919 PR tree-optimization/56310
23920 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23921 only matching indices and non-negative final offsets.
23922 (intersect_aggregates_with_edge): Pass src_idx to
23923 agg_replacements_to_vector. Pass src_idx insstead of index to
23924 intersect_with_agg_replacements.
23926 2013-02-21 Martin Jambor <mjambor@suse.cz>
23928 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23929 instead of hard-wired defaults.
23931 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23933 * doc/invoke.texi (MIPS Options): Update documentation of the
23934 floating-point multiply-accumulate instruction restrictions.
23936 2013-02-21 Kostya Serebryany <kcc@google.com>
23938 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
23939 asan_shadow_offset on x86_64 linux.
23941 2013-02-21 Richard Biener <rguenther@suse.de>
23943 PR tree-optimization/56415
23945 2013-02-11 Richard Biener <rguenther@suse.de>
23947 PR tree-optimization/56273
23948 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23951 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23954 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
23957 PR inline-asm/56405
23958 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
23959 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
23961 2013-02-20 Jan Hubicka <jh@suse.cz>
23963 PR tree-optimization/56265
23964 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
23965 when target is referenced for first time.
23967 2013-02-20 Richard Biener <rguenther@suse.de>
23969 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
23970 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
23971 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
23972 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
23973 not return anything.
23974 (rename_ssa_copies): Do not remove unused locals.
23975 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
23976 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
23977 * passes.c (execute_function_todo): Do not schedule unused locals
23978 removal if cleanup_tree_cfg did something.
23979 * tree-ssa-live.c (remove_unused_locals): Dump statistics
23980 about the number of removed locals.
23982 2013-02-20 Richard Biener <rguenther@suse.de>
23984 PR tree-optimization/56398
23985 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
23987 2013-02-20 Martin Jambor <mjambor@suse.cz>
23989 PR tree-optimization/55334
23990 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
23991 restricted pointers to arrays.
23993 2013-02-20 Richard Biener <rguenther@suse.de>
23994 Jakub Jelinek <jakub@redhat.com>
23996 PR tree-optimization/56396
23997 * tree-ssa-ccp.c (n_const_val): New static variable.
23998 (get_value): Return NULL for SSA names we don't have a lattice
24000 (ccp_initialize): Initialize n_const_val.
24001 * tree-ssa-copy.c (n_copy_of): New static variable.
24002 (init_copy_prop): Initialize n_copy_of.
24003 (get_value): Return NULL_TREE for SSA names we don't have a
24006 2013-02-20 Martin Jambor <mjambor@suse.cz>
24008 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24010 2013-02-20 Richard Biener <rguenther@suse.de>
24012 * genpreds.c (write_lookup_constraint): Do not compare first
24013 letter of the constraint again.
24015 2013-02-20 Richard Biener <rguenther@suse.de>
24017 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24019 (get_use_iv_cost): Terminate hashtable walk when coming across
24022 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24024 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24025 reassociation for avx2 targets.
24027 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24029 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24030 Add version check for v8.10.a to enable microblaze_has_clz
24031 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24032 version and TARGET_PATTERN_COMPARE check
24033 * config/microblaze/microblaze.md: New clzsi2 instruction
24035 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24037 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24038 function before branching.
24040 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24042 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24044 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24046 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24048 PR middle-end/55889
24049 * sel-sched.c: Include ira.h.
24050 (implicit_clobber_conflict_p): New function.
24051 (moveup_expr): Use it.
24052 * Makefile.in (sel-sched.o): Depend on ira.h.
24054 2013-02-19 Richard Biener <rguenther@suse.de>
24056 PR tree-optimization/56384
24057 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24058 (vn_hash_type): Split out from ...
24059 (vn_hash_constant_with_type): ... here.
24060 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24061 (vn_phi_eq): Compare types from vn_phi_s structure.
24062 (vn_phi_lookup): Populate vn_phi_s type.
24063 (vn_phi_insert): Likewise.
24065 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24067 PR tree-optimization/56350
24068 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24069 if haven't found reduction or nested cycle operand, rather than
24070 asserting we must find it.
24072 PR tree-optimization/56381
24073 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24076 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24077 Jakub Jelinek <jakub@redhat.com>
24080 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24081 (swap_optab_enable): Same.
24082 (init_all_optabs): Use argument instead of global.
24083 * tree.h (struct tree_optimization_option): New field target_optabs.
24084 * expr.h (init_all_optabs): Add argument to prototype.
24085 (TREE_OPTIMIZATION_OPTABS): New.
24086 (save_optabs_if_changed): Protoize.
24087 * optabs.h: Declare this_fn_optabs.
24088 * optabs.c (save_optabs_if_changed): New.
24089 Declare this_fn_optabs.
24090 (init_optabs): Add argument to init_all_optabs() call.
24091 * function.c (invoke_set_current_function_hook): Handle per
24093 * function.h (struct function): New field optabs.
24094 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24095 optimization_current_node has changed.
24096 * target-globals.h (save_target_globals_default_opts): Protoize.
24097 * target-globals.c (save_target_globals_default_opts): New.
24099 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24102 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24103 registers %fr12 and %fr12R as call used.
24106 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24107 and HImode, require all displacements to be an integer multiple of
24109 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24110 only allow QImode and HImode when reload is in progress and strict is
24111 true. Likewise for symbolic addresses. Use base14_operand to check
24112 displacements in REG+BASE addresses.
24114 2013-02-18 Richard Biener <rguenther@suse.de>
24116 PR tree-optimization/56366
24117 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24118 handle sign-conversion of outer-loop initial induction value.
24120 2013-02-18 Richard Biener <rguenther@suse.de>
24122 PR middle-end/56349
24123 * cfghooks.c (merge_blocks): If we merge a latch into another
24124 block adjust references to it.
24125 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24126 (verify_loop_structure): Verify that a recorded latch is in fact
24129 2013-02-18 Richard Biener <rguenther@suse.de>
24131 PR tree-optimization/56321
24132 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24133 order SSA name release and virtual operand unlinking.
24135 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24137 * config/microblaze/microblaze.md (save_stack_block): Define.
24138 (restore_stack_block): Likewise.
24140 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24142 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24143 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24144 * config/microblaze/microblaze.c (microblaze_option_override):
24145 Bail out early for PIC modes when target does not support PIC.
24147 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24149 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24150 Replace with a microblaze version.
24151 (microblaze_trampoline_init): Adapt for microblaze.
24152 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24155 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24156 Dodji Seketeli <dodji@redhat.com>
24159 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24160 (instrument_mem_region_access): Do not forget to always put
24161 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24162 0) statement, even for cases where either 'base' or 'base + len'
24163 are not instrumented -- because they have been previously
24164 instrumented. Simplify the logic by putting all the statements
24165 instrument 'base + len' inside a sequence, and then insert that
24166 sequence right before the current insertion point. Then, to
24167 instrument 'base + len', just get an iterator on that statement.
24168 And do not forget to update the pointer to iterator the function
24169 received as argument.
24171 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24173 PR rtl-optimization/56348
24174 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24176 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24178 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24179 (clean_graph_dump_file): Pass base to start_graph_dump.
24181 2013-02-14 Richard Henderson <rth@redhat.com>
24184 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24186 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24188 * collect2-aix.h: Define F_LOADONLY.
24190 2013-02-14 Richard Biener <rguenther@suse.de>
24193 * varasm.c (output_constant_def_1): Get the decl representing
24194 the constant as argument.
24195 (output_constant_def): Wrap output_constant_def_1.
24196 (make_decl_rtl): Use output_constant_def_1 with the decl
24197 representing the constant.
24198 (build_constant_desc): Optionally re-use a decl already
24199 representing the constant.
24200 (tree_output_constant_def): Adjust.
24202 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24205 * asan.c (instrument_builtin_call): Really put the length of the
24206 second source argument into src1_len.
24208 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24210 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24211 argument. If it is false, don't create edge from then_bb to
24213 (insert_if_then_before_iter): Pass true to it.
24214 (build_check_stmt): Pass false to it.
24215 (transform_statements): Flush hash table only on extended basic
24216 block boundaries, rather than at the beginning of every bb.
24217 Don't flush hash table on nonfreeing_call_p calls.
24218 * tree-flow.h (nonfreeing_call_p): New prototype.
24219 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24221 2013-02-13 David S. Miller <davem@davemloft.net>
24223 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24225 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24228 * ira.c (max_regno_before_ira): Move from ...
24230 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24233 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24235 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24237 2013-02-13 Richard Biener <rguenther@suse.de>
24240 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24241 globals in MEM_REFs.
24243 2013-02-13 Richard Biener <rguenther@suse.de>
24245 * loop-init.c (loop_optimizer_init): Clear loop state when
24246 re-initializing preserved loops.
24247 * loop-unswitch.c (unswitch_single_loop): Return whether
24248 we unswitched the loop. Do not verify loop state here.
24249 (unswitch_loops): When we unswitched a loop discover new loops.
24251 2013-02-13 Kostya Serebryany <kcc@google.com>
24253 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24255 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24257 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24259 Avoid instrumenting duplicated memory access in the same basic block
24260 * Makefile.in (asan.o): Add new dependency on hash-table.h
24261 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24262 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24263 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24264 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24265 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24266 (get_mem_ref_of_assignment): New functions.
24267 (get_mem_refs_of_builtin_call): Extract from
24268 instrument_builtin_call and tweak a little bit to make it fit with
24270 (instrument_builtin_call): Use the new
24271 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24272 of is_gimple_builtin_call.
24273 (instrument_derefs, instrument_mem_region_access): Insert the
24274 instrumented memory reference into the hash table.
24275 (maybe_instrument_assignment): Renamed instrument_assignment into
24276 this, and change it to advance the iterator when instrumentation
24277 actually happened and return true in that case. This makes it
24278 homogeneous with maybe_instrument_assignment, and thus give a
24279 chance to callers to be more 'regular'.
24280 (transform_statements): Clear the memory reference hash table
24281 whenever we enter a new BB, when we cross a function call, or when
24282 we are done transforming statements. Use
24283 maybe_instrument_assignment instead of instrumentation. No more
24284 need to special case maybe_instrument_assignment and advance the
24285 iterator after calling it; it's now handled just like
24286 maybe_instrument_call. Update comment.
24288 2013-02-13 Richard Biener <rguenther@suse.de>
24290 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24291 Fix loop discovery code.
24293 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24295 PR inline-asm/56148
24296 * lra-constraints.c (process_alt_operands): Match early clobber
24297 operand with itself. Check conflicts with earlyclobber only if
24298 the operand is not reloaded. Prefer to reload conflicting operand
24299 if earlyclobber and matching operands are the same.
24301 2013-02-12 Richard Biener <rguenther@suse.de>
24304 * lto-streamer-out.c (write_symbol): Do not output symbols
24305 for hard register variables.
24307 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24310 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24311 (umulsidi3_insn, mulsidi3_insn): New insns.
24313 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24315 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24316 (struct tune_params): Add vec_costs field.
24317 * config/arm/arm.c (arm_builtin_vectorization_cost)
24318 (arm_add_stmt_cost): New functions.
24319 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24320 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24321 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24322 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24323 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24324 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24325 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24327 2013-02-12 Richard Biener <rguenther@suse.de>
24330 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24331 decls again if possible.
24333 2013-02-12 Richard Biener <rguenther@suse.de>
24335 PR middle-end/56288
24336 * tree-ssa.c (verify_ssa_name): Fix check, move
24337 SSA_NAME_IN_FREE_LIST check up.
24339 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24340 Steven Bosscher <steven@gcc.gnu.org>
24342 PR rtl-optimization/56151
24343 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24344 equal to op0 or op1, and last_insn pattern is CODE operation
24345 with MEM dest and one of the operands matches that MEM.
24347 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24349 * doc/extend.texi: Document Function Multiversioning and "default"
24350 parameter string to target attribute.
24351 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24352 target attribute parameter is "default".
24353 (ix86_compare_version_priority): Remove checks for target attribute.
24354 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24355 Remove check for target attribute equal to NULL. Add assert.
24356 (ix86_generate_version_dispatcher_body): Change error to sorry.
24358 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24359 Jack Howarth <howarth@bromo.med.uc.edu>
24360 Patrick Marlier <patrick.marlier@gmail.com>
24363 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24364 define ENDFILE_SPEC as TM_DESTRUCTOR.
24365 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24367 2013-02-11 Alexander Potapenko <glider@google.com>
24368 Jack Howarth <howarth@bromo.med.uc.edu>
24369 Jakub Jelinek <jakub@redhat.com>
24372 * config/darwin.c (cdtor_record): Rename ctor_record.
24373 (sort_cdtor_records): Rename sort_ctor_records.
24374 (finalize_dtors): New routine to sort destructors by
24375 priority before use in assemble_integer.
24376 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24378 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24380 PR rtl-optimization/56275
24381 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24382 offset is non-negative and less than cmode size before
24383 calling simplify_subreg.
24385 2013-02-11 Richard Biener <rguenther@suse.de>
24387 PR tree-optimization/56264
24388 * cfgloop.h (fix_loop_structure): Adjust prototype.
24389 * loop-init.c (fix_loop_structure): Return the number of
24390 newly discovered loops.
24391 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24392 are discovered, do a full loop-closed SSA rewrite.
24394 2013-02-11 Richard Biener <rguenther@suse.de>
24396 PR tree-optimization/56273
24397 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24399 (check_array_ref): Fix missing newline in dumps.
24400 (search_for_addr_array): Likewise.
24402 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24404 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24406 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24409 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24411 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24413 PR rtl-optimization/56246
24414 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24416 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24417 constraints are satisfied.
24419 2013-02-08 Jeff Law <law@redhat.com>
24422 * emit-rtl.c (reg_is_parm_p): New function.
24423 * regs.h (reg_is_parm_p): New prototype.
24424 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24425 callee-clobbered registers.
24427 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24430 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24431 If there is no implicit builtin declaration, just return NULL.
24433 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24435 * config/i386/sse.md (FMAMODEM): New mode iterator.
24436 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24437 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24439 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24441 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24442 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24443 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24445 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24447 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24448 (microblaze*-*-elf): Likewise.
24449 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24451 * config/microblaze/microblaze-c.c: Add builtin defines for
24452 _LITTLE_ENDIAN and _BIG_ENDIAN.
24453 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24454 add to TARGET_DEFAULT flags.
24455 Expand ASM_SPEC and LINK_SPEC.
24456 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24457 * config/microblaze/microblaze.md: Update extendsidi2 and
24458 movdi_internal instructions to use low-order / high-order reg
24460 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24461 options and inversemask / mask of LITTLE_ENDIAN.
24462 * config/microblaze/t-microblaze: Expand multilib options to
24463 include mlittle-endian (le) and update exceptions patterns.
24465 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24467 PR rtl-optimization/56195
24468 * lra-constraints.c (get_reload_reg): Don't reuse regs
24469 if they have smaller mode than requested, if they have
24470 wider mode than requested, try to return a SUBREG.
24472 PR tree-optimization/56250
24473 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24474 if type is unsigned and code isn't MULT_EXPR.
24476 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24478 PR tree-optimization/56064
24479 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24480 bits according to mode.
24481 * fixed-value.h (fixed_from_double_int)
24482 (const_fixed_from_double_int): Adjust comments.
24484 2013-02-08 Richard Biener <rguenther@suse.de>
24487 * lto-streamer.h (struct data_in): Remove current_file, current_line
24488 and current_col members.
24489 * lto-streamer-out.c (lto_output_location): Stream changed bits
24490 en-block for efficiency.
24491 * lto-streamer-in.c (clear_line_info): Remove.
24492 (lto_input_location): Cache current file, line and column
24493 globally via local statics. Read changed bits en-block.
24494 (input_function): Do not call clear_line_info.
24495 (lto_read_body): Likewise.
24496 (lto_input_toplevel_asms): Likewise.
24498 2013-02-08 Michael Matz <matz@suse.de>
24500 PR tree-optimization/52448
24501 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24502 (nt_call_phase): New static.
24503 (add_or_mark_expr): Only mark accesses with newer phase than any
24505 (nonfreeing_call_p): New.
24506 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24507 (nt_fini_block): Keep blocks marked as visited.
24508 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24510 2013-02-08 Richard Biener <rguenther@suse.de>
24512 * ira.c (ira): Free broken dominator information.
24514 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24516 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24518 2013-02-08 Marek Polacek <polacek@redhat.com>
24520 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24522 2013-02-08 Richard Biener <rguenther@suse.de>
24524 PR middle-end/56181
24525 * cfgloop.h (flow_loops_find): Adjust.
24526 (bb_loop_header_p): Declare.
24527 * cfgloop.c (bb_loop_header_p): New function split out from ...
24528 (flow_loops_find): ... here. Adjust function signature,
24529 support incremental loop structure update.
24530 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24531 * cfgloopmanip.c (fix_loop_structure): Move ...
24532 * loop-init.c (fix_loop_structure): ... here.
24533 (apply_loop_flags): Split out from ...
24534 (loop_optimizer_init): ... here.
24535 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24536 in incremental mode, only remove dead loops here.
24538 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24541 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24542 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24543 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24544 (*round<mode>3.libgcc): New insns for fixed-modes.
24545 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24546 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24547 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24548 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24549 implementations. Define to __builtin_avr_absFX,
24550 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24551 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24552 __builtin_avr_countlsFX, respectively.
24553 * config/avr/avr-c.c (target.h): Include it.
24554 (enum avr_builtin_id): New enum.
24555 (avr_resolve_overloaded_builtin): New static function.
24556 (avr_register_target_pragmas): Use it to set
24557 targetm.resolve_overloaded_builtin.
24558 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24559 tree nodes used by DEF_BUILTIN.
24560 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24561 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24562 <AVR_BUILTIN_xxBITS>: Same.
24564 2013-02-08 Richard Biener <rguenther@suse.de>
24566 * cfgloop.c (verify_loop_structure): Properly handle
24567 a loop exiting to another loop header.
24568 * ira-int.h (ira_loops): Remove.
24569 * ira.c (ira_loops): Remove.
24570 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24571 (do_reload): Use loop_optimizer_finalize.
24572 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24573 number_of_loops to access the loop tree.
24574 (more_one_region_p): Likewise.
24575 (finish_loop_tree_nodes): Likewise.
24576 (rebuild_regno_allocno_maps): Likewise.
24577 (mark_loops_for_removal): Likewise.
24578 (mark_all_loops_for_removal): Likewise.
24579 (remove_unnecessary_regions): Likewise.
24580 (ira_build): Likewise.
24581 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24583 2013-02-08 Richard Biener <rguenther@suse.de>
24585 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24586 * ipa-pure-const.c (analyze_function): Avoid calling
24587 mark_irreducible_loops twice.
24588 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24590 2013-02-07 David S. Miller <davem@davemloft.net>
24592 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24594 * var-tracking.c (vt_add_function_parameter): Test the presence of
24595 HAVE_window_save properly and do not remap argument registers when
24596 we have a leaf function.
24598 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24601 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24603 * config/i386/i386.c (ix86_print_operand): Ditto.
24605 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24607 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24609 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24611 PR rtl-optimization/56225
24612 * lra-constraints.c (process_alt_operands): Check that reload hard
24613 reg can hold value for strict_low_part.
24615 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24618 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24619 dwarf2out_end_function.
24620 (in_first_function_p, maybe_at_text_label_p,
24621 first_loclabel_num_not_at_text_label): New variables.
24622 (dwarf2out_var_location): In the first function find out
24623 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24624 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24627 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24629 PR rtl-optimization/56178
24630 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24631 SUBREG of a register. Tidy up related block of code.
24632 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24633 note if the source is a register or a SUBREG of a register.
24635 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24638 * config/rs6000/rs6000.md (ptrm): New mode attr.
24639 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24640 call_value_indirect_aix<pttrsize>,
24641 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24644 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24646 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24647 if -bnortl. Convert to strcmp and strncmp.
24649 2013-02-07 Alan Modra <amodra@gmail.com>
24652 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24653 addresses won't wrap when offsetting.
24654 (rs6000_secondary_reload): Provide secondary reloads needed for
24655 wrapping LO_SUM addresses.
24657 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24659 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24660 MACH, just __MACH__.
24662 2013-02-06 Richard Biener <rguenther@suse.de>
24664 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24665 instead of calling fix_loop_structure.
24667 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24669 PR middle-end/56217
24670 * omp-low.c (use_pointer_for_field): Return false if
24671 lower_send_shared_vars doesn't generate any copy-out code.
24673 2013-02-06 Tom de Vries <tom@codesourcery.com>
24675 PR rtl-optimization/56131
24676 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24677 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24678 of the label is NULL. Add comment.
24680 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24682 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24685 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24686 (STATIC_LIBTSAN_LIBS): Likewise.
24687 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24688 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24689 is defined, don't add anything else beyond that.
24690 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24691 (LINK_COMMAND_SPEC): Use them.
24693 PR tree-optimization/56205
24694 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24695 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24696 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24698 2013-02-05 Richard Biener <rguenther@suse.de>
24700 PR tree-optimization/53342
24701 PR tree-optimization/53185
24702 * tree-vectorizer.h (vect_check_strided_load): Remove.
24703 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24704 not disallow peeling for vectorized strided loads.
24705 (vect_check_strided_load): Make static and simplify.
24706 (vect_analyze_data_refs): Adjust.
24707 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24708 correctly when vectorizing strided loads.
24710 2013-02-05 Richard Biener <rguenther@suse.de>
24712 * doc/install.texi: Refer to ISL, not PPL.
24714 2013-02-05 Jan Hubicka <jh@suse.cz>
24716 PR tree-optimization/55789
24717 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24719 2013-02-05 Jan Hubicka <jh@suse.cz>
24721 PR tree-optimization/55789
24722 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24723 the dead call anyway.
24725 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24728 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24730 2013-02-04 Alexander Potapenko <glider@google.com>
24731 Jack Howarth <howarth@bromo.med.uc.edu>
24732 Jakub Jelinek <jakub@redhat.com>
24735 * config/darwin.c (sort_ctor_records): Stabilized qsort
24736 on constructor priority by using original position.
24737 (finalize_ctors): New routine to sort constructors by
24738 priority before use in assemble_integer.
24739 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24741 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24744 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24745 about visibility on artificial decls.
24746 * config/sol2.c (solaris_assemble_visibility): Likewise.
24748 2013-02-04 Kai Tietz <ktietz@redhat.com>
24751 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24752 argument and improve checking of return-argument types for 16-byte
24754 (ix86_function_value_1): Add additional valtype argument on call
24755 of function_value_64.
24756 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24757 handling infunction_value_64 function.
24759 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24761 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24763 2013-02-04 Richard Biener <rguenther@suse.de>
24765 PR tree-optimization/56188
24766 * tree-ssa-structalias.c (label_visit): Consider case with
24767 initially non-empty points-to set.
24768 (perform_var_substitution): Dump node mapping and clean up.
24770 2013-02-04 Richard Guenther <rguenther@suse.de>
24773 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24774 node prevail as last resort.
24775 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24776 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24778 2013-02-04 Richard Biener <rguenther@suse.de>
24780 PR tree-optimization/56113
24781 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24783 (equiv_class_lookup_or_add): ... this.
24784 (label_visit): Adjust and fix error in previous patch.
24785 (perform_var_substitution): Adjust.
24787 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24789 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24790 * config/sh/sh4-300.md: Likewise.
24791 * config/sh/sh4a.md: Likewise.
24792 * config/sh/constraints.md: Likewise.
24793 * config/sh/sh.md: Likewise.
24794 * config/sh/netbsd-elf.h: Likewise.
24795 * config/sh/predicates.md: Likewise.
24796 * config/sh/sh-protos.h: Likewise.
24797 * config/sh/ushmedia.h: Likewise.
24798 * config/sh/linux.h: Likewise.
24799 * config/sh/sh.c: Likewise.
24800 * config/sh/superh.h: Likewise.
24801 * config/sh/elf.h: Likewise.
24802 * config/sh/sh4.md: Likewise.
24803 * config/sh/sh.h: Likewise.
24805 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24807 * config/pa/constraints.md: Adjust unused letters. Change "T"
24808 constraint to match_test floating_point_store_memory_operand().
24809 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24810 (base14_operand): New.
24811 (floating_point_store_memory_operand): New.
24812 (integer_store_memory_operand): Revise to use base14_operand and
24813 reg_plus_base_memory_operand.
24814 (move_dest_operand): Allow symbolic_memory_operands.
24815 (symbolic_memory_operand): Check for LO_SOM.
24816 (symbolic_operand): Change default case to break.
24817 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24818 CONST_DOUBLE values to be reloaded by putting them into memory when
24819 the destination is a floating point register.
24820 (movdf): Remove code to handle CONST_DOUBLE.
24822 (reload_indf_r1): New.
24823 (reload_insf_r1): New.
24824 Consistently use "Q" and "T" constraints with integer and floating
24825 point move instructions, respectively.
24826 (movdi): Remove FAIL.
24827 Change predicate for source operand unamed DImode move from
24828 general_operand to move_src_operand.
24829 (umulsidi3): Change predicate for destination operand to
24831 Likewise for similar unamed patterns.
24832 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24833 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24834 (hppa_legitimize_address): Simplify mask calculation.
24835 (pa_emit_move_sequence): Revised handling of secondary reloads from
24836 REG+D addresses for floating point loads and stores. Directly handle
24837 loading CONST0_RTX (mode) to a floating point register.
24838 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24839 to floating point registers. Don't restrict secondary reloads to
24840 floating point registers to integer modes. Revise some comments and
24842 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24843 (pa_legitimate_address_p): New.
24844 (pa_legitimize_reload_address): New.
24845 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24846 (STRICT_REG_OK_FOR_BASE_P): New.
24847 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24848 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24850 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
24851 Andrew Dixie <andrewd@gentrack.com>
24853 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24856 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24858 * expmed.c (extract_bit_field_1): Pass the full width of the
24859 structure to get_best_reg_extraction_insn.
24861 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24864 * configure.ac (use_cxa_atexit): Add AIX.
24865 * configure: Regenerate.
24867 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24869 2013-02-01 Jakub Jelinek <jakub@redhat.com>
24872 * final.c (need_profile_function): New variable.
24873 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24874 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24875 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24876 notes, targetm.asm_out.function_prologue doesn't emit anything,
24877 HAVE_prologue and profiler should be emitted before prologue,
24878 set need_profile_function instead of emitting it.
24879 (final_scan_insn): If need_profile_function, emit
24880 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24881 NOTE_INSN_FUNCTION_BEG note.
24883 2013-02-01 Richard Henderson <rth@redhat.com>
24885 * config/rs6000/rs6000.md (smulditi3): New.
24888 * config/alpha/alpha.md (umulditi3): New.
24890 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24892 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24893 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24895 2013-02-01 Richard Biener <rguenther@suse.de>
24897 PR tree-optimization/56113
24898 * tree-ssa-structalias.c (label_visit): Reduce work for
24899 single-predecessor nodes.
24901 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24903 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24904 range isn't testing for zero.
24906 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24908 PR middle-end/56113
24909 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24911 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24912 Nick Clifton <nickc@redhat.com>
24914 * config/v850/constraints.md (Q): Define as a memory constraint.
24915 * config/v850/predicates.md (label_ref_operand): New predicate.
24916 (e3v5_shift_operand): New predicate.
24917 (ior_operator): New predicate.
24918 * config/v850/t-v850: Add e3v5 multilib.
24919 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24920 (v850_gen_movdi): Prototype.
24921 * config/v850/v850.c: Add support for e3v5 architecture.
24922 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24924 (construct_save_jarl): Add e3v5 long JARL support.
24925 (v850_adjust_insn_length): New function. Adjust length of call
24926 insns when using e3v5 instructions.
24927 (v850_gen_movdi): New function: Generate instructions to move a
24929 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24930 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24931 (TARGET_USE_FPU): Enable for e3v5.
24932 (CONST_OK_FOR_W): New macro.
24933 (ADJUST_INSN_LENGTH): Define.
24934 * config/v850/v850.md (UNSPEC_LOOP): Define.
24935 (attr cpu): Add v850e3v5.
24936 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24937 (movdi): New pattern.
24938 (movdi_internal): New pattern.
24939 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
24940 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
24941 (cstoresf4): Likewise.
24942 (cstoredf4): Likewise.
24943 (insv): New pattern.
24944 (rotlso3_a): New pattern.
24945 (rotlsi3_b): New pattern
24946 (rotlsi3_v850e3v5): New pattern.
24947 (doloop_begin): New pattern.
24948 (fix_loop_counter): New pattern.
24949 (doloop_end): New pattern.
24950 (branch_normal): Add e3v5 long branch support.
24951 (branch_invert): Likewise.
24952 (branch_z_normal): Likewise.
24953 (branch_z_invert): Likewise.
24954 (branch_nz_normal): Likewise.
24955 (branch_nz_invert): Likewise.
24956 (call_internal_short): Add e3v5 register-indirect JARL support.
24957 (call_internal_long): Likewise.
24958 (call_value_internal_short): Likewise.
24959 (call_value_internal_long): Likewise.
24960 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
24961 (mloop): New option.
24962 * config.gcc: Add support for configuring v840e3v5 target.
24963 * doc/invoke.texi: Document new v850 specific command line options.
24965 2013-01-31 Paul Koning <ni1d@arrl.net>
24969 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
24970 children if parent is a class.
24971 (prune_unused_types_prune): Don't add DW_AT_declaration.
24973 2013-01-31 Richard Biener <rguenther@suse.de>
24975 PR tree-optimization/56157
24976 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
24977 match up operand with SLP child.
24979 2013-01-31 Jason Merrill <jason@redhat.com>
24982 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
24983 parameters the first time.
24984 (gen_scheduled_generic_parms_dies): Check completeness here.
24986 2013-01-31 Richard Biener <rguenther@suse.de>
24988 PR middle-end/53073
24989 * common.opt (faggressive-loop-optimizations): New flag,
24990 enabled by default.
24991 * doc/invoke.texi (faggressive-loop-optimizations): Document.
24992 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
24993 infer_loop_bounds_from_undefined by it.
24995 2013-01-31 Richard Biener <rguenther@suse.de>
24997 PR tree-optimization/56150
24998 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
24999 visit virtual operands.
25000 (find_uses_to_rename_bb): Likewise.
25002 2013-01-31 Richard Biener <rguenther@suse.de>
25004 PR tree-optimization/56150
25005 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25006 mixed store non-store stmts.
25008 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25011 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25012 LIBASAN_EARLY_SPEC is defined.
25013 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25014 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25016 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25019 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25020 invalid args instead of ICEing on it.
25021 (ix86_valid_target_attribute_tree): Return error_mark_node if
25022 ix86_valid_target_attribute_inner_p failed.
25023 (ix86_valid_target_attribute_p): Return false only if
25024 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25025 target("default") attribute.
25026 (sorted_attr_string): Change argument from const char * to tree,
25027 merge in all target attribute arguments rather than just one.
25028 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25029 instead of free. Avoid using strcat.
25030 (ix86_mangle_function_version_assembler_name): Mangle
25031 target("default") as if no target attribute is present. Adjust
25032 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25033 instead of xmalloc and XDELETEVEC instead of free.
25034 (ix86_function_versions): Don't return true if one of the decls
25035 doesn't have target attribute. If they don't and one of the decls
25036 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25037 sorted_attr_string caller. Use XDELETEVEC instead of free.
25038 (ix86_supports_function_versions): Remove.
25039 (make_name): Fix up formatting.
25040 (make_dispatcher_decl): Remove resolver_name and its initialization.
25041 Avoid leaking memory.
25042 (is_function_default_version): Return true if there is
25043 target("default") attribute rather than no target attribute at all.
25044 (make_resolver_func): Avoid leaking memory.
25045 (ix86_generate_version_dispatcher_body): Likewise.
25046 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25047 * target.def (supports_function_versions): Remove.
25048 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25049 * doc/tm.texi: Regenerated.
25051 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25053 PR rtl-optimization/56144
25054 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25055 for values with side effects.
25057 2013-01-30 Richard Biener <rguenther@suse.de>
25059 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25060 (sparseset_pop): Likewise.
25061 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25062 to be able to use quick_push in the worker loop.
25064 2013-01-30 Marek Polacek <polacek@redhat.com>
25066 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25068 2013-01-30 Richard Biener <rguenther@suse.de>
25071 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25073 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25075 PR tree-optimization/56064
25076 * fixed-value.c (fixed_from_double_int): New function.
25077 * fixed-value.h (fixed_from_double_int): New prototype.
25078 (const_fixed_from_double_int): New static inline function.
25079 * fold-const.c (native_interpret_fixed): New static function.
25080 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25081 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25082 (native_encode_fixed): New static function.
25083 (native_encode_expr) <FIXED_CST>: Use it.
25084 (native_interpret_int): Move double_int worker code to...
25085 * double-int.c (double_int::from_buffer): ...this new static method.
25086 * double-int.h (double_int::from_buffer): Prototype it.
25088 2013-01-30 Richard Biener <rguenther@suse.de>
25090 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25091 New pointer-map and obstack.
25092 (init_alias_vars): Allocate pointer-map and obstack.
25093 (delete_points_to_sets): Free them.
25094 (find_what_var_points_to): Cache result.
25095 (find_what_p_points_to): Adjust for changed interface of
25096 find_what_var_points_to.
25097 (compute_points_to_sets): Likewise.
25098 (ipa_pta_execute): Likewise.
25100 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25102 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25103 * configure: Regenerate.
25104 * config.in: Regenerate.
25105 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25106 #nobits/#progbits if supported.
25108 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25111 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25112 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25113 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25115 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25117 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25118 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25120 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25122 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25123 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25125 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25127 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25129 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25130 * config/arm/cortex-a7.md: New bypasses using
25131 arm_mac_accumulator_is_result.
25133 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25135 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25136 (cortex_a7_neon_mla): Likewise.
25137 (cortex_a7_fpfmad): New reservation.
25138 (cortex_a7_fpmacs): Use ffmas and update required units.
25139 (cortex_a7_fpmuld): Update required units and latency.
25140 (cortex_a7_fpmacd): Likewise.
25141 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25142 (cortex_a7_neon). Likewise.
25143 (bypass) Update participating units.
25145 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25147 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25148 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25150 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25151 (vfp_fmul): Use ffmad.
25152 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25153 (cortex_r4_fmacd): Use ffmad.
25154 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25155 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25156 (cortex_a9_fmacd): Use ffmad.
25157 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25158 (cortex_a8_vfp_macd): Use ffmad.
25159 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25160 (cortex_a5_fpmacd): Use ffmad.
25161 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25162 (cortex_a15_vfp_macd): Use ffmad.
25163 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25165 2013-01-29 Jason Merrill <jason@redhat.com>
25168 * varasm.c (default_assemble_visibility): Don't warn about
25169 visibility on artificial decls.
25171 2013-01-29 Richard Biener <rguenther@suse.de>
25173 PR tree-optimization/56113
25174 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25176 (label_visit): Free duplicate bitmaps and record the leader instead.
25177 (perform_var_substitution): Adjust.
25179 2013-01-29 Richard Biener <rguenther@suse.de>
25181 PR tree-optimization/55270
25182 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25183 the CFG, schedule loops for fixup.
25185 2013-01-29 Nick Clifton <nickc@redhat.com>
25187 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25190 2013-01-28 Leif Ekblad <leif@rdos.net>
25192 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25193 * config/i386/i386.h (TARGET_RDOS): New macro.
25194 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25195 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25196 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25197 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25198 DEFAULT_LARGE_SECTION_THRESHOLD.
25199 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25200 * config/i386/rdos.h: New file.
25201 * config/i386/rdos64.h: New file.
25203 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25206 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25209 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25211 PR rtl-optimization/56117
25212 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25213 call cselib_lookup_from_insn on the MEM before calling
25214 add_insn_mem_dependence.
25216 2013-01-28 Richard Biener <rguenther@suse.de>
25218 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25219 to a stmt that didn't have one.
25220 (copy_phis_for_bb): Likewise for PHI arguments.
25221 (copy_debug_stmt): Likewise for debug stmts.
25223 2013-01-28 Richard Biener <rguenther@suse.de>
25225 PR tree-optimization/56034
25226 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25227 (partition_builtin_p): Adjust.
25228 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25229 it is the last partition.
25230 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25231 up the vertex for the definition.
25232 (classify_partition): Classify whether a partition is a
25233 PKIND_REDUCTION, thus has uses outside of the loop.
25234 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25235 Merge all PKIND_REDUCTION partitions into the last partition.
25236 (tree_loop_distribution): Seed partitions from reductions as well.
25238 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25240 PR tree-optimization/56125
25241 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25242 pow(x,c) into sqrt(x) * powi(x, n/2) or
25243 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25244 optimizing for size.
25245 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25246 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25249 PR tree-optimization/56094
25250 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25251 to UNKNOWN_LOCATION while gimplifying expr.
25253 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25256 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25257 operand 0 in movabs insn template for -masm=intel asm alternative.
25258 (*movabs<mode>_2): Ditto for operand 1.
25260 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25263 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25266 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25268 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25271 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25273 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25274 Undef to avoid warning.
25276 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25278 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25279 * configure: Regenerate.
25281 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25283 PR tree-optimization/56098
25284 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25285 for stmts with volatile ops.
25286 (cond_store_replacement): Don't optimize if assign has volatile ops.
25287 (cond_if_else_store_replacement_1): Don't optimize if either
25288 then_assign or else_assign have volatile ops.
25289 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25292 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25294 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25296 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25298 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25299 missing ':' in asm example.
25301 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25303 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25304 entries into lane and laneq entries.
25305 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25306 Remove AdvSIMD scalar modes.
25307 (aarch64_sq<r>dmulh_laneq<mode>): New.
25308 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25310 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25311 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25312 * config/aarch64/iterators.md (VCOND): New.
25315 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25318 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25319 Add NULL LIBNAME argument to existing definitions.
25320 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25321 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25322 * config/avr/avr.c (DEF_BUILTIN): Same.
25323 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25324 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25325 implementation is available (DECL_ASSEMBLER_NAME is set).
25326 (avr_fold_absfx): New static function.
25327 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25328 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25329 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25330 AVR_BUILTIN_ABSLLK.
25331 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25332 (abshk, absk, abslk, absllk): Provide as static inline functions.
25334 2013-01-25 Marek Polacek <polacek@redhat.com>
25336 PR tree-optimization/56035
25337 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25339 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25341 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25342 (*movtf_internal_rex64): Add (!o,C) alternative
25343 (*movxf_internal_rex64): Ditto.
25344 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25346 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25348 * doc/invoke.texi: fix typo.
25349 * doc/objc.texi: fix typo.
25351 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25353 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25354 for the first two alternatives.
25356 2013-01-24 Diego Novillo <dnovillo@google.com>
25358 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25359 (ggc-zone.o): Remove.
25360 * configure.ac: Remove option --with-gc.
25361 * configure: Re-generate.
25362 * doc/install.texi: Remove documentation for --with-gc.
25363 * gengtype.c (write_enum_defn): Remove. Update all users.
25364 (write_Types_process_field): Remove generation of gt_e_* argument.
25365 (output_type_enum): Remove. Update all users.
25366 (write_enum_defn): Remove. Update all users.
25367 (enum alloc_zone): Remove. Update all users.
25368 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25369 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25370 Update all callers.
25371 (struct ptr_data): Remove field TYPE. Update all users.
25372 (gt_pch_note_object): Remove argument TYPE. Update all users.
25373 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25375 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25376 (struct alloc_zone): Remove.
25377 (ggc_internal_alloc_zone_stat): Remove.
25378 (ggc_internal_cleared_alloc_zone_stat): Remove.
25379 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25380 (ggc_pch_count_object): Remove last argument. Update all users.
25381 (ggc_pch_alloc_object): Remove last argument. Update all users.
25382 (struct alloc_zone): Remove.
25383 * ggc-zone.c: Remove.
25384 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25385 (struct alloc_zone): Remove.
25386 (ggc_alloc_typed_stat): Remove.
25387 (ggc_alloc_typed): Remove.
25388 (ggc_splay_alloc): Remove first argument.
25389 (rtl_zone): Remove. Update all users.
25390 (tree_zone): Remove. Update all users.
25391 (tree_id_zone): Remove. Update all users.
25392 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25393 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25394 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25395 * tree-ssanames.c: Remove references to zone allocator in comments.
25397 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25399 * config/avr/avr.c (avr_out_fract): Make register numbers that
25400 might be outside of source operand signed.
25402 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25404 * config/i386/constraints.md (Yf): New constraint.
25405 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25406 of f constraint to conditionaly disable x87 register preferences.
25407 (*movdf_internal): Ditto.
25408 (*movsf_internal): Ditto.
25410 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25412 PR inline-asm/55934
25413 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25414 that have operands with impossible constraints.
25415 Add a FIXME for a speed-up opportunity.
25416 * lra-constraints.c (process_alt_operands): Verify that a class
25417 selected from constraints on asms is valid for the operand mode.
25418 (curr_insn_transform): Remove incorrect comment.
25420 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25422 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25423 TOC operand is a valid symbol ref in the constant pool.
25425 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25427 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25429 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25432 * config/avr/stdfix.h: New file.
25433 * t-avr (stdfix-gcc.h): New rule to build it.
25434 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25436 2013-01-23 Kostya Serebryany <kcc@google.com>
25438 * config/darwin.h: remove dependency on
25439 CoreFoundation (asan on Mac OS).
25441 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25444 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25445 instead of cmpdi_operand for first comparison operand.
25446 Don't assert that comparison operands aren't both constants.
25448 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25450 * doc/install.texi (Downloading the Source): Update references to
25451 downloading separate components.
25453 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25455 * doc/extend.texi (__int128): Improve grammar.
25457 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25460 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25461 alternative to (o,r).
25462 (*movdi_internal_rex64): Remove (!o,n) alternative.
25463 (DImode immediate->memory splitter): Remove.
25464 (DImode immediate->memory peephole2): Remove.
25465 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25466 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25467 alternative to (!o,*r).
25468 (*movtf_internal_sse): New pattern.
25469 (*movxf_internal_rex64): New pattern.
25470 (*movxf_internal): Disable for TARGET_64BIT.
25471 (*movdf_internal_rex64): Remove (!o,F) alternative.
25473 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25475 PR middle-end/56074
25476 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25477 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25478 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25479 stmt locations where LOCATION_LOCUS of the stmt location is
25480 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25483 * config/i386/i386.md (UNSPEC_STOS): New.
25484 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25485 *strsetqi_1): Add UNSPEC_STOS.
25487 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25490 * doc/invoke.texi: Remove left over -Wsynth example.
25492 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25494 PR tree-optimization/56051
25495 * fold-const.c (fold_binary_loc): Don't fold
25496 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25497 a narrowing conversion, or widening conversion from signed
25500 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25502 PR rtl-optimization/56023
25503 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25504 dependent on debug instruction.
25506 2013-01-21 Martin Jambor <mjambor@suse.cz>
25508 PR middle-end/56022
25509 * function.c (allocate_struct_function): Call
25510 invoke_set_current_function_hook earlier.
25512 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25514 * reload1.c (init_reload): Only initialize reload_obstack
25515 during the first call.
25517 2013-01-21 Marek Polacek <polacek@redhat.com>
25519 * cfgloop.c (verify_loop_structure): Fix up grammar.
25521 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25523 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25524 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25526 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25529 * config/arm/marvell-pj4.md: Update copyright year.
25530 Fix up use of alu to alu_reg and simple_alu_imm.
25532 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25534 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25536 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25539 * lra-constraints.c (curr_insn_transform): Don't reuse original
25540 insn for secondary memory move when memory mode should be different.
25542 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25544 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25545 atomic_storedi_1): New patterns.
25547 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25549 btver2 pipeline descriptions.
25550 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25552 * config/i386/i386.md (btver2_decode): New type attributes.
25553 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25555 * config/i386/btver2.md: New file describing btver2 pipelines.
25557 2013-01-19 Andrew Pinski <apinski@cavium.com>
25559 PR tree-optimization/52631
25560 * tree-ssa-sccvn (visit_use): Before looking up the original
25561 statement, try looking up the simplified expression.
25563 2013-01-19 Anthony Green <green@moxielogic.com>
25565 * config/moxie/moxie.c (moxie_expand_prologue): Set
25566 current_function_static_stack_size.
25568 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25570 PR tree-optimization/56029
25571 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25572 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25574 2013-01-18 Sharad Singhai <singhai@google.com>
25576 PR tree-optimization/55995
25577 * dumpfile.c (dump_loc): Print location only if available.
25578 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25580 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25583 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25584 secondary memory move.
25585 (inherit_reload_reg): Use rclass instead of cl for
25586 check_secondary_memory_needed_p.
25588 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25590 PR middle-end/56015
25591 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25592 the case where writing real complex part of target modifies op1.
25594 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25596 * config/aarch64/aarch64-simd.md
25597 (aarch64_vcond_internal<mode>): Handle unordered cases.
25598 * config/aarch64/iterators.md (v_cmp_result): New.
25600 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25601 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25603 * config/arm/marvell-pj4.md: New file.
25604 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25605 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25606 (generic_vfp): Likewise.
25607 * config/arm/arm-cores.def: Add marvell-pj4.
25608 * config/arm/arm-tune.md: Regenerate.
25609 * config/arm/arm-tables.opt: Regenerate.
25610 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25611 * doc/invoke.texi: Document marvell-pj4.
25613 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25615 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25617 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25622 * var-tracking.c (negative_power_of_two_p): New.
25623 (global_get_addr_cache, local_get_addr_cache): New.
25624 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25625 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25627 (vt_stack_offset_p): Remove.
25628 (vt_canon_true_dep): Always canonicalize loc's address.
25629 (clobber_overlapping_mems): Make sure we have a MEM.
25630 (local_get_addr_clear_given_value): New.
25631 (val_reset): Clear local cached entries.
25632 (compute_bb_dataflow): Create and release the local cache.
25633 Disable duplicate MEMs clobbering.
25634 (emit_notes_in_bb): Clobber MEMs likewise.
25635 (vt_emit_notes): Create and release the local cache.
25636 (vt_initialize, vt_finalize): Create and release the global
25637 cache, respectively.
25638 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25640 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25642 PR libmudflap/53359
25643 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25644 not found in the symtab.
25646 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25649 PR rtl-optimization/55547
25650 PR rtl-optimization/53827
25653 * alias.c (offset_overlap_p): New, factored out of...
25654 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25655 the conservative special case for symbolic constants. Don't
25656 adjust zero sizes on alignment.
25658 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25660 PR rtl-optimization/52573
25661 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25662 REG_UNUSED for the same register.
25664 2013-01-17 Richard Biener <rguenther@suse.de>
25665 Marek Polacek <polacek@redhat.com>
25667 PR rtl-optimization/55833
25668 * loop-unswitch.c (unswitch_loops): Move loop verification...
25669 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25670 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25671 Set it to true when we're removing a loop from hierarchy tree in
25672 an irreducible region.
25673 (fix_bb_placements): Adjust caller.
25674 (fix_loop_placements): Likewise.
25676 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25678 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25679 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25680 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25681 Remove ID. Adjust comments.
25682 * config/avr/avr-c.c (avr_builtin_name): Remove.
25683 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25684 * config/avr/avr.c (avr_tolower): New static function.
25685 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25686 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25687 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25690 2013-01-17 Jan Hubicka <jh@suse.cz>
25692 PR tree-optimization/55273
25693 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25695 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25698 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25699 store through atomic_store<mode>_1.
25700 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25702 2013-01-17 Martin Jambor <mjambor@suse.cz>
25704 PR tree-optimizations/55264
25705 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25706 for virtual methods.
25707 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25708 virtual methods before inlining is over.
25709 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25711 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25714 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25716 PR rtl-optimization/56005
25717 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25718 pending reads for prefetch.
25720 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25722 * config/aarch64/aarch64.md
25723 (*cstoresi_neg_uxtw): New pattern.
25724 (*cmovsi_insn_uxtw): New pattern.
25725 (*<optab>si3_uxtw): New pattern.
25726 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25727 (*<optab>si3_insn_uxtw): New pattern.
25728 (*bswapsi2_uxtw): New pattern.
25730 2013-01-16 Richard Biener <rguenther@suse.de>
25732 * tree-inline.c (tree_function_versioning): Remove set but
25733 never used variable.
25735 2013-01-16 Richard Biener <rguenther@suse.de>
25737 PR tree-optimization/55964
25738 * tree-flow.h (rename_variables_in_loop): Remove.
25739 (rename_variables_in_bb): Likewise.
25740 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25741 (copy_loop_before): Adjust and delete update-ssa status.
25742 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25743 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25744 (rename_variables_in_loop): Remove.
25745 (slpeel_update_phis_for_duplicate_loop): Likewise.
25746 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25747 use available cfg machinery instead of duplicating it.
25748 Update PHI nodes and perform poor-mans SSA update here.
25749 (slpeel_tree_peel_loop_to_edge): Adjust.
25751 2013-01-16 Richard Biener <rguenther@suse.de>
25753 PR tree-optimization/54767
25754 PR tree-optimization/53465
25755 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25756 (vrp_visit_phi_node): For PHI arguments coming via backedges
25757 drop all symbolical range information.
25758 (execute_vrp): Compute backedges.
25760 2013-01-16 Richard Biener <rguenther@suse.de>
25762 * doc/install.texi: Update CLooG and ISL requirements to
25765 2013-01-16 Christian Bruel <christian.bruel@st.com>
25768 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25769 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25770 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25772 2013-01-16 DJ Delorie <dj@redhat.com>
25774 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25775 (UNSPECV_SP_SWITCH_E): New.
25776 (sp_switch_1): Change to an unspec.
25777 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25780 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25782 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25783 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25784 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25785 (expand_mem_thread_fence): Ditto.
25786 (expand_mem_signal_fence): Ditto.
25787 (expand_atomic_load): Ditto.
25788 (expand_atomic_store): Ditto.
25790 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25792 PR rtl-optimization/55547
25793 PR rtl-optimization/53827
25796 * alias.c (memrefs_conflict_p): Set sizes to negative after
25799 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25802 * function.c (thread_prologue_and_epilogue_insns): Always
25803 add crtl->drap_reg to set_up_by_prologue.set, even if
25804 stack_realign_drap is false.
25806 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25808 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25809 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25810 *call): Fix indention.
25812 2013-01-15 Tom de Vries <tom@codesourcery.com>
25815 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25818 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25820 PR rtl-optimization/55153
25821 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25823 2013-01-15 Martin Jambor <mjambor@suse.cz>
25825 PR tree-optimization/55920
25826 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25827 accesses as grp_to_be_debug_replaced.
25829 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25831 PR tree-optimization/55920
25832 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25833 there is non-useless type conversion needed from debug rhs to lhs,
25834 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25836 2013-01-15 Joseph Myers <joseph@codesourcery.com>
25837 Mikael Pettersson <mikpe@it.uu.se>
25840 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25842 (ASM_OUTPUT_CASE_LABEL): Remove.
25843 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25844 * final.c (shorten_branches): Update alignment of labels before
25845 jump tables if CASE_VECTOR_SHORTEN_MODE.
25847 2013-01-15 Richard Biener <rguenther@suse.de>
25850 * system.h: Do not include gmp.h for building host tools.
25852 2013-01-15 Richard Biener <rguenther@suse.de>
25854 PR middle-end/55882
25855 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25856 account for bitpos when computing alignment.
25858 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25860 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25861 (ix86_target_macros_internal): Likewise.
25863 * config/i386/i386.c (m_CORE2I7): Removed.
25864 (m_CORE_HASWELL): New macro.
25865 (m_CORE_ALL): Likewise.
25866 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25867 (initial_ix86_arch_features): Likewise.
25868 (processor_target_table): Initializations for Core avx2.
25869 (cpu_names): New names "core-avx2".
25870 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25871 PROCESSOR_CORE_HASWELL.
25872 (ix86_issue_rate): New case.
25873 (ia32_multipass_dfa_lookahead): Likewise.
25874 (ix86_sched_init_global): Likewise.
25876 * config/i386/i386.h (TARGET_HASWELL): New macro.
25877 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25878 (processor_type): New PROCESSOR_HASWELL.
25880 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25882 PR tree-optimization/55955
25883 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25884 *SHIFT_EXPR and *ROTATE_EXPR codes.
25886 PR tree-optimization/48766
25887 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25888 -ftrapv disable -fwrapv.
25890 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25893 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25894 etc. to 1 and not to __flash.
25895 Use LL suffix for __INT24_MAX__ with -mint8.
25896 Use ULL suffix for __UINT24_MAX__ with -mint8.
25898 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25900 * config/avr/avr-arch.h
25901 (struct base_arch_s): Use typedef avr_arch_t instead.
25902 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25903 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25904 * config/avr/avr.c: Same.
25905 * config/avr/avr-devices.c: Same.
25906 * config/avr/driver-avr.c: Same.
25907 * config/avr/gen-avr-mmcu-texi.c: Same.
25908 * config/avr/avr-mcus.def: Adjust comment.
25910 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25912 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25913 * config/aarch64/iterators.md (VALLDI): New.
25915 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
25916 Andi Kleen <ak@linux.intel.com>
25919 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25920 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25923 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25925 * config/avr/avr-stdint.h: Remove trailing blanks.
25926 * config/avr/avr-log.h: Same.
25927 * config/avr/avr-arch.h: Same.
25928 * config/avr/avr-devices.c: Same.
25929 * config/avr/avr-dimode.md: Same.
25930 * config/avr/predicates.md: Same.
25931 * config/avr/avr-c.c: Same. And fix typo.
25933 * config/avr/avr-protos.h: Same. And:
25934 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25935 (init_cumulative_args): Rename to avr_init_cumulative_args.
25936 (expand_prologue): Rename to avr_expand_prologue.
25937 (expand_epilogue): Rename to avr_expand_epilogue.
25938 (adjust_insn_length): Rename to avr_adjust_insn_length.
25939 (notice_update_cc): Rename to avr_notice_update_cc.
25940 (final_prescan_insn): Rename to avr_final_prescan_insn.
25941 * config/avr/avr.c: Same.
25942 * config/avr/avr.h: Same.
25943 * config/avr/avr.md: Remove trailing blanks.
25944 (prologue): Use avr_expand_prologue.
25945 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
25947 2013-01-14 Richard Biener <rguenther@suse.de>
25949 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
25950 verify_location, collect_subblocks): New functions.
25951 (verify_gimple_in_cfg): Verify that locations only reference
25952 BLOCKs in the functions BLOCK tree.
25954 2013-01-14 Richard Biener <rguenther@suse.de>
25956 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
25958 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
25960 (insert_out_of_ssa_copy_on_edge): Likewise.
25961 (rewrite_close_phi_out_of_ssa): Likewise.
25962 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
25964 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
25965 propagated constants.
25966 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
25969 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25971 * config/avr/avr-modes.def: Add GPL copyright notice.
25973 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
25975 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
25976 MEMMODEL_MASK to determine memory model.
25977 (atomic_store<mode>): Ditto from operands[2].
25978 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
25980 2013-01-13 Jakub Jelinek <jakub@redhat.com>
25983 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
25984 (fold_gimple_assign): Don't call unshare_expr here.
25985 (fold_ctor_reference): Call unshare_expr.
25987 2013-01-13 Terry Guo <terry.guo@arm.com>
25989 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
25990 * doc/fragments.texi: Document MULTILIB_REUSE.
25991 * gcc.c (multilib_reuse): New internal spec.
25992 (set_multilib_dir): Also search multilib from multilib_reuse.
25993 * genmultilib (tmpmultilib3): Refactor code.
25994 (tmpmultilib4): Ditto.
25995 (multilib_reuse): New multilib argument.
25997 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
25999 * Makefile.in: Update copyright.
26001 2013-01-12 Tom de Vries <tom@codesourcery.com>
26003 PR middle-end/55890
26004 * calls.c (expand_call): Check if arg_nr is valid.
26006 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26008 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26009 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26010 documentation. Add missing '__' in front of
26011 __builtin_ia32_packssdw256.
26013 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26016 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26017 NO_REGS for larl operands.
26018 (s390_reload_larl_operand): Use s390_load_address instead of
26021 2013-01-11 Richard Biener <rguenther@suse.de>
26023 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26024 (verify_node_sharing): ... here.
26025 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26027 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26029 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26030 Substitute TREECHECKING.
26031 * configure: Regenerate.
26032 * Makefile.in (TREECHECKING): New.
26034 2013-01-11 Richard Guenther <rguenther@suse.de>
26036 PR tree-optimization/44061
26037 * tree-vrp.c (extract_range_basic): Compute zero as
26038 value-range for __builtin_constant_p of function parameters.
26040 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26042 Update copyright years.
26044 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26046 PR rtl-optimization/55672
26047 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26048 const to be eliminable.
26050 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26052 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26053 * configure: Regenerate.
26055 2013-01-10 Richard Biener <rguenther@suse.de>
26057 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26059 2013-01-10 Richard Biener <rguenther@suse.de>
26062 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26063 locations for virtual PHI arguments.
26064 (rewrite_update_phi_arguments): Likewise.
26066 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26068 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26071 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26073 PR tree-optimization/55921
26074 * tree-complex.c (expand_complex_asm): New function.
26075 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26077 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26080 * config/s390/s390.c (s390_symref_operand_p)
26081 (s390_loadrelative_operand_p): Merge the two functions.
26082 (s390_check_qrst_address, print_operand_address): Add parameters
26083 to s390_loadrelative_operand_p invokation.
26084 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26085 (s390_reload_larl_operand, s390_secondary_reload): Use
26086 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26087 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26089 2013-01-09 Mike Stump <mikestump@comcast.net>
26091 * dse.c (record_store): Remove unnecessary assert.
26093 2013-01-09 Jan Hubicka <jh@suse.cz>
26095 PR tree-optimization/55569
26096 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26097 * cfgloop.h (scale_loop_profile): Likewise.
26099 2013-01-09 Jan Hubicka <jh@suse.cz>
26102 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26104 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26106 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26108 PR middle-end/55114
26109 * expr.h (maybe_emit_group_store): Declare.
26110 * expr.c (maybe_emit_group_store): New function.
26111 * builtins.c (expand_builtin_int_roundingfn): Call it.
26112 (expand_builtin_int_roundingfn_2): Likewise.
26114 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26116 PR rtl-optimization/55829
26117 * lra-constraints.c (match_reload): Add code for absent output.
26118 (curr_insn_transform): Add code for reloads of matched inputs
26121 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26123 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26124 attribute of movddup insn to DF.
26125 (*vec_interleave_lowv2df): Ditto.
26126 (vec_dupv2df): Ditto.
26128 2013-01-09 Jan Hubicka <jh@suse.cz>
26130 PR tree-optimiation/55875
26131 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26132 EVERY_ITERATION parameter.
26133 (number_of_iterations_exit): Check if exit is executed every iteration.
26134 (idx_infer_loop_bounds): Similarly here.
26135 (n_of_executions_at_most): Simplify
26136 to only test for cases where statement is dominated by the
26137 particular bound; handle correctly the "postdominance" test.
26138 (scev_probably_wraps_p): Use max loop iterations info
26139 as a global bound first.
26141 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26142 Nick Clifton <nickc@redhat.com>
26144 * config/v850/v850.md (cbranchsf4): New pattern.
26145 (cstoresf4): New pattern.
26146 (cbranchdf4): New pattern.
26147 (cstoredf4): New pattern.
26148 (movsicc): Disallow floating point comparisons.
26149 (cmpsf_le_insn): Fix order of operators.
26150 (cmpsf_lt_insn): Likewise.
26151 (cmpsf_eq_insn): Likewise.
26152 (cmpdf_le_insn): Likewise.
26153 (cmpdf_lt_insn): Likewise.
26154 (cmpdf_eq_insn): Likewise.
26155 (cmpsf_ge_insn): Use LE comparison.
26156 (cmpdf_ge_insn): Likewise.
26157 (cmpsf_gt_insn): Use LT comparison.
26158 (cmpdf_gt_insn): Likewise.
26159 (cmpsf_ne_insn): Delete pattern.
26160 (cmpdf_ne_insn): Delete pattern.
26161 * config/v850/v850.c (v850_gen_float_compare): Use
26162 gen_cmpdf_eq_insn for NE comparison.
26163 (v850_float_z_comparison_operator)
26164 (v850_float_nz_comparison_operator): Move from here ...
26165 * config/v850/predicates.md: ... to here. Move GT and GE
26166 comparisons into v850_float_z_comparison_operator.
26167 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26169 (v850_float_nz_comparison_operator): Likewise.
26171 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26173 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26174 with calls to gen_insvsi/gen_insvdi.
26176 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26178 * config/i386/i386.c (initial_ix86_tune_features): Set up
26179 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26181 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26182 Jakub Jelinek <jakub@redhat.com>
26184 PR tree-optimization/48189
26185 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26186 If nitercst is 0, don't predict the exit edge.
26188 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26190 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26191 in asm_fprintf with reg_names.
26192 (aarch64_print_operand_address): Likewise.
26193 (aarch64_return_addr): Likewise.
26194 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26196 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26198 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26199 (INT_U6_BITS): Likewise.
26200 * config/pa/predicates.md (uint6_operand): New predicate.
26201 (shift5_operand, shift6_operand): Likewise.
26202 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26204 (lshrdi3): Use shift6_operand.
26205 (shrpsi4, shrpdi4): New insn patterns.
26206 (extzv): Delete expander.
26207 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26208 predicates in unamed zero extract patterns. Tighten common constraint.
26209 (extv): Delete expander.
26210 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26211 predicates in unamed sign extract patterns. Tighten common constraint.
26212 (insv): Delete expander.
26213 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26214 predicates in unamed insert patterns. Tighten common constraint.
26215 Change uint32_operand predicate to uint6_operand predicate in unamed
26216 DImode pattern to insert constant values of type 1...1xxxx.
26218 2013-01-04 Jan Hubicka <jh@suse.cz>
26220 PR tree-optimization/55823
26221 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26224 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26225 Uros Bizjak <ubizjak@gmail.com>
26227 PR rtl-optimization/55845
26228 * df-problems.c (can_move_insns_across): Stop scanning at
26229 volatile_insn_p source instruction or give up if
26230 across_from .. across_to range contains any volatile_insn_p
26233 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26235 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26236 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26238 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26239 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26241 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26244 * asan.c (asan_clear_shadow): New function.
26245 (asan_emit_stack_protection): Use it.
26247 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26249 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26250 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26251 with tab instead of space.
26253 2013-01-08 Nick Clifton <nickc@redhat.com>
26255 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26256 register bank 0 at the start of an interrupt handler.
26257 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26260 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26262 * config/aarch64/aarch64-simd.md
26263 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26264 (aarch64_simd_bsl): Likewise.
26265 (aarch64_vcond_internal<mode>): Likewise.
26266 (vcond<mode><mode>): Likewise.
26267 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26268 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26270 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26272 * config/aarch64/aarch64-builtins.c
26273 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26275 2013-01-08 Martin Jambor <mjambor@suse.cz>
26278 * tree-sra.c (analyze_access_subtree): Return true also after
26279 potentially creating a debug-only replacement.
26281 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26283 PR middle-end/55890
26284 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26286 PR tree-optimization/54120
26287 * tree-vrp.c (range_fits_type_p): Don't allow
26288 src_precision < precision from signed vr to unsigned_p
26289 if vr->min or vr->max is negative.
26290 (simplify_float_conversion_using_ranges): Test can_float_p
26291 against CODE_FOR_nothing.
26293 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26294 Richard Biener <rguenther@suse.de>
26296 PR middle-end/55851
26297 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26298 types instead of just INTEGER_TYPE types.
26300 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26302 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26305 2013-01-07 Steve Ellcey <sellcey@mips.com>
26308 * config/mips/mips.opt: Change mad to mmad to match documentation.
26310 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26313 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26314 .progmemx.data now.
26316 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26319 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26320 (avr_addrspace_t): Add .section_name field.
26321 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26323 (avr_addrspace): Same. Initialize .section_name. Remove last
26324 NULL entry. Put __memx into .progmemx.data.
26325 (progmem_section_prefix): Remove.
26326 (avr_asm_init_sections): No need to initialize progmem_section.
26327 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26328 section name prefix.
26329 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26330 retrieve the progmem section.
26331 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26332 boundary to run over avr_addrspace[].
26333 (avr_register_target_pragmas): Ditto.
26335 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26337 * varasm.c (output_constant_def_contents): For asan_protect_global
26338 protected strings, adjust DECL_ALIGN if needed, before testing for
26340 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26341 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26343 (output_object_block): For asan protected decls, emit asan padding
26344 after their contents.
26345 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26346 (asan_finish_file): Test it here instead.
26348 2013-01-07 Nick Clifton <nickc@redhat.com>
26349 Matthias Klose <doko@debian.org>
26350 Doug Kwan <dougkwan@google.com>
26351 H.J. Lu <hongjiu.lu@intel.com>
26354 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26356 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26358 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26360 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26362 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26364 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26367 * doc/install.texi (Cross-Compiler-Specific Options): Document
26370 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26372 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26373 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26374 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26375 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26376 vqmovun_high_s64): Fix source operand number and update copyright.
26378 2013-01-07 Richard Biener <rguenther@suse.de>
26380 PR middle-end/55890
26381 * gimple.h (gimple_call_builtin_p): New overload.
26382 * gimple.c (validate_call): New function.
26383 (gimple_call_builtin_p): Likewise.
26384 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26385 Use gimple_call_builtin_p.
26386 (find_func_clobbers): Likewise.
26387 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26388 (strlen_optimize_stmt): Likewise.
26390 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26392 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26393 (vld1q_dup_*): Likewise.
26394 (vld1_*): Likewise.
26395 (vld1q_*): Likewise.
26396 (vld1_lane_*): Likewise.
26397 (vld1q_lane_*): Likewise.
26399 2013-01-07 Richard Biener <rguenther@suse.de>
26401 * lto-streamer.h (LTO_minor_version): Bump to 2.
26403 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26405 * config/aarch64/aarch64-protos.h
26406 (aarch64_const_double_zero_rtx_p): Rename to...
26407 (aarch64_float_const_zero_rtx_p): ...this.
26408 (aarch64_float_const_representable_p): New.
26409 (aarch64_output_simd_mov_immediate): Likewise.
26410 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26411 move immediate case.
26412 * config/aarch64/aarch64.c
26413 (aarch64_const_double_zero_rtx_p): Rename to...
26414 (aarch64_float_const_zero_rtx_p): ...this.
26415 (aarch64_print_operand): Allow printing of new constants.
26416 (aarch64_valid_floating_const): New.
26417 (aarch64_legitimate_constant_p): Check for valid floating-point
26419 (aarch64_simd_valid_immediate): Likewise.
26420 (aarch64_vect_float_const_representable_p): New.
26421 (aarch64_float_const_representable_p): Likewise.
26422 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26423 (aarch64_output_simd_mov_immediate): New.
26424 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26425 (*movdf_aarch64): Likewise.
26426 * config/aarch64/constraints.md (Ufc): New.
26427 (Y): call aarch64_float_const_zero_rtx.
26428 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26430 2013-01-07 Richard Biener <rguenther@suse.de>
26432 PR tree-optimization/55888
26433 PR tree-optimization/55862
26434 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26435 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26436 not if it is contained therein.
26438 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26440 * config/avr/t-avr: Typo.
26442 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26445 * config/avr/t-avr: Don't automatically rebuild
26446 $(srcdir)/config/avr/t-multilib
26447 $(srcdir)/config/avr/avr-tables.opt
26448 $(srcdir)/doc/avr-mmcu.texi
26449 (avr-mcus): New phony target to build them on request.
26450 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26451 * avr/avr-mcus.def: Adjust comments.
26453 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26455 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26457 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26459 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26461 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26463 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26465 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26467 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26468 to generate profiling.
26469 * config/rs6000/aix64.h (LIB_SPEC): Same.
26471 2013-01-04 Andrew Pinski <apinski@cavium.com>
26473 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26475 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26477 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26479 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26481 (ix86_expand_move): Ditto.
26482 (ix86_zero_extend_to_Pmode): Ditto.
26483 (ix86_expand_call): Ditto.
26484 (ix86_expand_special_args_builtin): Ditto.
26485 (ix86_expand_builtin): Ditto.
26487 2013-01-04 Richard Biener <rguenther@suse.de>
26489 PR tree-optimization/55862
26490 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26491 translating them through PHI nodes.
26493 2013-01-04 Martin Jambor <mjambor@suse.cz>
26495 PR tree-optimization/55755
26496 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26497 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26499 2013-01-04 Marek Polacek <polacek@redhat.com>
26501 PR middle-end/55859
26502 * opts.c (default_options_optimization): Clarify error message.
26504 2013-01-04 Richard Biener <rguenther@suse.de>
26506 PR middle-end/55863
26507 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26510 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26513 * config/pa/pa.md (movsi): Revert previous change.
26514 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26517 2013-01-03 Richard Henderson <rth@redhat.com>
26519 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26520 after eliminating TLS symbols.
26522 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26525 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26526 * graphite-poly.c (debug_gmp_value): Likewise.
26528 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26531 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26532 selected code model, define __code_mode_small__, __code_model_medium__,
26533 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26534 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26535 xchg temporary register with %k. Declare temporary register as
26537 [__x86_64__]: For medium and large code models, preserve %rbx register.
26539 2013-01-03 Richard Biener <rguenther@suse.de>
26541 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26542 (dump_subscript): Adjust.
26543 (finalize_ddr_dependent): Do not dump redundant info.
26544 (analyze_siv_subscript): Adjust.
26545 (subscript_dependence_tester): Likewise.
26546 (compute_affine_dependence): Likewise.
26548 2013-01-03 Richard Biener <rguenther@suse.de>
26551 2013-01-03 Richard Biener <rguenther@suse.de>
26553 PR tree-optimization/55857
26554 * tree-vect-stmts.c (vectorizable_load): Do not setup
26555 re-alignment for invariant loads.
26557 2013-01-02 Richard Biener <rguenther@suse.de>
26559 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26560 invariant load do not generate a vector load from the scalar location.
26562 2013-01-03 Richard Biener <rguenther@suse.de>
26564 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26565 for not vectorizing.
26566 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26567 not build INDIRECT_REFs, call get_name once only.
26568 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26569 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26571 2013-01-03 Richard Biener <rguenther@suse.de>
26573 PR tree-optimization/55857
26574 * tree-vect-stmts.c (vectorizable_load): Do not setup
26575 re-alignment for invariant loads.
26577 2013-01-03 Richard Biener <rguenther@suse.de>
26580 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26581 prefer a built-in decl.
26583 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26585 * gcc.c (process_command): Update copyright notice dates.
26586 * gcov.c (print_version): Likewise.
26587 * gcov-dump.c (print_version): Likewise.
26589 PR rtl-optimization/55838
26590 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26591 iv0.step, iv1.step and step.
26593 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26594 Marc Glisse <marc.glisse@inria.fr>
26596 PR tree-optimization/55832
26597 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26598 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26599 integer_{one,zero}_node.
26601 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26604 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26605 * var-tracking.c (reverse_op): Don't add reverse ops to
26606 VALUEs that have already
26607 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26609 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26611 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26613 2013-01-02 Teresa Johnson <tejohnson@google.com>
26615 * dumpfile.c (dump_loc): Print filename with location.
26616 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26617 new location_t parameter to emit complete unroll message with
26618 new dump framework.
26619 (canonicalize_loop_induction_variables): Compute loops location
26620 and pass to try_unroll_loop_completely.
26621 * loop-unroll.c (report_unroll_peel): New function.
26622 (peel_loops_completely): Use new dump format with location
26623 for main dumpfile message, and invoke report_unroll_peel on success.
26624 (decide_unrolling_and_peeling): Ditto.
26625 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26626 by report_unroll_peel.
26627 (decide_peel_completely): Ditto.
26628 (decide_unroll_constant_iterations): Ditto.
26629 (decide_unroll_runtime_iterations): Ditto.
26630 (decide_peel_simple): Ditto.
26631 (decide_unroll_stupid): Ditto.
26632 * cfgloop.c (get_loop_location): New function.
26633 * cfgloop.h (get_loop_location): Declare.
26635 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26637 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26640 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26642 PR middle-end/55198
26643 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26644 BLKmode objects when EXPAND_MEMORY is specified.
26646 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26648 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26650 (fold_builtin_cpu): Do not share cpu model decls across statements.
26652 2013-01-02 Jason Merrill <jason@redhat.com>
26655 * tree.c (build_array_type_1): Revert earlier change.
26657 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26659 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26661 * config/aarch64/aarch64-tune.md: Re-generate.
26663 2013-01-02 Richard Biener <rguenther@suse.de>
26665 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26666 invariant load do not generate a vector load from the scalar location.
26668 2013-01-02 Richard Biener <rguenther@suse.de>
26671 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26672 * configure: Regenerate.
26674 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26676 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26677 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26678 (expand_builtin_int_roundingfn_2): Keep the original target around
26679 for the fallback case.
26681 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26683 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26684 to be clear for sign changes.
26686 2013-01-01 Jan Hubicka <jh@suse.cz>
26688 * ipa-inline-analysis.c: Fix formatting.
26690 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26692 PR tree-optimization/55831
26693 * tree-vect-loop.c (get_initial_def_for_induction): Use
26694 gsi_after_labels instead of gsi_start_bb.
26696 Copyright (C) 2013 Free Software Foundation, Inc.
26698 Copying and distribution of this file, with or without modification,
26699 are permitted in any medium without royalty provided the copyright
26700 notice and this notice are preserved.