]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
projects / thirdparty / gcc.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Daily bump.
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
bf7b5747
ST		 12013-10-15 Sriraman Tallam <tmsriram@google.com>
2
3 PR target/57756
4 * optc-save-gen.awk: Add extra parameter to the save and restore
5 target calls.
6 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
7 * tree.c (build_optimization_node): New parameter. Add extra parameter
8 to call to cl_optimization_save.
9 (build_target_option_node): New parameter. Add extra parameter
10 to call to cl_target_option_save.
11 * tree.h (build_optimization_node): New parameter.
12 (build_target_option_node): New parameter.
13 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
14 build_optimization_node and build_target_option_node.
15 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
16 (handle_pragma_push_options): Ditto.
17 * toplev.c (process_options): Ditto.
18 * opts.c (init_options_struct): Check for opts_set non-null.
19 * target.def (target_option.save): New parameter.
20 (target_option.restore): New parameter.
21 * tm.texi: Generate.
22 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
23 (ix86_pragma_target_parse): Ditto.
24 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
25 parameters.
26 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
27 to build_optimization_node and build_target_option_node.
28 (rs6000_valid_attribute_p): Ditto.
29 (rs6000_pragma_target_parse): Ditto.
30 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
31 data.
32 * config/i386/i386.h:
33 TARGET_64BIT_P: New Macro
34 TARGET_MMX_P: New Macro.
35 TARGET_3DNOW_P: New Macro.
36 TARGET_3DNOW_A_P: New Macro.
37 TARGET_SSE_P: New Macro.
38 TARGET_SSE2_P: New Macro.
39 TARGET_SSE3_P: New Macro.
40 TARGET_SSSE3_P: New Macro.
41 TARGET_SSE4_1_P: New Macro.
42 TARGET_SSE4_2_P: New Macro.
43 TARGET_AVX_P: New Macro.
44 TARGET_AVX2_P: New Macro.
45 TARGET_AVX512F_P: New Macro.
46 TARGET_AVX512PF_P: New Macro.
47 TARGET_AVX512ER_P: New Macro.
48 TARGET_AVX512CD_P: New Macro.
49 TARGET_FMA_P: New Macro.
50 TARGET_SSE4A_P: New Macro.
51 TARGET_FMA4_P: New Macro.
52 TARGET_XOP_P: New Macro.
53 TARGET_LWP_P: New Macro.
54 TARGET_ABM_P: New Macro.
55 TARGET_BMI_P: New Macro.
56 TARGET_BMI2_P: New Macro.
57 TARGET_LZCNT_P: New Macro.
58 TARGET_TBM_P: New Macro.
59 TARGET_POPCNT_P: New Macro.
60 TARGET_SAHF_P: New Macro.
61 TARGET_MOVBE_P: New Macro.
62 TARGET_CRC32_P: New Macro.
63 TARGET_AES_P: New Macro.
64 TARGET_PCLMUL_P: New Macro.
65 TARGET_CMPXCHG16B_P: New Macro.
66 TARGET_FSGSBASE_P: New Macro.
67 TARGET_RDRND_P: New Macro.
68 TARGET_F16C_P: New Macro.
69 TARGET_RTM_P: New Macro.
70 TARGET_HLE_P: New Macro.
71 TARGET_RDSEED_P: New Macro.
72 TARGET_PRFCHW_P: New Macro.
73 TARGET_ADX_P: New Macro.
74 TARGET_FXSR_P: New Macro.
75 TARGET_XSAVE_P: New Macro.
76 TARGET_XSAVEOPT_P: New Macro.
77 TARGET_LP64_P: New Macro.
78 TARGET_X32_P: New Macro.
79 TARGET_FPMATH_DEFAULT_P: New Macro.
80 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
81 * config/i386/i386.c (ix86_option_override_internal): New parameters.
82 opts and opts_set.
83 Change ix86_tune_string to access opts->x_ix86_tune_string.
84 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
85 Change ix86_arch_string to access opts->x_ix86_arch_string.
86 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
87 Change ix86_pmode to access opts->x_ix86_pmode.
88 Change ix86_abi to access opts->x_ix86_abi.
89 Change ix86_cmodel to access opts->x_ix86_cmodel.
90 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
91 Change ix86_isa_flags_explicit to access
92 opts->x_ix86_isa_flags_explicit.
93 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
94 Change ix86_regparm to access opts->x_ix86_regparm.
95 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
96 Change ix86_preferred_stack_boundary_arg to access
97 opts->x_ix86_preferred_stack_boundary_arg.
98 Change ix86_force_align_arg_pointer to access
99 opts->x_ix86_force_align_arg_pointer.
100 Change ix86_incoming_stack_boundar_arg to access
101 opts->x_ix86_incoming_stack_boundar_arg.
102 Change ix86_fpmath to access opts->x_ix86_fpmath.
103 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
104 Change ix86_recip_name to access opts->x_ix86_recip_name.
105 Change ix86_stack_protector_guard to access
106 opts->x_ix86_stack_protector_guard.
107 Change ix86_tune_memcpy_strategy to access
108 opts->x_ix86_tune_memcpy_strategy.
109 Change ix86_tune_memset_strategy to access
110 opts->x_ix86_tune_memset_strategy.
111 Change global_options to access opts.
112 Change global_options_set to access opts_set.
113 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
114 Change TARGET_MMX to TARGET_MMX_P (opts->...)
115 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...)
116 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...)
117 Change TARGET_SSE to TARGET_SSE_P (opts->...)
118 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
119 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...)
120 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...)
121 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...)
122 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...)
123 Change TARGET_AVX to TARGET_AVX_P (opts->...)
124 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...)
125 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...)
126 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...)
127 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...)
128 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...)
129 Change TARGET_FMA to TARGET_FMA_P (opts->...)
130 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...)
131 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...)
132 Change TARGET_XOP to TARGET_XOP_P (opts->...)
133 Change TARGET_LWP to TARGET_LWP_P (opts->...)
134 Change TARGET_ABM to TARGET_ABM_P (opts->...)
135 Change TARGET_BMI to TARGET_BMI_P (opts->...)
136 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...)
137 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...)
138 Change TARGET_TBM to TARGET_TBM_P (opts->...)
139 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...)
140 Change TARGET_SAHF to TARGET_SAHF_P (opts->...)
141 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...)
142 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...)
143 Change TARGET_AES to TARGET_AES_P (opts->...)
144 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...)
145 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...)
146 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...)
147 Change TARGET_RDRND to TARGET_RDRND_P (opts->...)
148 Change TARGET_F16C to TARGET_F16C_P (opts->...)
149 Change TARGET_RTM to TARGET_RTM_P (opts->...)
150 Change TARGET_HLE to TARGET_HLE_P (opts->...)
151 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...)
152 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...)
153 Change TARGET_ADX to TARGET_ADX_P (opts->...)
154 Change TARGET_FXSR to TARGET_FXSR_P (opts->...)
155 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...)
156 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...)
157 Change TARGET_LP64 to TARGET_LP64_P (opts->...)
158 Change TARGET_X32 to TARGET_X32_P (opts->...)
159 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...)
160 Change TARGET_FLOAT_RETURNS_IN_80387 to
161 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...)
162 (ix86_function_specific_save): New parameter. Use opts-> fields
163 to replace global fields.
164 (ix86_function_specific_restore): Ditto.
165 (ix86_valid_target_attribute_inner_p): New parameters.
166 Fix recursive call.
167 Fix call to ix86_handle_option and set_option.
168 (ix86_valid_target_attribute_tree): New parameters.
169 Change global_options to access opts.
170 Change global_options_set to access opts_set.
171 Fix call to ix86_valid_target_attribute_inner_p.
172 Change ix86_tune_string to access opts->x_ix86_tune_string.
173 Change ix86_arch_string to access opts->x_ix86_arch_string.
174 Change ix86_fpmath to access opts->x_ix86_fpmath
175 Fix call to ix86_option_override_internal.
176 Fix call to ix86_add_new_builtins.
177 Fix calls to build_optimization_node and build_target_option_node.
178 (ix86_valid_target_attribute_p): Remove access to global_options.
179 Use new gcc_options structure func_options.
180 Fix call to ix86_valid_target_attribute_tree.
181 Fix call to build_optimization_node.
182 (get_builtin_code_for_version): Fix call to
183 ix86_valid_target_attribute_tree.
184
459260ec
DM		 1852013-10-15 David Malcolm <dmalcolm@redhat.com>
186
187 * Makefile.in (PICFLAG): New.
188 (enable_host_shared): New.
189 (INTERNAL_CFLAGS): Use PICFLAG.
190 (LIBIBERTY): Use pic build of libiberty.a if configured with
191 --enable-host-shared.
192 * configure.ac: Add --enable-host-shared, setting up new
193 PICFLAG variable.
194 * configure: Regenerate.
195 * doc/install.texi (--enable-shared): Add note contrasting it
196 with...
197 (--enable-host-shared): New option.
198
16022041
RS		 1992013-10-15 Richard Biener <rguenther@suse.de>
200
201 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
202 for built-in functions.
203
b114bfb4
ZC		 2042013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
205
206 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
207 (optimize_range_tests_1): New function,
208 extracted from optimize_range_tests.
209 (optimize_range_tests_xor): Similarly.
210 (optimize_range_tests_diff): New function.
211 (optimize_range_tests): Use optimize_range_tests_1.
212
69b8f2f9
CH		 2132013-10-15 Cong Hou <congh@google.com>
214
215 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
216 requirement of the reduction pattern so that one operand of the
217 reduction operation can come from outside of the loop.
218
8ff5a4e4
JG		 2192013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
220
221 * config/arm/neon-schedgen.ml: Remove.
222 * config/arm/cortex-a9-neon.md: Remove comment regarding
223 neon-schedgen.ml.
224
52b69ad2
JG		 2252013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
226
227 * config/arm/types: Remove old neon types.
228
e7c6a556
JG		 2292013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
230
231 * config/arm/cortex-a7.md
232 (cortex_a7_neon_type): New.
233 (cortex_a7_neon_mul): Update for new types.
234 (cortex_a7_neon_mla): Likewise.
235 (cortex_a7_neon): Likewise.
236
795dd8d5
JG		 2372013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
238
239 * config/arm/cortex-a15-neon.md
240 (cortex_a15_neon_type): New,
241
242 (cortex_a15_neon_int_1): Remove.
243 (cortex_a15_neon_int_2): Likewise.
244 (cortex_a15_neon_int_3): Likewise.
245 (cortex_a15_neon_int_4): Likewise.
246 (cortex_a15_neon_int_5): Likewise.
247 (cortex_a15_neon_vqneg_vqabs): Likewise.
248 (cortex_a15_neon_vmov): Likewise.
249 (cortex_a15_neon_vaba): Likewise.
250 (cortex_a15_neon_vaba_qqq): Likewise.
251 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
252 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
253 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
254 Likewise.
255 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
256 (cortex_a15_neon_mla_qqq_8_16): Likewise.
257 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
258 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
259 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
260 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
261 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
262 (cortex_a15_neon_shift_1): Likewise.
263 (cortex_a15_neon_shift_2): Likewise.
264 (cortex_a15_neon_shift_3): Likewise.
265 (cortex_a15_neon_vshl_ddd): Likewise.
266 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
267 (cortex_a15_neon_vsra_vrsra): Likewise.
268 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
269 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
270 (cortex_a15_neon_bp_3cycle): Likewise.
271 (cortex_a15_neon_ldm_2): Likewise.
272 (cortex_a15_neon_stm_2): Likewise.
273 (cortex_a15_neon_mcr): Likewise.
274 (cortex_a15_neon_mrc): Likewise.
275 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
276 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
277 (cortex_a15_neon_fp_vmul_ddd): Likewise.
278 (cortex_a15_neon_fp_vmul_qqd): Likewise.
279 (cortex_a15_neon_fp_vmla_ddd): Likewise.
280 (cortex_a15_neon_fp_vmla_qqq): Likewise.
281 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
282 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
283 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
284 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
285 (cortex_a15_neon_bp_simple): Likewise.
286 (cortex_a15_neon_bp_2cycle): Likewise.
287 (cortex_a15_neon_bp_3cycle): Likewise.
288 (cortex_a15_neon_vld1_1_2_regs): Likewise.
289 (cortex_a15_neon_vld1_3_4_regs): Likewise.
290 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
291 (cortex_a15_neon_vld2_4_regs): Likewise.
292 (cortex_a15_neon_vld3_vld4): Likewise.
293 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
294 (cortex_a15_neon_vst1_3_4_regs): Likewise.
295 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
296 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
297 (cortex_a15_neon_vst3_vst4): Rename to...
298 (cortex_a15_neon_vst4): This, update for new attributes.
299 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
300 (cortex_a15_neon_vld3_vld4_lane): Likewise.
301 (cortex_a15_neon_vst1_vst2_lane): Likewise.
302 (cortex_a15_neon_vst3_vst4_lane): Likewise.
303 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
304 (cortex_a15_neon_ldm_2): Likewise.
305 (cortex_a15_neon_stm_2): Likewise.
306 (cortex_a15_neon_mcr): Likewise.
307 (cortex_a15_neon_mcr_2_mcrr): Likewise.
308 (cortex_a15_neon_mrc): Likewise.
309 (cortex_a15_neon_mrrc): Likewise.
310
311 (cortex_a15_neon_abd): New.
312 (cortex_a15_neon_abd_q): Likewise.
313 (cortex_a15_neon_aba): Likewise.
314 (cortex_a15_neon_aba_q): Likewise.
315 (cortex_a15_neon_acc): Likewise.
316 (cortex_a15_neon_acc_q): Likewise.
317 (cortex_a15_neon_arith_basic): Likewise.
318 (cortex_a15_neon_arith_complex): Likewise.
319 (cortex_a15_neon_multiply): Likewise.
320 (cortex_a15_neon_multiply_q): Likewise.
321 (cortex_a15_neon_mla): Likewise.
322 (cortex_a15_neon_mla_q): Likewise.
323 (cortex_a15_neon_sat_mla_long): Likewise.
324 (cortex_a15_neon_shift_acc): Likewise.
325 (cortex_a15_neon_shift_imm_basic): Likewise.
326 (cortex_a15_neon_shift_imm_complex): Likewise.
327 (cortex_a15_neon_shift_reg_basic): Likewise.
328 (cortex_a15_neon_shift_reg_basic_q): Likewise.
329 (cortex_a15_neon_shift_reg_complex): Likewise.
330 (cortex_a15_neon_shift_reg_complex_q): Likewise.
331 (cortex_a15_neon_fp_negabs): Likewise
332 (cortex_a15_neon_fp_arith): Likewise
333 (cortex_a15_neon_fp_arith_q): Likewise
334 (cortex_a15_neon_fp_cvt_int): Likewise
335 (cortex_a15_neon_fp_cvt_int_q): Likewise
336 (cortex_a15_neon_fp_cvt_16): Likewise
337 (cortex_a15_neon_fp_mul): Likewise
338 (cortex_a15_neon_fp_mul_q): Likewise
339 (cortex_a15_neon_fp_mla): Likewise
340 (cortex_a15_neon_fp_mla_q): Likewise
341 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
342 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
343 (cortex_a15_neon_bitops): Likewise.
344 (cortex_a15_neon_bitops_q): Likewise.
345 (cortex_a15_neon_from_gp): Likewise.
346 (cortex_a15_neon_from_gp_q): Likewise.
347 (cortex_a15_neon_tbl3_tbl4): Likewise.
348 (cortex_a15_neon_zip_q): Likewise.
349 (cortex_a15_neon_to_gp): Likewise.
350 (cortex_a15_neon_load_a): Likewise.
351 (cortex_a15_neon_load_b): Likewise.
352 (cortex_a15_neon_load_c): Likewise.
353 (cortex_a15_neon_load_d): Likewise.
354 (cortex_a15_neon_load_e): Likewise.
355 (cortex_a15_neon_load_f): Likewise.
356 (cortex_a15_neon_store_a): Likewise.
357 (cortex_a15_neon_store_b): Likewise.
358 (cortex_a15_neon_store_c): Likewise.
359 (cortex_a15_neon_store_d): Likewise.
360 (cortex_a15_neon_store_e): Likewise.
361 (cortex_a15_neon_store_f): Likewise.
362 (cortex_a15_neon_store_g): Likewise.
363 (cortex_a15_neon_store_h): Likewise.
364 (cortex_a15_vfp_to_from_gp): Likewise.
365
bc5e57e3
JG		 3662013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
367
368 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
369
370 (cortex_a9_neon_vshl_ddd): Remove.
371 (cortex_a9_neon_vst3_vst4): Likewise.
372 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
373
374 (cortex_a9_neon_bit_ops_q): New.
375
376 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
377 (cortex_a9_neon_int_2): Likewise.
378 (cortex_a9_neon_int_3): Likewise.
379 (cortex_a9_neon_int_4): Likewise.
380 (cortex_a9_neon_int_5): Likewise.
381 (cortex_a9_neon_vqneg_vqabs): Likewise.
382 (cortex_a9_neon_vmov): Likewise.
383 (cortex_a9_neon_vaba): Likewise.
384 (cortex_a9_neon_vaba_qqq): Likewise.
385 (cortex_a9_neon_shift_1): Likewise.
386 (cortex_a9_neon_shift_2): Likewise.
387 (cortex_a9_neon_shift_3): Likewise.
388 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
389 (cortex_a9_neon_vsra_vrsra): Likewise.
390 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
391 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
392 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
393 Likewise.
394 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
395 (cortex_a9_neon_mla_qqq_8_16): Likewise.
396 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
397 Likewise.
398 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
399 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
400 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
401 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
402 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
403 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
404 (cortex_a9_neon_fp_vsum): Likewise.
405 (cortex_a9_neon_fp_vmul_ddd): Likewise.
406 (cortex_a9_neon_fp_vmul_qqd): Likewise.
407 (cortex_a9_neon_fp_vmla_ddd): Likewise.
408 (cortex_a9_neon_fp_vmla_qqq): Likewise.
409 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
410 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
411 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
412 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
413 (cortex_a9_neon_bp_simple): Likewise.
414 (cortex_a9_neon_bp_2cycle): Likewise.
415 (cortex_a9_neon_bp_3cycle): Likewise.
416 (cortex_a9_neon_ldr): Likewise.
417 (cortex_a9_neon_str): Likewise.
418 (cortex_a9_neon_vld1_1_2_regs): Likewise.
419 (cortex_a9_neon_vld1_3_4_regs): Likewise.
420 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
421 (cortex_a9_neon_vld2_4_regs): Likewise.
422 (cortex_a9_neon_vld3_vld4): Likewise.
423 (cortex_a9_neon_vld1_vld2_lane): Likewise.
424 (cortex_a9_neon_vld3_vld4_lane): Likewise.
425 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
426 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
427 (cortex_a9_neon_vst1_3_4_regs): Likewise.
428 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
429 (cortex_a9_neon_vst1_vst2_lane): Likewise.
430 (cortex_a9_neon_vst3_vst4_lane): Likewise.
431 (cortex_a9_neon_mcr): Likewise.
432 (cortex_a9_neon_mcr_2_mcrr): Likewise.
433 (cortex_a9_neon_mrc): Likewise.
434 (cortex_a9_neon_mrrc): Likewise.
435
9deda8cb
JG		 4362013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
437
438 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
439
440 (cortex_a8_neon_vshl_ddd): Remove.
441 (cortex_a8_neon_vst3_vst4): Likewise.
442 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
443
444 (cortex_a8_neon_bit_ops_q): New.
445
446 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
447 (cortex_a8_neon_int_2): Likewise..
448 (cortex_a8_neon_int_3): Likewise.
449 (cortex_a8_neon_int_5): Likewise.
450 (cortex_a8_neon_vqneg_vqabs): Likewise.
451 (cortex_a8_neon_int_4): Likewise.
452 (cortex_a8_neon_vaba): Likewise.
453 (cortex_a8_neon_vaba_qqq): Likewise.
454 (cortex_a8_neon_shift_1): Likewise.
455 (cortex_a8_neon_shift_2): Likewise.
456 (cortex_a8_neon_shift_3): Likewise.
457 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
458 (cortex_a8_neon_vsra_vrsra): Likewise.
459 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
460 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
461 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
462 Likewise.
463 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
464 (cortex_a8_neon_mla_qqq_8_16): Likewise.
465 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
466 Likewise.
467 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
468 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
469 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
470 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
471 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
472 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
473 (cortex_a8_neon_fp_vsum): Likewise.
474 (cortex_a8_neon_fp_vmul_ddd): Likewise.
475 (cortex_a8_neon_fp_vmul_qqd): Likewise.
476 (cortex_a8_neon_fp_vmla_ddd): Likewise.
477 (cortex_a8_neon_fp_vmla_qqq): Likewise.
478 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
479 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
480 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
481 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
482 (cortex_a8_neon_bp_simple): Likewise.
483 (cortex_a8_neon_bp_2cycle): Likewise.
484 (cortex_a8_neon_bp_3cycle): Likewise.
485 (cortex_a8_neon_ldr): Likewise.
486 (cortex_a8_neon_str): Likewise.
487 (cortex_a8_neon_vld1_1_2_regs): Likewise.
488 (cortex_a8_neon_vld1_3_4_regs): Likewise.
489 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
490 (cortex_a8_neon_vld2_4_regs): Likewise.
491 (cortex_a8_neon_vld3_vld4): Likewise.
492 (cortex_a8_neon_vld1_vld2_lane): Likewise.
493 (cortex_a8_neon_vld3_vld4_lane): Likewise.
494 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
495 (cortex_a8_neon_vst1_3_4_regs): Likewise.
496 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
497 (cortex_a8_neon_vst1_vst2_lane): Likewise.
498 (cortex_a8_neon_vst3_vst4_lane): Likewise.
499 (cortex_a8_neon_mcr): Likewise.
500 (cortex_a8_neon_mcr_2_mcrr): Likewise.
501 (cortex_a8_neon_mrc): Likewise.
502 (cortex_a8_neon_mrrc): Likewise.
503
0f686aa9
JG		 5042013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
505
506 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
507 (fp): New.
508 * config/aarch64/aarch64-simd.md (neon_type): Remove.
509 (aarch64_simd_dup<mode>): Add "type" attribute.
510 (aarch64_dup_lane<mode>): Likewise.
511 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
512 (*aarch64_simd_mov<mode>): Likewise.
513 (aarch64_simd_mov_from_<mode>low): Likewise.
514 (aarch64_simd_mov_from_<mode>high): Likewise.
515 (orn<mode>3): Likewise.
516 (bic<mode>3): Likewise.
517 (add<mode>3): Likewise.
518 (sub<mode>3): Likewise.
519 (mul<mode>3): Likewise.
520 (*aarch64_mul3_elt<mode>): Likewise.
521 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
522 (*aarch64_mul3_elt_to_128df): Likewise.
523 (*aarch64_mul3_elt_to_64v2df): Likewise.
524 (neg<mode>2): Likewise.
525 (abs<mode>2): Likewise.
526 (abd<mode>_3): Likewise.
527 (aba<mode>_3): Likewise.
528 (fabd<mode>_3): Likewise.
529 (*fabd_scalar<mode>3): Likewise.
530 (and<mode>3): Likewise.
531 (ior<mode>3): Likewise.
532 (xor<mode>3): Likewise.
533 (one_cmpl<mode>2): Likewise.
534 (aarch64_simd_vec_set<mode>): Likewise.
535 (aarch64_simd_lshr<mode>): Likewise.
536 (aarch64_simd_ashr<mode>): Likewise.
537 (aarch64_simd_imm_shl<mode>): Likewise.
538 (aarch64_simd_reg_sshl<mode): Likewise.
539 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
540 (aarch64_simd_reg_shl<mode>_signed): Likewise.
541 (aarch64_simd_vec_setv2di): Likewise.
542 (aarch64_simd_vec_set<mode>): Likewise.
543 (aarch64_mla<mode>): Likewise.
544 (*aarch64_mla_elt<mode>): Likewise.
545 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
546 (aarch64_mls<mode>): Likewise.
547 (*aarch64_mls_elt<mode>): Likewise.
548 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
549 (<su><maxmin><mode>3): Likewise.
550 (move_lo_quad_<mode>): Likewise.
551 (aarch64_simd_move_hi_quad_<mode>): Likewise.
552 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
553 (vec_pack_trunc_<mode>): Likewise.
554 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
555 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
556 (*aarch64_<su>mlal_lo<mode>): Likewise.
557 (*aarch64_<su>mlal_hi<mode>): Likewise.
558 (*aarch64_<su>mlsl_lo<mode>): Likewise.
559 (*aarch64_<su>mlsl_hi<mode>): Likewise.
560 (*aarch64_<su>mlal<mode>): Likewise.
561 (*aarch64_<su>mlsl<mode>): Likewise.
562 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
563 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
564 (add<mode>3): Likewise.
565 (sub<mode>3): Likewise.
566 (mul<mode>3): Likewise.
567 (div<mode>3): Likewise.
568 (neg<mode>2): Likewise.
569 (abs<mode>2): Likewise.
570 (fma<mode>4): Likewise.
571 (*aarch64_fma4_elt<mode>): Likewise.
572 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
573 (*aarch64_fma4_elt_to_128df): Likewise.
574 (*aarch64_fma4_elt_to_64v2df): Likewise.
575 (fnma<mode>4): Likewise.
576 (*aarch64_fnma4_elt<mode>): Likewise.
577 (*aarch64_fnma4_elt_<vswap_width_name><mode>
578 (*aarch64_fnma4_elt_to_128df): Likewise.
579 (*aarch64_fnma4_elt_to_64v2df): Likewise.
580 (<frint_pattern><mode>2): Likewise.
581 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
582 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
583 (vec_unpacks_lo_v4sf): Likewise.
584 (aarch64_float_extend_lo_v2df): Likewise.
585 (vec_unpacks_hi_v4sf): Likewise.
586 (aarch64_float_truncate_lo_v2sf): Likewise.
587 (aarch64_float_truncate_hi_v4sf): Likewise.
588 (aarch64_vmls<mode>): Likewise.
589 (<su><maxmin><mode>3): Likewise.
590 (<maxmin_uns><mode>3): Likewise.
591 (reduc_<sur>plus_<mode>): Likewise.
592 (reduc_<sur>plus_v2di): Likewise.
593 (reduc_<sur>plus_v2si): Likewise.
594 (reduc_<sur>plus_<mode>): Likewise.
595 (aarch64_addpv4sf): Likewise.
596 (clz<mode>2): Likewise.
597 (reduc_<maxmin_uns>_<mode>): Likewise.
598 (reduc_<maxmin_uns>_v2di): Likewise.
599 (reduc_<maxmin_uns>_v2si): Likewise.
600 (reduc_<maxmin_uns>_<mode>): Likewise.
601 (reduc_<maxmin_uns>_v4sf): Likewise.
602 (aarch64_simd_bsl<mode>_internal): Likewise.
603 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
604 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
605 (aarch64_get_lane<mode>): Likewise.
606 (*aarch64_combinez<mode>): Likewise.
607 (aarch64_combine<mode>): Likewise.
608 (aarch64_simd_combine<mode>): Likewise.
609 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
610 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
611 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
612 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
613 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
614 (aarch64_<sur>h<addsub><mode>): Likewise.
615 (aarch64_<sur><addsub>hn<mode>): Likewise.
616 (aarch64_<sur><addsub>hn2<mode>): Likewise.
617 (aarch64_pmul<mode>): Likewise.
618 (aarch64_<su_optab><optab><mode>): Likewise.
619 (aarch64_<sur>qadd<mode>): Likewise.
620 (aarch64_sqmovun<mode>): Likewise.
621 (aarch64_<sur>qmovn<mode>): Likewise.
622 (aarch64_s<optab><mode>): Likewise.
623 (aarch64_sq<r>dmulh<mode>): Likewise.
624 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
625 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
626 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
627 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
628 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
629 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
630 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
631 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
632 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
633 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
634 (aarch64_sqdmull<mode>): Likewise.
635 (aarch64_sqdmull_lane<mode>_internal): Likewise.
636 (aarch64_sqdmull_n<mode>): Likewise.
637 (aarch64_sqdmull2<mode>_internal): Likewise.
638 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
639 (aarch64_sqdmull2_n<mode>_internal): Likewise.
640 (aarch64_<sur>shl<mode>): Likewise.
641 (aarch64_<sur>q<r>shl<mode>
642 (aarch64_<sur>shll_n<mode>): Likewise.
643 (aarch64_<sur>shll2_n<mode>): Likewise.
644 (aarch64_<sur>shr_n<mode>): Likewise.
645 (aarch64_<sur>sra_n<mode>): Likewise.
646 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
647 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
648 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
649 (aarch64_cm<optab><mode>): Likewise.
650 (aarch64_cm<optab>di): Likewise.
651 (aarch64_cm<optab><mode>): Likewise.
652 (aarch64_cm<optab>di): Likewise.
653 (aarch64_cmtst<mode>): Likewise.
654 (aarch64_cmtstdi): Likewise.
655 (aarch64_cm<optab><mode>): Likewise.
656 (*aarch64_fac<optab><mode>): Likewise.
657 (aarch64_addp<mode>): Likewise.
658 (aarch64_addpdi): Likewise.
659 (sqrt<mode>2): Likewise.
660 (vec_load_lanesoi<mode>): Likewise.
661 (vec_store_lanesoi<mode>): Likewise.
662 (vec_load_lanesci<mode>): Likewise.
663 (vec_store_lanesci<mode>): Likewise.
664 (vec_load_lanesxi<mode>): Likewise.
665 (vec_store_lanesxi<mode>): Likewise.
666 (*aarch64_mov<mode>): Likewise.
667 (aarch64_ld2<mode>_dreg): Likewise.
668 (aarch64_ld2<mode>_dreg): Likewise.
669 (aarch64_ld3<mode>_dreg): Likewise.
670 (aarch64_ld3<mode>_dreg): Likewise.
671 (aarch64_ld4<mode>_dreg): Likewise.
672 (aarch64_ld4<mode>_dreg): Likewise.
673 (aarch64_tbl1<mode>): Likewise.
674 (aarch64_tbl2v16qi): Likewise.
675 (aarch64_combinev16qi): Likewise.
676 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
677 (aarch64_st2<mode>_dreg): Likewise.
678 (aarch64_st2<mode>_dreg): Likewise.
679 (aarch64_st3<mode>_dreg): Likewise.
680 (aarch64_st3<mode>_dreg): Likewise.
681 (aarch64_st4<mode>_dreg): Likewise.
682 (aarch64_st4<mode>_dreg): Likewise.
683 (*aarch64_simd_ld1r<mode>): Likewise.
684 (aarch64_frecpe<mode>): Likewise.
685 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
686 (aarch64_frecps<mode>): Likewise.
687
f7379e5e
JG		 6882013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
689
690 * config/arm/iterators.md (V_elem_ch): New.
691 (q): Likewise.
692 (VQH_type): Likewise.
693 * config/arm/arm.md (is_neon_type): New.
694 (conds): Use is_neon_type.
695 (anddi3_insn): Update type attribute.
696 (xordi3_insn): Likewise.
697 (one_cmpldi2): Likewise.
698 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
699 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
700 (*movmisalign<mode>_neon_store): Likewise.
701 (*movmisalign<mode>_neon_load): Likewise.
702 (vec_set<mode>_internal): Likewise.
703 (vec_set<mode>_internal): Likewise.
704 (vec_setv2di_internal): Likewise.
705 (vec_extract<mode>): Likewise.
706 (vec_extract<mode>): Likewise.
707 (vec_extractv2di): Likewise.
708 (*add<mode>3_neon): Likewise.
709 (adddi3_neon): Likewise.
710 (*sub<mode>3_neon): Likewise.
711 (subdi3_neon): Likewise.
712 (fma<VCVTF:mode>4): Likewise.
713 (fma<VCVTF:mode>4_intrinsic): Likewise.
714 (*fmsub<VCVTF:mode>4): Likewise.
715 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
716 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
717 (ior<mode>3): Likewise.
718 (and<mode>3): Likewise.
719 (orn<mode>3_neon): Likewise.
720 (orndi3_neon): Likewise.
721 (bic<mode>3_neon): Likewise.
722 (bicdi3_neon): Likewise.
723 (xor<mode>3): Likewise.
724 (one_cmpl<mode>2): Likewise.
725 (abs<mode>2): Likewise.
726 (neg<mode>2): Likewise.
727 (negdi2_neon): Likewise.
728 (*umin<mode>3_neon): Likewise.
729 (*umax<mode>3_neon): Likewise.
730 (*smin<mode>3_neon): Likewise.
731 (*smax<mode>3_neon): Likewise.
732 (vashl<mode>3): Likewise.
733 (vashr<mode>3_imm): Likewise.
734 (vlshr<mode>3_imm): Likewise.
735 (ashl<mode>3_signed): Likewise.
736 (ashl<mode>3_unsigned): Likewise.
737 (neon_load_count): Likewise.
738 (ashldi3_neon_noclobber): Likewise.
739 (ashldi3_neon): Likewise.
740 (signed_shift_di3_neon): Likewise.
741 (unsigned_shift_di3_neon): Likewise.
742 (ashrdi3_neon_imm_noclobber): Likewise.
743 (lshrdi3_neon_imm_noclobber): Likewise.
744 (<shift>di3_neon): Likewise.
745 (widen_ssum<mode>3): Likewise.
746 (widen_usum<mode>3): Likewise.
747 (quad_halves_<code>v4si): Likewise.
748 (quad_halves_<code>v4sf): Likewise.
749 (quad_halves_<code>v8hi): Likewise.
750 (quad_halves_<code>v16qi): Likewise.
751 (reduc_splus_v2di): Likewise.
752 (neon_vpadd_internal<mode>): Likewise.
753 (neon_vpsmin<mode>): Likewise.
754 (neon_vpsmax<mode>): Likewise.
755 (neon_vpumin<mode>): Likewise.
756 (neon_vpumax<mode>): Likewise.
757 (*ss_add<mode>_neon): Likewise.
758 (*us_add<mode>_neon): Likewise.
759 (*ss_sub<mode>_neon): Likewise.
760 (*us_sub<mode>_neon): Likewise.
761 (neon_vadd<mode>_unspec): Likewise.
762 (neon_vaddl<mode>): Likewise.
763 (neon_vaddw<mode>): Likewise.
764 (neon_vhadd<mode>): Likewise.
765 (neon_vqadd<mode>): Likewise.
766 (neon_vaddhn<mode>): Likewise.
767 (neon_vmul<mode>): Likewise.
768 (neon_vfms<VCVTF:mode>): Likewise.
769 (neon_vmlal<mode>): Likewise.
770 (neon_vmls<mode>): Likewise.
771 (neon_vmlsl<mode>): Likewise.
772 (neon_vqdmulh<mode>): Likewise.
773 (neon_vqdmlal<mode>): Likewise.
774 (neon_vqdmlsl<mode>): Likewise.
775 (neon_vmull<mode>): Likewise.
776 (neon_vqdmull<mode>): Likewise.
777 (neon_vsub<mode>_unspec): Likewise.
778 (neon_vsubl<mode>): Likewise.
779 (neon_vsubw<mode>): Likewise.
780 (neon_vqsub<mode>): Likewise.
781 (neon_vhsub<mode>): Likewise.
782 (neon_vsubhn<mode>): Likewise.
783 (neon_vceq<mode>): Likewise.
784 (neon_vcge<mode>): Likewise.
785 (neon_vcgeu<mode>): Likewise.
786 (neon_vcgt<mode>): Likewise.
787 (neon_vcgtu<mode>): Likewise.
788 (neon_vcle<mode>): Likewise.
789 (neon_vclt<mode>): Likewise.
790 (neon_vcage<mode>): Likewise.
791 (neon_vcagt<mode>): Likewise.
792 (neon_vtst<mode>): Likewise.
793 (neon_vabd<mode>): Likewise.
794 (neon_vabdl<mode>): Likewise.
795 (neon_vaba<mode>): Likewise.
796 (neon_vabal<mode>): Likewise.
797 (neon_vmax<mode>): Likewise.
798 (neon_vmin<mode>): Likewise.
799 (neon_vpaddl<mode>): Likewise.
800 (neon_vpadal<mode>): Likewise.
801 (neon_vpmax<mode>): Likewise.
802 (neon_vpmin<mode>): Likewise.
803 (neon_vrecps<mode>): Likewise.
804 (neon_vrsqrts<mode>): Likewise.
805 (neon_vqabs<mode>): Likewise.
806 (neon_vqneg<mode>): Likewise.
807 (neon_vcls<mode>): Likewise.
808 (clz<mode>2): Likewise.
809 (popcount<mode>2): Likewise.
810 (neon_vrecpe<mode>): Likewise.
811 (neon_vrsqrte<mode>): Likewise.
812 (neon_vget_lane<mode>_sext_internal): Likewise.
813 (neon_vget_lane<mode>_zext_internal): Likewise.
814 (neon_vdup_n<mode>): Likewise.
815 (neon_vdup_n<mode>): Likewise.
816 (neon_vdup_nv2di): Likewise.
817 (neon_vdup_lane<mode>_interal): Likewise.
818 (*neon_vswp<mode>): Likewise.
819 (neon_vcombine<mode>): Likewise.
820 (float<mode><V_cvtto>2): Likewise.
821 (floatuns<mode><V_cvtto>2): Likewise.
822 (fix_trunc<mode><V_cvtto>2): Likewise.
823 (fixuns_trunc<mode><V_cvtto>2
824 (neon_vcvt<mode>): Likewise.
825 (neon_vcvt<mode>): Likewise.
826 (neon_vcvtv4sfv4hf): Likewise.
827 (neon_vcvtv4hfv4sf): Likewise.
828 (neon_vcvt_n<mode>): Likewise.
829 (neon_vcvt_n<mode>): Likewise.
830 (neon_vmovn<mode>): Likewise.
831 (neon_vqmovn<mode>): Likewise.
832 (neon_vqmovun<mode>): Likewise.
833 (neon_vmovl<mode>): Likewise.
834 (neon_vmul_lane<mode>): Likewise.
835 (neon_vmul_lane<mode>): Likewise.
836 (neon_vmull_lane<mode>): Likewise.
837 (neon_vqdmull_lane<mode>): Likewise.
838 (neon_vqdmulh_lane<mode>): Likewise.
839 (neon_vqdmulh_lane<mode>): Likewise.
840 (neon_vmla_lane<mode>): Likewise.
841 (neon_vmla_lane<mode>): Likewise.
842 (neon_vmlal_lane<mode>): Likewise.
843 (neon_vqdmlal_lane<mode>): Likewise.
844 (neon_vmls_lane<mode>): Likewise.
845 (neon_vmls_lane<mode>): Likewise.
846 (neon_vmlsl_lane<mode>): Likewise.
847 (neon_vqdmlsl_lane<mode>): Likewise.
848 (neon_vext<mode>): Likewise.
849 (neon_vrev64<mode>): Likewise.
850 (neon_vrev32<mode>): Likewise.
851 (neon_vrev16<mode>): Likewise.
852 (neon_vbsl<mode>_internal): Likewise.
853 (neon_vshl<mode>): Likewise.
854 (neon_vqshl<mode>): Likewise.
855 (neon_vshr_n<mode>): Likewise.
856 (neon_vshrn_n<mode>): Likewise.
857 (neon_vqshrn_n<mode>): Likewise.
858 (neon_vqshrun_n<mode>): Likewise.
859 (neon_vshl_n<mode>): Likewise.
860 (neon_vqshl_n<mode>): Likewise.
861 (neon_vqshlu_n<mode>): Likewise.
862 (neon_vshll_n<mode>): Likewise.
863 (neon_vsra_n<mode>): Likewise.
864 (neon_vsri_n<mode>): Likewise.
865 (neon_vsli_n<mode>): Likewise.
866 (neon_vtbl1v8qi): Likewise.
867 (neon_vtbl2v8qi): Likewise.
868 (neon_vtbl3v8qi): Likewise.
869 (neon_vtbl4v8qi): Likewise.
870 (neon_vtbl1v16qi): Likewise.
871 (neon_vtbl2v16qi): Likewise.
872 (neon_vcombinev16qi): Likewise.
873 (neon_vtbx1v8qi): Likewise.
874 (neon_vtbx2v8qi): Likewise.
875 (neon_vtbx3v8qi): Likewise.
876 (neon_vtbx4v8qi): Likewise.
877 (*neon_vtrn<mode>_insn): Likewise.
878 (*neon_vzip<mode>_insn): Likewise.
879 (*neon_vuzp<mode>_insn): Likewise.
880 (neon_vld1<mode>): Likewise.
881 (neon_vld1_lane<mode>): Likewise.
882 (neon_vld1_lane<mode>): Likewise.
883 (neon_vld1_dup<mode>): Likewise.
884 (neon_vld1_dup<mode>): Likewise.
885 (neon_vld1_dupv2di): Likewise.
886 (neon_vst1<mode>): Likewise.
887 (neon_vst1_lane<mode>): Likewise.
888 (neon_vst1_lane<mode>): Likewise.
889 (neon_vld2<mode>): Likewise.
890 (neon_vld2<mode>): Likewise.
891 (neon_vld2_lane<mode>): Likewise.
892 (neon_vld2_lane<mode>): Likewise.
893 (neon_vld2_dup<mode>): Likewise.
894 (neon_vst2<mode>): Likewise.
895 (neon_vst2<mode>): Likewise.
896 (neon_vst2_lane<mode>): Likewise.
897 (neon_vst2_lane<mode>): Likewise.
898 (neon_vld3<mode>): Likewise.
899 (neon_vld3qa<mode>): Likewise.
900 (neon_vld3qb<mode>): Likewise.
901 (neon_vld3_lane<mode>): Likewise.
902 (neon_vld3_lane<mode>): Likewise.
903 (neon_vld3_dup<mode>): Likewise.
904 (neon_vst3<mode>): Likewise.
905 (neon_vst3qa<mode>): Likewise.
906 (neon_vst3qb<mode>): Likewise.
907 (neon_vst3_lane<mode>): Likewise.
908 (neon_vst3_lane<mode>): Likewise.
909 (neon_vld4<mode>): Likewise.
910 (neon_vld4qa<mode>): Likewise.
911 (neon_vld4qb<mode>): Likewise.
912 (neon_vld4_lane<mode>): Likewise.
913 (neon_vld4_lane<mode>): Likewise.
914 (neon_vld4_dup<mode>): Likewise.
915 (neon_vst4<mode>): Likewise.
916 (neon_vst4qa<mode>): Likewise.
917 (neon_vst4qb<mode>): Likewise.
918 (neon_vst4_lane<mode>): Likewise.
919 (neon_vst4_lane<mode>): Likewise.
920 (neon_vec_unpack<US>_lo_<mode>): Likewise.
921 (neon_vec_unpack<US>_hi_<mode>): Likewise.
922 (neon_vec_<US>mult_lo_<mode>): Likewise.
923 (neon_vec_<US>mult_hi_<mode>): Likewise.
924 (neon_vec_<US>shiftl_<mode>): Likewise.
925 (neon_unpack<US>_<mode>): Likewise.
926 (neon_vec_<US>mult_<mode>): Likewise.
927 (vec_pack_trunc_<mode>): Likewise.
928 (neon_vec_pack_trunc_<mode>): Likewise.
929 (neon_vabd<mode>_2): Likewise.
930 (neon_vabd<mode>_3): Likewise.
931
a9e66678
JG		 9322013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
933
934 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
935 (load_pair): Update type attribute.
936 (store_pair): Update type attribute.
937 * config/aarch64/iterators.md (q): New.
938
20445ed6
JG		 9392013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
940
941 * config/arm/types.md: Add new types for Neon insns.
942
afb4ac68
AI		 9432013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
944 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
945 Sergey Lega <sergey.s.lega@intel.com>
946 Anna Tikhonova <anna.tikhonova@intel.com>
947 Ilya Tocar <ilya.tocar@intel.com>
948 Andrey Turetskiy <andrey.turetskiy@intel.com>
949 Ilya Verbin <ilya.verbin@intel.com>
950 Kirill Yukhin <kirill.yukhin@intel.com>
951 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
952
953 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
954 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
955 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
956 (rcp14<mode>): New.
957 (srcp14<mode>): Ditto.
958 (rsqrt14<mode>): Ditto.
959 (rsqrt14<mode>): Ditto.
960 (avx512f_vmscalef<mode>): Ditto.
961 (avx512f_scalef<mode>): Ditto.
962 (avx512f_getexp<mode>): Ditto.
963 (avx512f_sgetexp<mode>): Ditto.
964 (avx512f_fixupimm<mode>): Ditto.
965 (avx512f_sfixupimm<mode>): Ditto.
966 (avx512f_rndscale<mode>): Ditto.
967 (*avx512er_exp2<mode>): Ditto.
968 (*avx512er_rcp28<mode>): Ditto.
969 (avx512er_rsqrt28<mode>): Ditto.
970 (avx512f_getmant<mode>): Ditto.
971 (avx512f_getmant<mode>): Ditto.
972 (avx512f_rndscale<mode>): Fix formatting.
973
a1b499a3
JL		 9742013-10-15 Martin Jambor <mjambor@suse.cz>
975
976 * ipa-utils.h (ipa_edge_within_scc): Declare.
977 * ipa-cp.c (edge_within_scc): Moved...
978 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 979
2e2206fa
AI		 9802013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
981 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
982 Sergey Lega <sergey.s.lega@intel.com>
983 Anna Tikhonova <anna.tikhonova@intel.com>
984 Ilya Tocar <ilya.tocar@intel.com>
985 Andrey Turetskiy <andrey.turetskiy@intel.com>
986 Ilya Verbin <ilya.verbin@intel.com>
987 Kirill Yukhin <kirill.yukhin@intel.com>
988 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
989
990 * config/i386/predicates.md (const_8_to_15_operand): New.
991 (const_16_to_31_operand): Ditto.
992 * config/i386/sse.md (V8FI): New.
993 (V16FI): Ditto.
994 (reduc_splus_v8df): Ditto.
995 (reduc_splus_v16sf): Ditto.
996 (avx512f_vextract<shuffletype>32x4_1): Ditto.
997 (vec_extract_hi_<mode>): Ditto.
998 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
999 (vec_set_lo_<mode>): Ditto.
1000 (vec_set_hi_<mode>): Ditto.
1001 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1002 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1003 (avx512f_pshufd_1): Ditto.
1004 (avx512f_broadcast<mode>): Ditto.
1005 (avx512f_broadcast<mode>): Ditto.
1006 (define_split): Split vec_extract_lo into move.
1007 (ssequartermode): Ditto.
1008 (ssedoublemode): Extened with wider modes.
1009 (vec_extract_lo_<mode>): Ditto.
1010
0fe65b75
AI		 10112013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1012 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1013 Sergey Lega <sergey.s.lega@intel.com>
1014 Anna Tikhonova <anna.tikhonova@intel.com>
1015 Ilya Tocar <ilya.tocar@intel.com>
1016 Andrey Turetskiy <andrey.turetskiy@intel.com>
1017 Ilya Verbin <ilya.verbin@intel.com>
1018 Kirill Yukhin <kirill.yukhin@intel.com>
1019 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1020
1021 * config/i386/predicates.md (register_or_constm1_operand): New.
1022 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1023 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1024 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1025 UNSPEC_SCATTER_PREFETCH
1026 (VI48_512): New.
1027 (avx512f_ucmp<mode>3): Ditto.
1028 (avx512f_vternlog<mode>): Ditto.
1029 (avx512f_align<mode>): Ditto.
1030 (<shift_insn><mode>3): Ditto.
1031 (avx512f_<rotate>v<mode>): Ditto.
1032 (avx512f_<rotate><mode>): Ditto.
1033 (avx512f_eq<mode>3): Ditto.
1034 (avx512f_eq<mode>3_1): Ditto.
1035 (avx512f_gt<mode>3): Ditto.
1036 (avx512f_testm<mode>3): Ditto.
1037 (avx512f_testnm<mode>3): Ditto.
1038 (avx512pf_gatherpf<mode>): Ditto.
1039 (*avx512pf_gatherpf<mode>_mask): Ditto.
1040 (*avx512pf_gatherpf<mode>): Ditto.
1041 (avx512pf_scatterpf<mode>): Ditto.
1042 (*avx512pf_scatterpf<mode>_mask): Ditto.
1043 (*avx512pf_scatterpf<mode>): Ditto.
1044 (avx512f_vec_dup_gpr<mode>): Ditto.
1045 (clz<mode>2): Ditto.
1046 (conflict<mode>): Ditto.
1047 (REDUC_SMINMAX_MODE): Extened with wider modes.
1048 (reduc_<code>_<mode>): Ditto.
1049 (vlshr<mode>3): Ditto.
1050 (vashl<mode>3): Ditto.
1051
ab931c71
AI		 10522013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1053 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1054 Sergey Lega <sergey.s.lega@intel.com>
1055 Anna Tikhonova <anna.tikhonova@intel.com>
1056 Ilya Tocar <ilya.tocar@intel.com>
1057 Andrey Turetskiy <andrey.turetskiy@intel.com>
1058 Ilya Verbin <ilya.verbin@intel.com>
1059 Kirill Yukhin <kirill.yukhin@intel.com>
1060 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1061
1062 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1063 UNSPEC_SCATTER.
1064 (VI48F_512): New.
1065 (avx512fmaskmode): Ditto.
1066 (bcstscalarsuff): Ditto.
1067 (avx512f_blendm<mode>): Ditto.
1068 (cmp_imm_predicate): Ditto.
1069 (avx512f_cmp<mode>3): Ditto.
1070 (avx512f_vec_dup<mode>): Ditto.
1071 (avx512f_vec_dup_mem<mode>): Ditto.
1072 (avx512f_vpermi2var<mode>3): Ditto.
1073 (avx512f_vpermt2var<mode>3): Ditto.
1074 (vec_init<mode>): Ditto.
1075 (avx512f_gathersi<mode>): Ditto.
1076 (*avx512f_gathersi<mode>): Ditto.
1077 (*avx512f_gathersi<mode>_2): Ditto.
1078 (avx512f_gatherdi<mode>): Ditto.
1079 (*avx512f_gatherdi<mode>): Ditto.
1080 (*avx512f_gatherdi<mode>_2): Ditto.
1081 (avx512f_scattersi<mode>): Ditto.
1082 (*avx512f_scattersi<mode>): Ditto.
1083 (avx512f_scatterdi<mode>): Ditto.
1084 (*avx512f_scatterdi<mode>): Ditto.
1085 (sseintprefix): Extened with wider modes.
1086 (VEC_GATHER_IDXSI): Ditto.
1087 (VEC_GATHER_IDXDI): Ditto.
1088 (VEC_GATHER_SRCDI): Ditto.
1089
c9acb877
MGD		 10902013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1091 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1092
1093 * config/arm/t-aprofile: New file.
1094 * config.gcc: Handle --with-multilib-list option.
1095
ecbda484
BS		 10962013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1097
1098 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
1099 r2.
1100
24f161fd
RB		 11012013-10-15 Richard Biener <rguenther@suse.de>
1102
1103 * tree-loop-distribution.c (build_empty_rdg): Inline into
1104 single user.
1105 (rdg_flag_vertex): Inline into single user.
1106 (rdg_flag_vertex_and_dependent): Likewise.
1107 (build_rdg_partition_for_vertex): Remove processed bitmap.
1108 (rdg_build_partitions): Simplify.
1109
5de989ed
RB		 11102013-10-15 Richard Biener <rguenther@suse.de>
1111
1112 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1113 Restructure forwarding through conversions and copies to
1114 avoid performing copy-propagation the wrong way. Adjust
1115 recursion invocations.
1116 (forward_propagate_addr_expr): Add argument stating if we
1117 are recursing from a single-use.
1118 (ssa_forward_propagate_and_combine): Adjust.
1119
47e0da37
DM		 11202013-10-14 David Malcolm <dmalcolm@redhat.com>
1121
1122 * dumpfile.h (gcc::dump_manager): New class, to hold state
1123 relating to dumpfile management.
1124 (get_dump_file_name): Remove in favor of method of dump_manager.
1125 (dump_initialized_p): Likewise.
1126 (dump_start): Likewise.
1127 (dump_finish): Likewise.
1128 (dump_switch_p): Likewise.
1129 (dump_register): Likewise.
1130 (get_dump_file_info): Likewise.
1131 * context.c (gcc::context::context): Construct the dump_manager
1132 instance.
1133 * context.h (gcc::context::get_dumps): New.
1134 (gcc::context::m_dumps): New.
1135 * coverage.c (coverage_init): Port to dump_manager API.
1136 * dumpfile.c (extra_dump_files): Convert to field of
1137 gcc::dump_manager.
1138 (extra_dump_files_in_use): Likewise.
1139 (extra_dump_files_alloced): Likewise.
1140 (gcc::dump_manager::dump_manager): New.
1141 (dump_register): Convert to...
1142 (gcc::dump_manager::dump_register): ...method, replacing
1143 function-static next_dump with m_next_dump field.
1144 (get_dump_file_info): Convert to...
1145 (gcc::dump_manager::get_dump_file_info): ...method.
1146 (get_dump_file_name): Convert to...
1147 (gcc::dump_manager::get_dump_file_name): ...method.
1148 (dump_start): Convert to...
1149 (gcc::dump_manager::dump_start): ...method.
1150 (dump_finish): Convert to...
1151 (gcc::dump_manager::dump_finish): ...method.
1152 (dump_begin): Replace body with...
1153 (gcc::dump_manager::dump_begin): ...new method.
1154 (dump_phase_enabled_p): Convert to...
1155 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1156 (dump_phase_enabled_p): Convert to...
1157 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1158 (dump_initialized_p): Convert to...
1159 (gcc::dump_manager::dump_initialized_p): ...method.
1160 (dump_flag_name): Replace body with...
1161 (gcc::dump_manager::dump_flag_name): ...new method.
1162 (dump_enable_all): Convert to...
1163 (gcc::dump_manager::dump_enable_all): ...new method.
1164 (opt_info_enable_passes): Convert to...
1165 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1166 (dump_switch_p_1): Convert to...
1167 (gcc::dump_manager::dump_switch_p_1): ...new method.
1168 (dump_switch_p): Convert to...
1169 (gcc::dump_manager::dump_switch_p): ...new method.
1170 (opt_info_switch_p): Port to dump_manager API.
1171 (enable_rtl_dump_file): Likewise.
1172 * opts-global.c (handle_common_deferred_options): Port to new
1173 dump_manager API.
1174 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1175 (pass_manager::register_one_dump_file): Likewise.
1176 (pass_manager::register_pass): Likewise.
1177 (pass_init_dump_file): Likewise.
1178 (pass_fini_dump_file): Likewise.
1179 * statistics.c (statistics_early_init): Likewise.
1180
83fd5d11
RB		 11812013-10-14 Richard Biener <rguenther@suse.de>
1182
1183 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1184 iterative_hash_canonical_type, gimple_canonical_type_hash,
1185 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1186 gimple_register_canonical_type, print_gimple_types_stats,
1187 free_gimple_type_tables): Move to lto/lto.c
1188 (gt-gimple.h): Do not include.
1189 * gimple.h (gimple_register_canonical_type,
1190 print_gimple_types_stats, free_gimple_type_tables): Remove.
1191 * Makefile.in (GTFILES): Remove gimple.c.
1192
a6dbd0f7
TS		 11932013-10-14 Travis Snoozy <quandary@remstate.com>
1194
1195 PR target/58716
1196 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1197 scanning for msp430x targets.
1198
b5ccb9ed
EB		 11992013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1200
1201 PR bootstrap/58509
1202 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1203 (registers_ok_for_ldd_peep): Move around.
1204 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1205 * config/sparc/sparc.md (widening peepholes): Use it.
1206
fd7e5084
RB		 12072013-10-14 Richard Biener <rguenther@suse.de>
1208
1209 PR middle-end/58712
1210 PR middle-end/55358
1211 * gimple.c (iterative_hash_canonical_type): Make sure to
1212 record the hash into the correct hashtable slot.
1213
3246afaa
EB		 12142013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1215
1216 PR rtl-optimization/58662
1217 * combine.c (try_combine): Take into account death nodes on I2 when
1218 splitting a PARALLEL of two independent SETs. Fix dump message.
1219
5d30dc5b
OE		 12202013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1221
1222 PR target/51244
1223 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1224 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1225 * config/sh/t-sh (sh_treg_combine.o): New entry.
1226 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1227 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1228 (register_sh_passes): New function. Register sh_treg_combine pass.
1229 (sh_option_override): Invoke it.
1230 (sh_canonicalize_comparison): Handle op0_preserve_value.
1231 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1232 opportunities. Canonicalize branch condition.
1233 (nott): Allow only if pseudos can be created for non-SH2A.
1234
37d1c5db
L		 12352013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1236
1237 PR target/58690
1238 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1239 (ix86_expand_movmem): Replace copy_addr_to_reg with
1240 ix86_copy_addr_to_reg.
1241 (ix86_expand_setmem): Likewise.
1242
b86a70ff
AM		 12432013-10-12 Alexander Monakov <amonakov@ispras.ru>
1244
1245 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1246 provided by ix86_fp_compare_mode instead of CCFPUmode.
1247
6e1d5961
JG		 12482013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1249
1250 * config/aarch64/arm_neon.h
1251 (vtbx<1,3>_<psu>8): Fix register constriants.
1252
ef3cfba2
JL		 12532013-10-11 Jeff Law <law@redhat.com>
1254
1255 PR tree-optimization/58640
adcfd489
UB		 1256 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1257 threading paths that cross over two loop entry points.
ef3cfba2 1258
6e8b7d9c
BS		 12592013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1260
1261 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1262 handle vector float as well.
1263 (*vsx_le_perm_load_v4si): Likewise.
1264 (*vsx_le_perm_store_v2di): Likewise.
1265 (*vsx_le_perm_store_v4si): Likewise.
1266
f200869a
BS		 12672013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1268
1269 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1270 directly to circumvent subtract from splat{31} workaround.
1271 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1272 prototype.
1273 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1274 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1275 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1276 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1277 separate big and little endian logic.
1278 (*altivec_vperm_<mode>_internal): New define_insn.
1279 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1280 separate big and little endian logic.
1281 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1282 (vec_permv16qi): Add little endian logic.
1283
8d28e3fc
MG		 12842013-10-11 Marc Glisse <marc.glisse@inria.fr>
1285
1286 * doc/extend.texi (returns_nonnull): Remove arguments.
1287
3c87b77b
AI		 12882013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1289 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1290 Sergey Lega <sergey.s.lega@intel.com>
1291 Anna Tikhonova <anna.tikhonova@intel.com>
1292 Ilya Tocar <ilya.tocar@intel.com>
1293 Andrey Turetskiy <andrey.turetskiy@intel.com>
1294 Ilya Verbin <ilya.verbin@intel.com>
1295 Kirill Yukhin <kirill.yukhin@intel.com>
1296 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1297
1298 * config/i386/sse.md (VI48F_256_512): New.
1299 (avx2_permvar<mode>): Change to ...
1300 (<avx2_avx512f>_permvar<mode>): This.
1301
50e60d7d
AI		 13022013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1303 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1304 Sergey Lega <sergey.s.lega@intel.com>
1305 Anna Tikhonova <anna.tikhonova@intel.com>
1306 Ilya Tocar <ilya.tocar@intel.com>
1307 Andrey Turetskiy <andrey.turetskiy@intel.com>
1308 Ilya Verbin <ilya.verbin@intel.com>
1309 Kirill Yukhin <kirill.yukhin@intel.com>
1310 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1311
1312 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1313 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1314 * config/i386/sse.md (VI4_AVX): New.
1315 (sf2simodelower): Ditto.
1316 (sse2_cvtps2dq): Change to ...
1317 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1318
f62ce24f
AI		 13192013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1320 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1321 Sergey Lega <sergey.s.lega@intel.com>
1322 Anna Tikhonova <anna.tikhonova@intel.com>
1323 Ilya Tocar <ilya.tocar@intel.com>
1324 Andrey Turetskiy <andrey.turetskiy@intel.com>
1325 Ilya Verbin <ilya.verbin@intel.com>
1326 Kirill Yukhin <kirill.yukhin@intel.com>
1327 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1328
1329 * config/i386/sse.md (V_512): New.
1330 (VI_512): Ditto.
1331 (vcond<V_512:mode><VF_512:mode>): Ditto.
1332 (vcond<V_512:mode><VI_512:mode>): Ditto.
1333 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1334
558d9f79
AI		 13352013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1336 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1337 Sergey Lega <sergey.s.lega@intel.com>
1338 Anna Tikhonova <anna.tikhonova@intel.com>
1339 Ilya Tocar <ilya.tocar@intel.com>
1340 Andrey Turetskiy <andrey.turetskiy@intel.com>
1341 Ilya Verbin <ilya.verbin@intel.com>
1342 Kirill Yukhin <kirill.yukhin@intel.com>
1343 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1344
1345 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1346 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1347 (FMAMODE): Ditto.
1348 (fma<mode>4): Removed condition.
1349 (fms<mode>4): Ditto.
1350 (fnma<mode>4): Ditto.
1351 (fnms<mode>4): Ditto.
1352 (fma4i_fmadd_<mode>): Ditto.
1353 (*fma_fmadd_<mode>): Ditto.
1354 (*fma_fmsub_<mode>): Ditto.
1355 (*fma_fnmadd_<mode>): Ditto.
1356 (*fma_fnmsub_<mode>): Ditto.
1357 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1358 (*fma_fmaddsub_<mode>): Ditto.
1359 (*fma_fmsubadd_<mode>): Ditto.
1360 (*fmai_fmadd_<mode>): Ditto.
1361 (*fmai_fmsub_<mode>): Ditto.
1362 (*fmai_fnmadd_<mode>): Ditto.
1363 (*fmai_fnmsub_<mode>): Ditto.
1364
e8d08206
AI		 13652013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1366 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1367 Sergey Lega <sergey.s.lega@intel.com>
1368 Anna Tikhonova <anna.tikhonova@intel.com>
1369 Ilya Tocar <ilya.tocar@intel.com>
1370 Andrey Turetskiy <andrey.turetskiy@intel.com>
1371 Ilya Verbin <ilya.verbin@intel.com>
1372 Kirill Yukhin <kirill.yukhin@intel.com>
1373 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1374
1375 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1376 (VI124_256): Changed to ...
1377 (VI124_256_48_512): This.
1378 (ssepackmode): Extended with wider modes.
1379 (<code><mode>3): Changed iterator.
1380 (*avx2_<code><mode>3): Ditto.
1381 (vec_pack_trunc_<mode>): Ditto.
1382
16821545
AI		 13832013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1384 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1385 Sergey Lega <sergey.s.lega@intel.com>
1386 Anna Tikhonova <anna.tikhonova@intel.com>
1387 Ilya Tocar <ilya.tocar@intel.com>
1388 Andrey Turetskiy <andrey.turetskiy@intel.com>
1389 Ilya Verbin <ilya.verbin@intel.com>
1390 Kirill Yukhin <kirill.yukhin@intel.com>
1391 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1392
1393 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1394 (VI8F_256_512): Ditto.
1395 (abs<mode>2): Changed iterator.
1396 (avx2_perm<mode>): Changed to ...
1397 (<avx2_avx512f>_perm<mode>): This.
1398 (avx2_perm<mode>_1): Changed to ...
1399 (<avx2_avx512f>_perm<mode>_1): This.
1400
5348cff8
AI		 14012013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1402 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1403 Sergey Lega <sergey.s.lega@intel.com>
1404 Anna Tikhonova <anna.tikhonova@intel.com>
1405 Ilya Tocar <ilya.tocar@intel.com>
1406 Andrey Turetskiy <andrey.turetskiy@intel.com>
1407 Ilya Verbin <ilya.verbin@intel.com>
1408 Kirill Yukhin <kirill.yukhin@intel.com>
1409 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1410
1411 * config/i386/sse.md (VI48_AVX512F): New.
1412 (VI48_AVX2): Changed to ...
1413 (VI48_AVX2_48_AVX512F): This.
1414 (avx2_ashrv<mode>): Changed to ...
1415 (<avx2_avx512f>_ashrv<mode>): This.
1416 (avx2_<shift_insn>v<mode>): Changed to ...
1417 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1418
f5f41d88
AI		 14192013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1420 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1421 Sergey Lega <sergey.s.lega@intel.com>
1422 Anna Tikhonova <anna.tikhonova@intel.com>
1423 Ilya Tocar <ilya.tocar@intel.com>
1424 Andrey Turetskiy <andrey.turetskiy@intel.com>
1425 Ilya Verbin <ilya.verbin@intel.com>
1426 Kirill Yukhin <kirill.yukhin@intel.com>
1427 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1428
1429 * config/i386/sse.md (VI4_AVX512F): New.
1430 (VI8_AVX2_AVX512F): Ditto.
1431 (mul<mode>3): Extended with wider modes.
1432 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1433 (mul<mode>3): Ditto.
1434 (vec_widen_<s>mult_odd_<mode>): Ditto.
1435
3bdf6340
AI		 14362013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1437 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1438 Sergey Lega <sergey.s.lega@intel.com>
1439 Anna Tikhonova <anna.tikhonova@intel.com>
1440 Ilya Tocar <ilya.tocar@intel.com>
1441 Andrey Turetskiy <andrey.turetskiy@intel.com>
1442 Ilya Verbin <ilya.verbin@intel.com>
1443 Kirill Yukhin <kirill.yukhin@intel.com>
1444 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1445
1446 * config/i386/sse.md (VI2_AVX512F): New.
1447 (VI124_AVX512F): Ditto.
1448 (sseunpackmode): Extended with wider modes.
1449 (sseunpackfltmode): Ditto.
1450 (vec_unpacks_float_hi_<mode>): Ditto.
1451 (vec_unpacks_float_lo_<mode>): Ditto.
1452 (vec_unpacku_float_hi_<mode>): Ditto.
1453 (vec_unpacku_float_lo_<mode>): Ditto.
1454 (vec_unpacks_lo_<mode>): Ditto.
1455 (vec_unpacks_hi_<mode>): Ditto.
1456 (vec_unpacku_lo_<mode>): Ditto.
1457 (vec_unpacku_hi_<mode>): Ditto.
1458
67f783cb
AI		 14592013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1460 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1461 Sergey Lega <sergey.s.lega@intel.com>
1462 Anna Tikhonova <anna.tikhonova@intel.com>
1463 Ilya Tocar <ilya.tocar@intel.com>
1464 Andrey Turetskiy <andrey.turetskiy@intel.com>
1465 Ilya Verbin <ilya.verbin@intel.com>
1466 Kirill Yukhin <kirill.yukhin@intel.com>
1467 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1468
1469 * config/i386/i386.md (multdiv): New.
1470 (multdiv_mnemonic): Ditto.
1471 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1472 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1473 (<sse>_vmdiv<mode>3): Removed.
1474
ec5e777c
AI		 14752013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1476 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1477 Sergey Lega <sergey.s.lega@intel.com>
1478 Anna Tikhonova <anna.tikhonova@intel.com>
1479 Ilya Tocar <ilya.tocar@intel.com>
1480 Andrey Turetskiy <andrey.turetskiy@intel.com>
1481 Ilya Verbin <ilya.verbin@intel.com>
1482 Kirill Yukhin <kirill.yukhin@intel.com>
1483 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1484
1485 * config/i386/sse.md (V): Extended with wider modes.
1486 (VF2): Ditto.
1487 (ssehalfvecmode): Ditto.
1488 (i128): Ditto.
1489 (ssepackfltmode): Ditto.
1490 (avx_vec_concat<mode>): Ditto.
1491 (V_256_512): New iterator.
1492 (VF2_512_256): Ditto.
1493 (si2dfmode): New attribute.
1494 (si2dfmodelower): Ditto.
1495 (sf2dfmode): Ditto.
1496 (concat_tg_mode): Ditto.
1497 (floatv4siv4df2): Changed to ...
1498 (float<si2dfmodelower><mode>2): This.
1499 (avx_cvtps2pd256): Changed to ...
1500 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1501 (vec_pack_trunc_v4df): Changed to ...
1502 (vec_pack_trunc_<mode>): This.
1503 (avx_vpermil<mode>): Changed to ...
1504 (<sse2_avx_avx512f>_vpermil<mode>): This.
1505 (<fixsuffix>fix_truncv8dfv8si2): New.
1506 (vec_pack_sfix_trunc_v8df): Ditto.
1507 (avx512f_rndscale<mode>): Ditto.
1508 (avx512f_roundpd512): Ditto.
1509 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1510
a9ccbba2
AI		 15112013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1512 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1513 Sergey Lega <sergey.s.lega@intel.com>
1514 Anna Tikhonova <anna.tikhonova@intel.com>
1515 Ilya Tocar <ilya.tocar@intel.com>
1516 Andrey Turetskiy <andrey.turetskiy@intel.com>
1517 Ilya Verbin <ilya.verbin@intel.com>
1518 Kirill Yukhin <kirill.yukhin@intel.com>
1519 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1520
1521 * config/i386/i386.md (any_fix): New iterator.
1522 (fixsuffix): New attribute.
1523 * config/i386/sse.md (VF1): Extened with wider modes.
1524 (VI): Ditto.
1525 (VI_AVX2): Ditto.
1526 (VI8): Ditto.
1527 (sseintvecmodelower): Ditto.
1528 (ssescalarmode): Ditto.
1529 (ssescalarnum): Ditto.
1530 (VF1_128_256): New.
1531 (ssexmmmode): Ditto.
1532 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1533 (<sse>_rcp<mode>2): Change iterator.
1534 (rsqrt<mode>2): Ditto.
1535 (<sse>_rsqrt<mode>2): Ditto.
1536 (avx2_vec_dup<mode>): Ditto.
1537 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1538 (round<mode>2_sfix): Ditto.
1539 (avx2_pbroadcast<mode>): Ditto.
1540 (*andnot<mode>3): Handle XI mode.
1541 (*<code><mode>3): Ditto.
1542 (AVXTOSSEMODE): Removed.
1543 (avx_vpermil<mode>): Changed to ...
1544 (<sse2_avx_avx512f>_vpermil<mode>): This.
1545
2b1ebb0c
AI		 15462013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1547 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1548 Sergey Lega <sergey.s.lega@intel.com>
1549 Anna Tikhonova <anna.tikhonova@intel.com>
1550 Ilya Tocar <ilya.tocar@intel.com>
1551 Andrey Turetskiy <andrey.turetskiy@intel.com>
1552 Ilya Verbin <ilya.verbin@intel.com>
1553 Kirill Yukhin <kirill.yukhin@intel.com>
1554 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1555
1556 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1557 (<sse>_comi): Ditto.
1558 (<sse>_ucomi): Ditto.
1559 (sse_cvtss2siq_2): Ditto.
1560 (sse2_cvtsd2si): Ditto.
1561 (sse2_cvtsd2siq): Ditto.
1562 (sse2_cvttsd2si): Ditto.
1563 (sse2_cvttsd2siq): Ditto.
1564 (<shift_insn><mode>3): Ditto.
1565 (sse2_cvtsi2sdq): Update constraint and prefix.
1566 (sse_cvtsi2ss): Update prefix.
1567 (sse_cvtsi2ssq): Ditto.
1568
d89124ed
JJ		 15692013-10-11 Jakub Jelinek <jakub@redhat.com>
1570
1571 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1572 ignore internal calls.
1573
e6ad28c3
RB		 15742013-10-11 Richard Biener <rguenther@suse.de>
1575
1576 * tree-pretty-print.c (dump_generic_node): Allow to dump
1577 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1578 (OVF) if TREE_OVERFLOW is set.
1579
d1227447
TS		 15802013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1581
1ac4b8e4
TS		 1582 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1583
7d05cebb
TS		 1584 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1585 * gimple.h: Likewise.
1586
adcfd489 1587 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 1588
1a79fb8a
TS		 1589 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1590 macros.
1591
3afd2873
TS		 1592 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1593
adcfd489 1594 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS		 1595 * configure: Regenerate.
1596
acf0174b
JJ		 15972013-10-11 Jakub Jelinek <jakub@redhat.com>
1598
1599 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1600 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1601 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1602 similarly pp_right_paren.
1603 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1604 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1605 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1606 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1607 clauses.
1608 (omp_declare_simd_clauses_equal,
1609 omp_remove_redundant_declare_simd_attrs): New functions.
1610 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1611 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1612 * tree.h (OMP_LOOP_CHECK): Define.
1613 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1614 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1615 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1616 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1617 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1618 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1619 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1620 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1621 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1622 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1623 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1624 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1625 (omp_remove_redundant_declare_simd_attrs): New prototype.
1626 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1627 GIMPLE_OMP_TEAMS): New codes.
1628 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1629 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1630 fields.
1631 (target_nesting_level): New variable.
1632 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1633 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1634 for collapse > 1 static schedule unless ordered.
1635 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1636 (determine_parallel_type): Adjust get_ws_args_for caller.
1637 (install_var_field): Handle mask & 4 for double indirection.
1638 (scan_sharing_clauses): Ignore shared clause on teams construct.
1639 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1640 (create_omp_child_function): If inside target or declare target
adcfd489 1641 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ		 1642 (find_combined_for): New function.
1643 (scan_omp_parallel): Handle combined loops.
1644 (scan_omp_target, scan_omp_teams): New functions.
1645 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1646 restrictions and set ctx->cancellable for cancellable constructs.
1647 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1648 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1649 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 1650 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b
JJ		 1651 (omp_clause_aligned_alignment): New function.
1652 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1653 on decls.
1654 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1655 on teams constructs. Handle user defined reductions and new
1656 OpenMP 4.0 clauses.
1657 (lower_reduction_clauses): Don't set placeholder to address of ref
1658 if it has already the right type.
1659 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1660 (expand_parallel_call): Use the new non-_start suffixed builtins,
1661 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1662 and GOMP_parallel_end after the call.
1663 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1664 (expand_omp_for_init_counts): Handle combined loops.
1665 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1666 loops.
1667 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1668 end of cancellable loops.
1669 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1670 Likewise. Handle collapse > 1 loops.
1671 (expand_omp_simd): Handle combined loops.
1672 (expand_omp_for): Add inner_stmt argument, adjust callers of
1673 expand_omp_for* functions, use expand_omp_for_static*chunk even
1674 for collapse > 1 unless ordered.
1675 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1676 of cancellable sections.
1677 (expand_omp_single): Remove need_barrier variable, just rely on
1678 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1679 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1680 (expand_omp_atomic_load, expand_omp_atomic_store,
1681 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1682 (expand_omp_target): New function.
1683 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1684 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1685 (build_omp_regions_1): Immediately close region for
1686 GF_OMP_TARGET_KIND_UPDATE.
1687 (maybe_add_implicit_barrier_cancel): New function.
1688 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1689 cancellation.
1690 (lower_omp_single): Likewise. Add clobber after the barrier.
1691 (lower_omp_taskgroup): New function.
1692 (lower_omp_for): Handle combined loops. Adjust
1693 lower_rec_input_clauses caller. Handle cancellation.
1694 (lower_depend_clauses): New function.
1695 (lower_omp_taskreg): Lower depend clauses. Adjust
1696 lower_rec_input_clauses caller. Add clobber after the call. Handle
1697 cancellation.
1698 (lower_omp_target, lower_omp_teams): New functions.
1699 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1700 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1701 and GOMP_cancellation_point calls.
1702 (lower_omp): Fold stmts inside of target region.
1703 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1704 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1705 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1706 (BT_FN_VOID_OMPFN_PTR_UINT,
1707 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1708 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1709 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1710 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1711 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1712 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1713 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1714 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1715 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1716 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1717 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1718 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1719 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1720 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1721 BUILT_IN_GOMP_PARALLEL_END.
1722 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1723 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1724 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1725 GF_OMP_FOR_KIND_DISTRIBUTE.
1726 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1727 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1728 (dump_gimple_omp_return): Print lhs if it has any.
1729 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1730 gimple_omp_atomic_seq_cst_p.
1731 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1732 and GIMPLE_OMP_TEAMS.
1733 * langhooks.c (lhd_omp_mappable_type): New function.
1734 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1735 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1736 hook.
1737 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1738 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1739 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1740 ORT_TARGET.
1741 (struct gimplify_omp_ctx): Add combined_loop field.
1742 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1743 on stmts inside of target region.
1744 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1745 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1746 ORT_TARGET and ORT_TARGET_DATA.
1747 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1748 Handle GOVD_MAP.
1749 (omp_notice_threadprivate_variable): Complain about threadprivate
1750 variables in target region.
1751 (omp_notice_variable): Complain about vars with non-mappable type
1752 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1753 (omp_check_private): Ignore ORT_TARGET* regions.
1754 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1755 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1756 (find_combined_omp_for): New function.
1757 (gimplify_omp_for): Handle gimplification of combined loops.
1758 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1759 OMP_TEAMS.
1760 (gimplify_omp_target_update): New function.
1761 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1762 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1763 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1764 (gimplify_body): If fndecl has "omp declare target" attribute, add
1765 implicit ORT_TARGET context around it.
1766 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1767 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1768 * tree-nested.c (convert_nonlocal_reference_stmt,
1769 convert_local_reference_stmt, convert_gimple_call): Handle
1770 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1771 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1772 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1773 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1774 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1775 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1776 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1777 BUILT_IN_GOMP_LOOP_END_CANCEL,
1778 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1779 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1780 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1781 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1782 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1783 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1784 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1785 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1786 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1787 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1788 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1789 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1790 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1791 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1792 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1793 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1794 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1795 gimple_build_omp_teams): New functions.
1796 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1797 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1798 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1799 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1800 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1801 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1802 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1803 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1804 GF_OMP_ATOMIC_SEQ_CST): New.
1805 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1806 gimple_build_omp_teams): New prototypes.
1807 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1808 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1809 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1810 GIMPLE_OMP_SINGLE as end of range.
1811 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1812 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1813 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1814 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1815 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1816 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1817 gimple_omp_target_kind, gimple_omp_target_set_kind,
1818 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1819 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1820 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1821 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1822 gimple_omp_teams_set_clauses): New inlines.
1823 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1824 and GIMPLE_OMP_TASKGROUP.
adcfd489 1825 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ		 1826 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1827 enum omp_clause_proc_bind_kind): New.
1828 (union omp_clause_subcode): Add depend_kind, map_kind and
1829 proc_bind_kind fields.
1830 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1831 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1832 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1833 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1834 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1835
67fa7880
TJ		 18362013-10-10 Teresa Johnson <tejohnson@google.com>
1837
1838 * predict.c (tree_estimate_probability): Add new parameter
1839 for estimate_bb_frequencies.
1840 (estimate_bb_frequencies): Add new parameter to force estimation.
1841 (rebuild_frequencies): When max frequency in function is small,
1842 recompute counts from frequencies.
1843 * predict.h (estimate_bb_frequencies): New parameter.
1844
b591a8b7
DM		 18452013-10-10 David Malcolm <dmalcolm@redhat.com>
1846
1847 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1848 optimizations are disabled.
1849
782f0db2
DM		 18502013-10-10 David Malcolm <dmalcolm@redhat.com>
1851
1852 * coverage.c (coverage_finish): Fix leak of da_file_name.
1853
13ef00fa
JH		 18542013-10-10 Jan Hubicka <jh@suse.cz>
1855
1856 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1857 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1858 Bobcat and generic.
1859
a1d50386
JJ		 18602013-10-10 Jakub Jelinek <jakub@redhat.com>
1861
1862 PR middle-end/58670
1863 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1864 if any labels are in FALLTHRU_BB, use a special label emitted
1865 immediately after the asm goto insn rather than label_rtx
1866 of the LABEL_DECL.
1867 (expand_asm_stmt): Adjust caller.
1868 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1869 edge if the last insn in predecessor is a jump with single successor,
1870 but it isn't simplejump_p.
1871
e2c2fde2
RB		 18722013-10-10 Richard Biener <rguenther@suse.de>
1873
1874 PR tree-optimization/58656
1875 * tree-ssa-pre.c (phi_translate): Do not cache failed
1876 translations.
1877
de6a5608
AM		 18782013-10-10 Andrew MacLeod <amacleod@redhat.com>
1879
adcfd489
UB		 1880 * gimplify.c: Include expr.h and tm_p.h for targets with special
1881 va-arg padding requirements.
de6a5608 1882
b184c8f1
AM		 18832013-10-10 Andrew MacLeod <amacleod@redhat.com>
1884
1885 * tree-flow.h: Move some prototypes to gimple.h.
1886 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1887 * gimple.h: Relocate some prototypes from tree-flow.h
1888 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1889 Move to gimplify.c.
1890 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1891 (build_va_arg_indirect_ref): Relocate and make static.
1892 (std_gimplify_va_arg_expr): Relocate here.
1893 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1894 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1895
13b72c22
AK		 18962013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1897
1898 * doc/md.texi: Document the mnemonic attribute.
1899
5db40447
AK		 19002013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1901
1902 PR target/57377
1903 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1904 (set_attr_alternative x ...) when searching for user defined
1905 mnemonic attribute.
1906
ed018d4b
AM		 19072013-10-10 Andrew MacLeod <amacleod@redhat.com>
1908
1909 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1910
4d5b5e9f
ER		 19112013-10-09 Easwaran Raman <eraman@google.com>
1912
1913 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1914 * cfgexpand.c (defer_stack_allocation): ...use here
1915 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1916
c9ef86a1
ZC		 19172013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1918
1919 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1920 (operand_equal_for_value_replacement): New function, extracted from
1921 value_replacement and enhanced to catch more cases.
1922 (value_replacement): Use operand_equal_for_value_replacement.
1923
2431114f
AM		 19242013-10-09 Andrew MacLeod <amacleod@redhat.com>
1925
adcfd489 1926 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM		 1927 get_max_loop_iterations.
1928
1dd2a9a0
KT		 19292013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1930
1931 * config/arm/aarch-common.c (arm_early_load_addr_dep):
1932 Place comment above function.
1933
c1bf2a39
AM		 19342013-10-09 Andrew MacLeod <amacleod@redhat.com>
1935
1936 * tree-flow.h: Remove all remaining prototypes, enums and structs that
1937 are not related to tree-cfg.c.
1938 * tree-ssa-address.h: New file. Relocate prototypes.
1939 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
1940 (addr_for_mem_ref): New. Combine call to get_address_description and
1941 return addr_for_mem_ref.
1942 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
1943 * tree-ssa-live.h: Adjust prototypes.
1944 * passes.c: Include tree-ssa-live.h.
1945 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
1946 * graphite.c (graphite_transform_loops): Make static.
1947 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 1948 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM		 1949 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
1950 * ipa-pure-const.c (warn_function_noreturn): Make static.
1951 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1952 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1953 Relocate from tree-cfg.c
1954 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
1955 static.
1956 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1957 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1958 Move to ipa-pure-const.c.
adcfd489
UB		 1959 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
1960 Relocate from tree-optimize.c.
c1bf2a39
AM		 1961 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
1962 make_pass_fixup_cfg): Move to tree-cfg.c.
1963 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
1964 Relocate some prototypes.
1965 * tree-data-ref.h (tree_check_data_deps) Add prototype.
1966 * tree-dump.c (dump_function_to_file): Remove prototype.
1967 Add tree-flow.h to the include file.
1968 * tree-dump.h: Remove prototype.
1969 * tree-parloops.h: New File. Add prototypes.
1970 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
1971 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
1972 from tree-ssa-loop.c.
1973 * tree-predcom.c (run_tree_predictive_commoning,
1974 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
1975 Relocate here from tree-ssa-loop.c.
adcfd489 1976 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM		 1977 ssa_name_values.release ().
1978 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
1979 (ssa_name_values): Relocate from tree-flow.h.
1980 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
1981 * tree-ssa-loop.c (run_tree_predictive_commoning,
1982 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
1983 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1984 make_pass_graphite, pass_data_graphite_transforms,
1985 make_pass_graphite_transforms, gate_tree_parallelize_loops,
1986 tree_parallelize_loops, pass_data_parallelize_loops,
1987 make_pass_parallelize_loops): Move to other files.
1988 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
1989 moved here.
1990 * tree.h: Remove prototypes from tree-address.c.
1991
cc524fc7
AM		 19922013-10-09 Andrew MacLeod <amacleod@redhat.com>
1993
1994 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
1995 (struct int_tree_map): Move to tree-hasher.h
1996 (SCALE, LABEL, PERCENT): Move to gimple.h
1997 * tree-flow-inline.h: Delete. Move functions to other files.
1998 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
1999 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2000 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2001 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2002 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2003 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2004 * tree-hasher.h: Don't include tree-flow.h.
2005 (struct int_tree_map): Relocate from tree-flow.h.
2006 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2007 tree-flow-inline.h and make static.
adcfd489
UB		 2008 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2009 tree-flow-inline.h.
cc524fc7
AM		 2010 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2011 tree-flow-inline.h and make static.
2012 * tree.h (is_global_var, may_be_aliased): Relocate from
2013 tree-flow-inline.h.
2014 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2015 * value-prof.c: No longer include tree-flow-inline.h.
2016 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2017
71343877
AM		 20182013-10-09 Andrew MacLeod <amacleod@redhat.com>
2019
2020 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2021 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2022 (enum move_pos): Move to tree-ssa-loop-im.h
2023 * cfgloop.h: Move some prototypes.
2024 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2025 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2026 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2027 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2028 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2029 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2030 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2031 (enum move_pos): Relocate here.
2032 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2033 tree-ssa-loop.c.
2034 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB		 2035 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2036 make_pass_lim): Relocate here from tree-ssa-loop.c.
2037 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM		 2038 tree-ssa-loop.c.
2039 (loop_edge_to_cancel, unloop_loops): Make static.
2040 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2041 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2042 (tree_complete_unroll, gate_tree_complete_unroll,
2043 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2044 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2045 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2046 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2047 (stmt_invariant_in_loop_p): Remove unused function.
2048 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2049 * tree-ssa-loop-manip.h: New file. Add prototypes.
2050 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2051 (gcov_type_to_double_int): Move to cfgloop.h.
2052 (double_int_cmp, bound_index,
2053 estimate_numbers_of_iterations_loop): Make static.
2054 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2055 (max_loop_iterations): Factor out get_max_loop_iterations.
2056 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2057 cfgloop.c.
2058 * tree-ssa-loop-niter.h: New file. Add prototypes.
2059 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2060 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2061 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2062 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2063 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2064 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2065 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2066 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2067 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2068 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2069 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2070 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2071 pass_data_complete_unroll, make_pass_complete_unroll,
2072 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2073 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2074 tree-ssa-loop-ivcanon.c.
2075 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2076 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2077 tree-ssa-loop-prefetch.c.
2078 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2079 tree-ssa-loop-im.c.
2080 (get_lsm_tmp_name): Relocate and add suffix parameter.
2081 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2082 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2083 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2084 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 2085 (get_estimated_loop_iterations): Factor out accessor from
71343877 2086 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB		 2087 (get_max_loop_iterations): Factor out accessor from
2088 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM		 2089 * loop-unroll.c (decide_unroll_constant_iterations,
2090 decide_unroll_runtime_iterations, decide_peel_simple,
2091 decide_unroll_stupid): Use new get_* accessors.
2092
826cacfe
MG		 20932013-10-09 Marc Glisse <marc.glisse@inria.fr>
2094
2095 PR tree-optimization/20318
2096 * doc/extend.texi (returns_nonnull): New function attribute.
2097 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2098 attribute.
2099 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2100 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2101
378f8976
EB		 21022013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2103
2104 PR middle-end/58570
2105 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2106 false if both components are bitfields.
2107
0fe04f5c
AV		 21082013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2109
2110 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2111 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2112 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2113 * config/aarch64/aarch64.h
2114 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2115 * config/aarch64/aarch64-simd-builtins.def
2116 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2117
bed9bae4
AV		 21182013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2119
2120 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2121
7f3d8b19
AV		 21222013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2123
2124 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2125
7df625a6
AV		 21262013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2127
2128 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2129 (vneg_f64): New intrinsic.
2130 (vneg_s8): Asm replaced with C.
2131 (vneg_s16): Likewise.
2132 (vneg_s32): Likewise.
2133 (vneg_s64): New intrinsic.
2134 (vnegq_f32): Asm replaced with C.
2135 (vnegq_f64): Likewise.
2136 (vnegq_s8): Likewise.
2137 (vnegq_s16): Likewise.
2138 (vnegq_s32): Likewise.
2139 (vnegq_s64): Likewise.
2140
2b86fca7
RL		 21412013-10-09 Renlin Li <Renlin.Li@arm.com>
2142
2143 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2144
e179df83
AK		 21452013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2146
2147 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2148 packed stack special handling.
2149 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2150 back to fixed stack slots for FPRs saved due to stdarg.
2151
74129172
AK		 21522013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2153
2154 * config/s390/s390.c (s390_frame_info): Restructure function.
2155
6455a49e
AK		 21562013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2157
2158 * config/s390/s390.c (struct s390_frame_layout): New field
2159 gpr_save_slots.
2160 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2161 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2162 regs_ever_clobbered to char*.
adcfd489
UB		 2163 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2164 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK		 2165 (s390_register_info): Enable FPR save slots. Move/Copy some
2166 functionality into ...
2167 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2168 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2169 function.
2170 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2171 not imply a stack frame.
2172 (s390_init_frame_layout): Remove variable clobbered_regs.
2173 (s390_update_register_info): Remove function.
2174 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2175 cannot be used for register renaming.
2176 (s390_hard_regno_scratch_ok): New function.
2177 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2178 (s390_initial_elimination_offset): Change offset calculation of
2179 the return address pointer.
adcfd489 2180 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 2181 (restore_gprs): Set frame related flag.
adcfd489 2182 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK		 2183 (s390_emit_prologue): Call s390_register_info instead of
2184 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2185 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2186 (s390_optimize_prologue): Call s390_optimize_register_info.
2187 Try to remove also FPR slot save/restore INSNs. Remove frame
2188 related flags from restore INSNs.
2189
59d96342
DD		 21902013-10-08 DJ Delorie <dj@redhat.com>
2191
0dc7adad
DD		 2192 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2193 (movhi): Likewise.
2194
59d96342
DD		 2195 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2196 avoid conflict with the MI use of %c.
2197 * config/rl78/rl78-real.md: change %c to %C throughout.
2198 * config/rl78/rl78-virt.md: Likewise.
2199
a81169d8
JH		 22002013-10-08 Jan Hubicka <jh@suse.cz>
2201
2202 * config/i386/i386.c (ix86_option_override_internal): Switch
2203 to SSE math for -ffast-math when target ISA supports SSE2.
2204
481d1b81
AM		 22052013-10-08 Andrew MacLeod <amacleod@redhat.com>
2206
2207 * tree-flow.h: Remove some prototypes.
2208 * tree.h: Remove some protypes, add a couple.
2209 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2210 using_eh_for_cleanups_p): Add interface routines for front ends.
2211 * tree-eh.h: New file. Add protoptyes.
2212 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2213 (add_stmt_to_eh_lp_fn): Make static.
2214 (lower_try_finally): Use new using_eh_for_cleanups_p.
2215 * emit-rtl.c: Include tree-eh.h.
2216 * gimple.h: Include tree-eh.h.
2217
0e6a0e48
MG		 22182013-10-08 Marc Glisse <marc.glisse@inria.fr>
2219
2220 PR tree-optimization/58480
2221 * tree-vrp.c (infer_nonnull_range): New function.
2222 (infer_value_range): Call infer_nonnull_range.
2223
ef6179d1
DC		 22242013-10-08 Dehao Chen <dehao@google.com>
2225
2226 PR tree-optimization/58619
2227 * tree-inline.c (copy_phis_for_bb): Combine location data
2228 only if non-null.
2229
07d964d5
ZC		 22302013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2231
2232 PR target/58423
2233 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2234 RTX_FRAME_RELATED_P on INSN.
2235
19e9b2a3
BS		 22362013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2237
2238 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2239 (altivec_expand_vec_perm_const): Call it.
2240
0cf68694
BS		 22412013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2242
2243 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2244 sequences for LE VSX loads and stores at expand time.
2245 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2246 prototype.
2247 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2248 (rs6000_gen_le_vsx_permute): New.
2249 (rs6000_gen_le_vsx_load): New.
2250 (rs6000_gen_le_vsx_store): New.
2251 (rs6000_gen_le_vsx_move): New.
2252 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2253 (*vsx_le_perm_load_v4si): New.
2254 (*vsx_le_perm_load_v8hi): New.
2255 (*vsx_le_perm_load_v16qi): New.
2256 (*vsx_le_perm_store_v2di): New.
2257 (*vsx_le_perm_store_v4si): New.
2258 (*vsx_le_perm_store_v8hi): New.
2259 (*vsx_le_perm_store_v16qi): New.
2260 (*vsx_xxpermdi2_le_<mode>): New.
2261 (*vsx_xxpermdi4_le_<mode>): New.
2262 (*vsx_xxpermdi8_le_V8HI): New.
2263 (*vsx_xxpermdi16_le_V16QI): New.
2264 (*vsx_lxvd2x2_le_<mode>): New.
2265 (*vsx_lxvd2x4_le_<mode>): New.
2266 (*vsx_lxvd2x8_le_V8HI): New.
2267 (*vsx_lxvd2x16_le_V16QI): New.
2268 (*vsx_stxvd2x2_le_<mode>): New.
2269 (*vsx_stxvd2x4_le_<mode>): New.
2270 (*vsx_stxvd2x8_le_V8HI): New.
2271 (*vsx_stxvd2x16_le_V16QI): New.
2272
9520e1eb
RL		 22732013-10-07 Renlin Li <Renlin.Li@arm.com>
2274
2275 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2276
36c0bd4f
AK		 22772013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2278
2279 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2280 loop to work also for 31bit ABI.
2281 Save the stack pointer for frame_size > 0.
2282
ee163e72
AK		 22832013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2284
2285 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2286 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2287 constraint letters from expanders.
2288 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2289 retry count to general_operand.
2290 ("tabort"): Give operand 0 a mode.
2291 ("tabort_1"): Add mode and constraint letter for operand 0.
2292 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2293
45d99234
JL		 22942013-10-04 Jeff Law <law@redhat.com>
2295
2296 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2297
adcfd489
UB		 2298 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2299 out of ...
45d99234
JL		 2300 (thread_across_edge): Here. Call it.
2301
7c327f7b
CC		 23022013-10-04 Cary Coutant <ccoutant@google.com>
2303
2304 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2305 discarding a location list expression (or a piece of one).
7c327f7b 2306
6867e128
JH		 23072013-10-03 Jan Hubicka <jh@suse.cz>
2308
65efa7e7
UB		 2309 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2310 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2311 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2312
c7f36d55
JH		 23132013-10-03 Jan Hubicka <jh@suse.cz>
2314
65efa7e7 2315 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH		 2316 accumulate-outgoing-args when producing unwind info.
2317
4f0bee4c
WM		 23182013-10-03 Wei Mi <wmi@google.com>
2319
2320 * lra-constraints.c (insert_move_for_subreg): New function
2321 extracted from simplify_operand_subreg.
2322 (simplify_operand_subreg): Add reload for paradoxical subreg.
2323
111c3f39
RX		 23242013-10-03 Rong Xu <xur@google.com>
2325
65efa7e7
UB		 2326 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2327 the candidate of builtin_expect such that we should fix the
2328 size/time estimation.
2329 (estimate_function_body_sizes): Do the acutally size/time fix-up
2330 for builtin_expect.
111c3f39 2331
942df739
RX		 23322013-10-03 Rong Xu <xur@google.com>
2333
65efa7e7
UB		 2334 * predict.c (tree_predict_by_opcode): Get the probability
2335 for builtin_expect from param builtin_expect_probability.
2336 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2337 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2338 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2339
2284b034
MG		 23402013-10-03 Marc Glisse <marc.glisse@inria.fr>
2341
2342 PR c++/19476
2343 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2344 * calls.c (alloca_call_p): Use get_callee_fndecl.
2345 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2346 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2347 Likewise.
2348 (vrp_visit_stmt): Remove duplicated code.
2349
0609bdf2
MM		 23502013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2351
2352 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2353 ceildf2, btruncdf2, instead of vsx_* name.
2354
2355 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2356 iterators to only do V2DF and V4SF here. Move the DF code to
2357 rs6000.md where it is combined with SF mode. Replace <VSv> with
2358 just 'v' since only vector operations are handled with these insns
2359 after moving the DF support to rs6000.md.
2360 (vsx_sub<mode>3): Likewise.
2361 (vsx_mul<mode>3): Likewise.
2362 (vsx_div<mode>3): Likewise.
2363 (vsx_fre<mode>2): Likewise.
2364 (vsx_neg<mode>2): Likewise.
2365 (vsx_abs<mode>2): Likewise.
2366 (vsx_nabs<mode>2): Likewise.
2367 (vsx_smax<mode>3): Likewise.
2368 (vsx_smin<mode>3): Likewise.
2369 (vsx_sqrt<mode>2): Likewise.
2370 (vsx_rsqrte<mode>2): Likewise.
2371 (vsx_fms<mode>4): Likewise.
2372 (vsx_nfma<mode>4): Likewise.
2373 (vsx_copysign<mode>3): Likewise.
2374 (vsx_btrunc<mode>2): Likewise.
2375 (vsx_floor<mode>2): Likewise.
2376 (vsx_ceil<mode>2): Likewise.
2377 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2378 (vsx_sminsf3): Likewise.
2379 (vsx_fmadf4): Likewise.
2380 (vsx_fmsdf4): Likewise.
2381 (vsx_nfmadf4): Likewise.
2382 (vsx_nfmsdf4): Likewise.
2383 (vsx_cmpdf_internal1): Likewise.
2384
2385 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2386 simpler to select whether a target has SPE or traditional floating
2387 point support in iterators.
2388 (TARGET_DF_SPE): Likewise.
2389 (TARGET_SF_FPR): Likewise.
2390 (TARGET_DF_FPR): Likewise.
2391 (TARGET_SF_INSN): Macros to say whether floating point support
2392 exists for a given operation for expanders.
2393 (TARGET_DF_INSN): Likewise.
2394
2395 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2396 combining of SF/DF mode operations, using both traditional and VSX
2397 registers.
2398 (Fvsx): Likewise.
2399 (Ff): Likewise.
2400 (Fv): Likewise.
2401 (Fs): Likewise.
2402 (Ffre): Likewise.
2403 (FFRE): Likewise.
2404 (abs<mode>2): Combine SF/DF modes using traditional floating point
2405 instructions. Add support for using the upper DF registers with
2406 VSX support, and SF registers with power8-vector support. Update
2407 expanders for operations supported by both the SPE and traditional
2408 floating point units.
2409 (abs<mode>2_fpr): Likewise.
2410 (nabs<mode>2): Likewise.
2411 (nabs<mode>2_fpr): Likewise.
2412 (neg<mode>2): Likewise.
2413 (neg<mode>2_fpr): Likewise.
2414 (add<mode>3): Likewise.
2415 (add<mode>3_fpr): Likewise.
2416 (sub<mode>3): Likewise.
2417 (sub<mode>3_fpr): Likewise.
2418 (mul<mode>3): Likewise.
2419 (mul<mode>3_fpr): Likewise.
2420 (div<mode>3): Likewise.
2421 (div<mode>3_fpr): Likewise.
2422 (sqrt<mode>3): Likewise.
2423 (sqrt<mode>3_fpr): Likewise.
2424 (fre<Fs>): Likewise.
2425 (rsqrt<mode>2): Likewise.
2426 (cmp<mode>_fpr): Likewise.
2427 (smax<mode>3): Likewise.
2428 (smin<mode>3): Likewise.
2429 (smax<mode>3_vsx): Likewise.
2430 (smin<mode>3_vsx): Likewise.
2431 (negsf2): Delete SF operations that are merged with DF.
2432 (abssf2): Likewise.
2433 (addsf3): Likewise.
2434 (subsf3): Likewise.
2435 (mulsf3): Likewise.
2436 (divsf3): Likewise.
2437 (fres): Likewise.
2438 (fmasf4_fpr): Likewise.
2439 (fmssf4_fpr): Likewise.
2440 (nfmasf4_fpr): Likewise.
2441 (nfmssf4_fpr): Likewise.
2442 (sqrtsf2): Likewise.
2443 (rsqrtsf_internal1): Likewise.
2444 (smaxsf3): Likewise.
2445 (sminsf3): Likewise.
2446 (cmpsf_internal1): Likewise.
2447 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2448 (negdf2): Delete DF operations that are merged with SF.
2449 (absdf2): Likewise.
2450 (nabsdf2): Likewise.
2451 (adddf3): Likewise.
2452 (subdf3): Likewise.
2453 (muldf3): Likewise.
2454 (divdf3): Likewise.
2455 (fred): Likewise.
2456 (rsqrtdf_internal1): Likewise.
2457 (fmadf4_fpr): Likewise.
2458 (fmsdf4_fpr): Likewise.
2459 (nfmadf4_fpr): Likewise.
2460 (nfmsdf4_fpr): Likewise.
2461 (sqrtdf2): Likewise.
2462 (smaxdf3): Likewise.
2463 (smindf3): Likewise.
2464 (cmpdf_internal1): Likewise.
2465 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2466 (btrunc<mode>2): Delete separate expander, and combine with the
2467 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2468 (btrunc<mode>2_fpr): Likewise.
2469 (ceil<mode>2): Likewise.
2470 (ceil<mode>2_fpr): Likewise.
2471 (floor<mode>2): Likewise.
2472 (floor<mode>2_fpr): Likewise.
2473 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2474 Add support for using the upper registers with VSX and
2475 power8-vector. Move insns to be closer to the define_expands. On
2476 VSX systems, prefer the traditional form of FMA over the VSX
2477 version, since the traditional form allows the target not to
2478 overlap with the inputs.
2479 (fms<mode>4_fpr): Likewise.
2480 (nfma<mode>4_fpr): Likewise.
2481 (nfms<mode>4_fpr): Likewise.
2482
5bea0c6c 24832013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 2484 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT		 2485
2486 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2487 (struct mult_cost_table): Likewise.
2488 (struct mem_cost_table): Likewise.
2489 (struct fp_cost_table): Likewise.
2490 (struct vector_cost_table): Likewise.
2491 (cpu_cost_table): Likewise.
2492 * config/arm/arm.opt (mold-rts-costs): New option.
2493 (mnew-generic-costs): Likewise.
2494 * config/arm/arm.c (generic_extra_costs): New table.
2495 (cortexa15_extra_costs): Likewise.
2496 (arm_slowmul_tune): Use NULL as new costs.
2497 (arm_fastmul_tune): Likewise.
2498 (arm_strongarm_tune): Likewise.
2499 (arm_xscale_tune): Likewise.
2500 (arm_9e_tune): Likewise.
2501 (arm_v6t2_tune): Likewise.
2502 (arm_cortex_a5_tune): Likewise.
2503 (arm_cortex_a9_tune): Likewise.
2504 (arm_v6m_tune): Likewise.
2505 (arm_fa726te_tune): Likewise.
2506 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2507 (arm_cortex_tune): Use generict_extra_costs.
2508 (shifter_op_p): New function.
2509 (arm_unspec_cost): Likewise.
2510 (LIBCALL_COST): Define.
2511 (arm_new_rtx_costs): New function.
2512 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2513 table is available. Use old costs otherwise unless mnew-generic-costs
2514 is specified.
2515 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2516 (cpu_cost_table): Declare.
2517
24c56925
MS		 25182013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2519
2520 PR target/58460
2521 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2522 (*subs_mul_imm_<mode>)
2523 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2524 (*sub_<shift>_<mode>)
2525 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2526 Remove k constraint.
2527
2ab8f063
IB		 25282013-10-03 Ian Bolton <ian.bolton@arm.com>
2529
2530 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2531 code.
2532 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2533
79221839
TJ		 25342013-10-02 Teresa Johnson <tejohnson@google.com>
2535
2536 * predict.c (probably_never_executed): New function.
2537 (probably_never_executed_bb_p): Invoke probably_never_executed.
2538 (probably_never_executed_edge_p): Ditto.
2539 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2540 Treat profile insanities conservatively.
2541
d441afe0
JDA		 25422013-10-02 John David Anglin <danglin@gcc.gnu.org>
2543
2544 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2545
80f466c4
VM		 25462013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2547
2548 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2549 use it. Use smaller increase for scratch. Don't increase reject
2550 for early clobber scratch.
2551 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2552 setting eliminated regs except setting fp from hfp.
2553 (lra_eliminate): Check lra_insn_recog_data on NULL.
2554
6e228b4b
MM		 25552013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2556
2557 PR target/58587
2558 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 2559 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM		 2560 (RS6000_CPU, power7 defaults): Likewise.
2561
11b54a5a
UB		 25622013-10-02 Uros Bizjak <ubizjak@gmail.com>
2563
2564 * config/x-linux (host-linux.o): Remove header dependencies.
2565 Use $(COMPILE) and $(POSTCOMPILE).
2566 * config/t-linux-android (linux-android.o): Ditto.
2567
bbc02b69
UB		 25682013-10-02 Uros Bizjak <ubizjak@gmail.com>
2569
2570 * Makefile.in (expmed.o-warn): Remove.
2571
440917de
AM		 25722013-10-02 Andrew MacLeod <amacleod@redhat.com>
2573
2574 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2575 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2576
99206ca9
TJ		 25772013-10-02 Teresa Johnson <tejohnson@google.com>
2578
2579 * dojump.c (do_jump_1): Divide probability between
2580 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2581
56e82b14
TT		 25822013-10-02 Tom Tromey <tromey@redhat.com>
2583
2584 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2585
3d9c733e
AM		 25862013-10-02 Andrew MacLeod <amacleod@redhat.com>
2587
2588 * tree-flow.h: Remove some prototypes.
2589 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2590 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2591 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2592 mark_virtual_phi_result_for_renaming): Relocate here.
2593 * tree-into-ssa.h: Add prototypes.
bbc02b69 2594 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM		 2595 single_pred_before_succ_order.
2596 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2597 (nonfreeing_call_p) Move to gimple.c.
2598 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2599 tree-ssa-phiopt.c.
2600 * basic-block.h (single_pred_before_succ_order): Add prototype.
2601 * gimple.c (nonfreeing_call_p): Relocate here.
2602 * gimple.h: Add prototype.
2603 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2604 * tree-ssa-dom.h: New file. Relocate prototypes here.
2605 * tree-ssa.h: Include tree-ssa-dom.h.
2606
78cedfb1
UB		 26072013-10-02 Uros Bizjak <ubizjak@gmail.com>
2608
2609 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2610 Use $(COMPILE) and $(POSTCOMPILE).
2611
2612 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2613
744730a4
AM		 26142013-10-02 Andrew MacLeod <amacleod@redhat.com>
2615
2616 * tree-flow.h: Remove some prototypes.
2617 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2618 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2619 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2620 propagate_tree_value*): Move from here to...
2621 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2622 propagate_tree_value*): Relocate here.
2623 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2624 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2625 * gimple-fold.c: Remove gimple-fold.h from include list.
2626 * tree-vrp.c: Remove gimple-fold.h from include list.
2627 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2628 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2629 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2630 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2631 * sese.c: Add tree-ssa-propagate.h to include list.
2632
826a536d
RB		 26332013-10-02 Richard Biener <rguenther@suse.de>
2634
2635 * tree-loop-distribution.c: Include tree-vectorizer.h for
2636 find_loop_location.
2637 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 2638 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB		 2639 (partition_alloc): Adjust.
2640 (partition_builtin_p): Likewise.
2641 (partition_has_writes): Remove.
2642 (partition_reduction_p): New function.
2643 (partition_merge_into): Likewise.
2644 (generate_code_for_partition): Commonize builtin partition
2645 handling tail.
2646 (rdg_cannot_recompute_vertex_p): Remove.
2647 (already_processed_vertex_p): Likewise.
2648 (rdg_flag_vertex): Do not set has_writes.
2649 (classify_partition): Adjust.
2650 (rdg_build_partitions): Do not set has_writes, treat all
2651 partitions as useful.
78cedfb1 2652 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB		 2653 (tree_loop_distribution): Report number of loops and library
2654 calls generated as opt-info.
2655
4b403ece
AM		 26562013-10-02 Andrew MacLeod <amacleod@redhat.com>
2657
2658 * tree-flow.h: Include new .h files. Move prototypes.
2659 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2660 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2661 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2662 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2663 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2664 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2665 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2666 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 2667 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM		 2668 tree-dfa.h.
2669 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2670 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2671 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2672 * tree-scalar-evolution.c: Include tree.h.
2673 * sese.c: Include tree.h.
2674 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2675 * dwarf2out.c: Include tree-dfa.h.
2676 * tree-chrec.c: Include tree.h.
2677 * tree-data-ref.c: Include tree.h.
2678
1d2151c6
YZ		 26792013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2680
2681 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2682 Fix whitespace.
2683
441ad147
RO		 26842013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2685
2686 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2687 Use $(COMPILE) and $(POSTCOMPILE).
2688 (sol2-cxx.o): Likewise.
2689 (sol2-stubs.o): Likewise.
2690 (sol2.o): Likewise.
2691 * config/x-solaris (host-solaris.o): Likewise.
2692
2693 * config/sparc/t-sparc (sparc.o): Remove.
2694 (sparc-c.o): Remove header dependencies.
2695 Use $(COMPILE) and $(POSTCOMPILE).
2696 * config/sparc/x-sparc: Likewise.
2697
2deaf8b0
JR		 26982013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2699
2700 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2701 * config/arc/arc700.md: Likewise.
2702 * config/arc/arc-modes.def: Likewise.
2703 * config/arc/arc-simd.h: Likewise.
2704 * config/arc/t-arc-uClibc: Likewise.
2705 * config/arc/t-arc-newlib: Likewise.
2706
f6fe771a
RL		 27072013-10-02 Renlin Li <renlin.li@arm.com>
2708
78cedfb1
UB		 2709 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2710 plus_constant.
f6fe771a
RL		 2711 (aarch64_expand_epilogue): Likewise.
2712
0916f876
YZ		 27132013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2714 Yufeng Zhang <yufeng.zhang@arm.com>
2715
2716 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2717 declaration.
2718 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2719 'base_in' represent a conversion and legal_cast_p_1 holds; set
2720 'base_in' with the returned value from get_unwidened.
2721
ebfcd719
KT		 27222013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2723
2724 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2725 plus_constant is not used.
2726
157ca3e9
WM		 27272013-10-01 Wei Mi <wmi@google.com>
2728
78cedfb1 2729 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM		 2730 * config/i386/i386.md: Add define_peephole2 to
2731 break partial reg stall for cvtss2sd/cvtsd2ss.
2732
cd4dd8f0
JR		 27332013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2734
2735 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 2736 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 2737
1465e5cf
JL		 27382013-10-01 Jeff Law <law@redhat.com>
2739
2740 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2741 outgoing_edge and intermediate_edge fields. Instead store the path.
2742 (redirection_data::hash): Hash on the last edge's destination index.
2743 (redirection_data::equal): Check the entire thread path.
2744 (lookup_redirectio_data): Corresponding changes.
2745 (create_edge_and_update_destination_phis): Likewise.
2746 (thread_single_edge): Likewise.
2747
25c606cb 27482013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 2749 Diego Novillo <dnovillo@google.com>
25c606cb
JR		 2750
2751 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2752 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2753 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2754 versions of these patterns.
2755
0ccbc132
JR		 2756 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2757 (__builtin_arc_aligned): Likewise.
2758
f55d4a20
JR		 2759 * config/arc/arc.md: Expand adc_0 comment stating the intended
2760 purpose and why it isn't ready.
2761 Replace commented out call_value_via_label_mixed with a
2762 plain comment about bl_s.
2763
5719867d 2764 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR		 2765 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2766 Remove if (1) condition.
2767 (arc_encode_section_info): Fix comment.
2768
6462fab0
JR		 27692013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2770
2771 * config/arc/arc.c (arc_conditional_register_usage):
2772 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2773 Also set reg_alloc_order for DMA config regs.
2774
5d5f6720
JR		 27752013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2776 Jeremy Bennett <jeremy.bennett@embecosm.com>
2777
2778 * doc/install.texi (--with-cpu): Mention ARC.
2779 (arc-*-elf32): New paragraph.
2780 (arc-linux-uclibc): Likewise.
2781 * doc/md.texi (Machine Constraints): Add ARC part.
2782 * doc/invoke.texi: (menu): Add ARC Options.
2783 (Machine Dependent Options) <ARC Options>: Add synopsis.
2784 (node ARC Options): Add.
2785 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2786 (ARC Built-in Functions): New section defining
2787 generic ARC built-in functions.
2788 (ARC SIMD Built-in Functions): New section defining SIMD specific
2789 built-in functions.
2790 (Declaring Attributes of Functions): Extended
2791 description of short_call and long_call attributes for ARC and
2792 added index entries.
2793
526b7aee
SV		 27942013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2795 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2796 Joern Rennecke <joern.rennecke@embecosm.com>
2797 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2798 Brendan Kehoe <brendan@zen.org>
2799 Michael Eager <eager@eagercon.com>
2800 Simon Cook <simon.cook@embecosm.com>
2801 Jeremy Bennett <jeremy.bennett@embecosm.com>
2802
2803 * config/arc, common/config/arc: New directories.
2804
53426f6c
JR		 28052013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2806 Brendan Kehoe <brendan@zen.org>
2807 Simon Cook <simon.cook@embecosm.com>
2808
2809 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2810
1df9f5a9
AM		 28112013-10-01 Andrew MacLeod <amacleod@redhat.com>
2812
2813 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2814 * tree-ssa-coalesce.h: New. Move prototype to here.
2815 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2816 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2817 (gimple_can_coalesce_p): Move to...
2818 * gimple.c (gimple_can_coalesce_p): Here.
2819
80560f95
AM		 28202013-10-01 Andrew MacLeod <amacleod@redhat.com>
2821
2822 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2823 (dump_decl_set): Move to gimple.c.
2824 * gimple.h: Don't include tree-ssa-operands.h.
2825 (dump_decl_set): Add prototype.
2826 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2827 Move to gimple-ssa.h.
2828 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2829 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2830 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2831 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2832 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2833 than PHI_ARG_DEF.
2834 (dump_decl_set): Relocate here.
2835 * gimple-ssa.h: New file.
2836 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2837 Relocate from gimple.h.
2838 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2839 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2840 swap_tree_operands and remove non-ssa path.
2841 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2842 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2843 swap_ssa_operands.
2844 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2845 vect_is_simple_reduction_1): Use swap_ssa_operands.
2846 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2847 (enum need_phi_state): Move to tree-into-ssa.c.
2848 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2849 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2850 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2851 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2852 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2853 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2854 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2855 num_imm_uses): Move to ssa-iterators.h.
2856 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2857 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2858 tree-phinodes.h.
2859 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2860 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2861 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2862 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2863 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2864 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2865 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2866 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2867 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2868 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2869 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2870 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2871 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2872 (set_phi_nodes): Move to tree-phinodes.h.
2873 * tree-ssa-operands.h (enum ssa_op_iter_type,
2874 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2875 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2876 (dump_decl_set): Remove prototype.
2877 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2878 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2879 (set_phi_nodes): Relocate from tree-flow-inline.h.
2880 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2881 tree-flow-inline.h
2882 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2883 include list. Temporarily add gimple.h to include list.
2884 * ssa-iterators.h: New file.
2885 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2886 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2887 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2888 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2889 Relocate from tree-ssa-operands.h.
2890 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2891 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2892 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2893 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2894 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2895 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2896 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2897 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2898 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2899 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2900 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2901 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2902 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2903 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2904 Relocate from tree-flow-inline.h.
2905 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2906
8da00d65
VP		 29072013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2908
2909 * aarch64-simd.md
2910 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2911 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2912 Insert '\t' to output template.
2913 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2914 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2915 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2916 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2917
e2ebe1c2
UB		 29182013-10-01 Uros Bizjak <ubizjak@gmail.com>
2919
2920 * doc/install.texi (Host/target specific installation notes for GCC):
2921 Put @anchor before @heading.
2922 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2923 Use @email for email addresses.
2924
aee2d611
JL		 29252013-10-01 Jeff Law <law@redhat.com>
2926
2927 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
2928 a vec. Only delete the path if we create one without successfully
2929 registering a jump thread.
2930 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
2931 as a pointer.
2932 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
2933 (paths): New vector of jump threading paths.
2934 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
2935 (THREAD_PATH): New accessor macro for the entire thread path.
2936 (lookup_redirection_data): Get intermediate and final outgoing edge
2937 from the thread path.
2938 (create_edge_and_update_destination_phis): Copy the threading path.
2939 (ssa_fix_duplicate_block_edges): Get edges and block types from the
2940 jump threading path.
2941 (ssa_redirect_edges): Get edges and block types from the jump threading
2942 path. Free the path vector.
2943 (thread_block): Get edges from the jump threading path. Look at the
2944 entire path to see if we thread to a loop exit. If we cancel a jump
2945 thread request, then free the path vector.
2946 (thread_single_edge): Get edges and block types from the jump threading
2947 path. Free the path vector.
2948 (thread_through_loop_header): Get edges and block types from the jump
2949 threading path. Free the path vector.
2950 (mark_threaded_blocks): Iterate over the vector of paths and store
2951 the path on the appropriate edge. Get edges and block types from the
2952 jump threading path.
2953 (mark_threaded_blocks): Get edges and block types from the jump
2954 threading path. Free the path vector.
2955 (thread_through_all_blocks): Use the vector of paths rather than
2956 a vector of 3-edge sets.
2957 (register_jump_thread): Accept pointer to a path vector rather
2958 than the path vector itself. Store the path vector for later use.
2959 Simplify.
2960
966f97ac 29612013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 2962 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ		 2963
2964 PR target/58574
2965 * config/s390/s390.c (s390_split_branches): Modify check for table
2966 jump insns.
2967 (s390_chunkify_start): Rearrange table jump insn check in order to
2968 deal with compare and branch insns correctly.
2969
3a323a38
KV		 29702013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
2971
2972 PR target/58578
2973 Revert
2974 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
2975 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
2976 define_insn_and_split.
2977 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
2978 (shiftsi3_compare): New pattern.
2979 (rrx): New pattern.
2980 * config/arm/unspecs.md (UNSPEC_RRX): New.
2981
90444831
AM		 29822013-10-01 Alan Modra <amodra@gmail.com>
2983
2984 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
2985 casing inout operands.
2986
818625cf
RB		 29872013-10-01 Richard Biener <rguenther@suse.de>
2988
2989 PR tree-optimization/58553
2990 * tree-loop-distribution.c (struct partition_s): Add niter member.
2991 (classify_partition): Populate niter member for the partition
2992 and properly identify whether the relevant store happens before
2993 or after the loop exit.
2994 (generate_memset_builtin): Use niter member from the partition.
2995 (generate_memcpy_builtin): Likewise.
2996
30f641cd
RS		 29972013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
2998
2999 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3000 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3001
65d3284b
RS		 30022013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3003
3004 * basic-block.h (edge_list): Prefix member names with "m_".
3005 * context.h (context): Likewise.
3006 * domwalk.h (dom_walker): Likewise.
3007 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3008 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3009 * hash-table.h (hash_table): Likewise.
3010 * machmode.h (bit_field_mode_iterator): Likewise.
3011 * pass_manager.h (pass_list): Likewise.
3012 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3013 * tree-pass.h (pass_data): Likewise.
3014 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3015 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3016 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3017 * asan.c (pass_data_asan): Update accordingly.
3018 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3019 (control_dependences::control_dependences): Likewise.
3020 (control_dependences::~control_dependences): Likewise.
3021 (control_dependences::~control_dependences): Likewise.
3022 (control_dependences::get_edges_dependent_on): Likewise.
3023 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3024 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3025 * context.c (gcc::context::context): Likewise.
3026 * cprop.c (pass_rtl_cprop::clone): Likewise.
3027 * domwalk.c (dom_walker::walk): Likewise.
3028 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3029 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3030 * mode-switching.c (pass_mode_switching::clone): Likewise.
3031 * passes.c (opt_pass::opt_pass): Likewise.
3032 (pass_manager::pass_manager): Likewise.
3033 * predict.c (pass_strip_predict_hints::clone): Likewise.
3034 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3035 (pass_split_all_insns::clone): Likewise.
3036 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3037 Likewise.
3038 (bit_field_mode_iterator::next_mode): Likewise.
3039 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3040 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3041 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3042 * tree-complex.c (pass_lower_complex::clone): Likewise.
3043 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3044 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3045 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3046 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3047 (pass_fold_builtins::clone): Likewise.
3048 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3049 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3050 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3051 (pass_cd_dce::clone): Likewise.
3052 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3053 (pass_phi_only_cprop::clone): Likewise.
3054 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3055 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3056 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3057 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3058 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3059 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3060 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3061 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3062 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3063 * tree-vrp.c (pass_vrp::clone): Likewise.
3064 * tsan.c (pass_tsan::clone): Likewise.
3065
f66d0891
JJ		 30662013-09-30 Jakub Jelinek <jakub@redhat.com>
3067
d2a365a8
JJ		 3068 PR middle-end/58564
3069 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3070 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3071
f66d0891
JJ		 3072 PR middle-end/58564
3073 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3074 optimization, punt if sign_bit_p looked through any zero extension.
3075
05357ac3
TJ		 30762013-09-30 Teresa Johnson <tejohnson@google.com>
3077
3078 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3079 Update redirected out edge count in joiner case.
3080 (ssa_redirect_edges): Common the joiner and non-joiner cases
3081 so that joiner case gets profile updates.
3082
ca406576
RB		 30832013-09-30 Richard Biener <rguenther@suse.de>
3084
3085 PR tree-optimization/58554
e2ebe1c2
UB		 3086 * tree-loop-distribution.c (classify_partition): Require
3087 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB		 3088 (tree_loop_distribution): Calculate dominance info.
3089
92d649c4
VK		 30902013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3091
3092 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3093 (NO_PROFILE_COUNTERS): Likewise.
3094 (PROFILE_HOOK): Likewise.
3095 (FUNCTION_PROFILER): Likewise.
3096 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3097
bd9534e2
IS		 30982013-09-30 Iain Sandoe <iain@codesourcery.com>
3099
3100 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3101 calls and defines in TARGET_MACHO conditional.
3102 (load_macho_picbase_di): Likewise.
3103 (reload_macho_picbase): Likewise.
3104 (reload_macho_picbase_si): Likewise.
3105 (reload_macho_picbase_di): Likewise.
3106 (nonlocal_goto_receiver): Likewise.
3107
4494fbc9
NC		 31082013-09-30 Nick Clifton <nickc@redhat.com>
3109
3110 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3111 that use the MSP430X ISA.
3112 (msp430_option_override): Scan -mmcu command line option for any
3113 MCU name that supports the MSP430X ISA.
3114 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3115 -mmcu options which enable the MSP430X ISA.
3116
791d9044
RB		 31172013-09-30 Richard Biener <rguenther@suse.de>
3118
3119 PR middle-end/58532
3120 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3121 before looking for setjmp-like calls.
3122
e1c5c877
IS		 31232013-09-29 Iain Sandoe <iain@codesourcery.com>
3124
3125 PR target/10901
3126 * config/darwin-protos.h (machopic_get_function_picbase): New.
3127 * config/darwin.c (machopic_get_function_picbase): New.
3128 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3129 label for a new func. (load_macho_picbase_di): Likewise.
3130 (reload_macho_picbase): New expand.
3131 (reload_macho_picbase_si): New insn.
3132 (reload_macho_picbase_di): New insn.
3133 (nonlocal_goto_receiver): New define and split.
3134 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3135 (unspecv enum): Add UNSPECV_NLGR.
3136
749278c5
IS		 31372013-09-29 Iain Sandoe <iain@codesourcery.com>
3138
3139 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3140 that altivec registers are correctly sized on Darwin.
3141
2c43a51c
IS		 31422013-09-29 Iain Sandoe <iain@codesourcery.com>
3143
3144 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3145 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3146 * config/x-darwin (host-darwin.o): Likewise.
3147 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3148 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3149 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3150
7a5911d3
UB		 31512013-09-29 Uros Bizjak <ubizjak@gmail.com>
3152
3153 * doc/invoke.texi: Fix usage of @tie{} command.
3154
3f67a73d
EB		 31552013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3156
3157 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3158
9ac2f538
JH		 31592013-09-28 Jan Hubicka <jh@suse.cz>
3160
3161 * config/i386/x86-tune.def: Add documentation for each of the options;
3162 add whitespace.
3163
b0ff06da
JH		 31642013-09-28 Jan Hubicka <jh@suse.cz>
3165
7a5911d3
UB		 3166 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3167 generic.
b0ff06da
JH		 3168 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3169 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3170 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3171 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3172 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3173
c3284718
RS		 31742013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3175
3176 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3177 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3178 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3179 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3180 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3181 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3182 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3183 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3184 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3185 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3186 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3187 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3188 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3189 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3190 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3191 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3192 gimple.h, godump.c, graphite-clast-to-gimple.c,
3193 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3194 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3195 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3196 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3197 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3198 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3199 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3200 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3201 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3202 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3203 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3204 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3205 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3206 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3207 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3208 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3209 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3210 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3211 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3212 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3213 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3214 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3215 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3216 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3217 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3218 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3219 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3220 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3221 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3222 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3223 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3224 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3225 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3226 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3227 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3228 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3229 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3230 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3231 whitespace before "(".
3232
c6285bd7
SL		 32332013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3234
3235 * expr.h (extract_bit_field): Remove packedp parameter.
3236 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3237 from forward declaration.
3238 (store_split_bit_field): Remove packedp arg from calls to
3239 extract_fixed_bit_field.
3240 (extract_bit_field_1): Remove packedp parameter and packedp
3241 argument from recursive calls and calls to extract_fixed_bit_field.
3242 (extract_bit_field): Remove packedp parameter and corresponding
3243 arg to extract_bit_field_1.
3244 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3245 to issue warnings.
3246 (extract_split_bit_field): Remove packedp arg from call to
3247 extract_fixed_bit_field.
3248 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3249 (copy_blkmode_from_reg): Likewise.
3250 (copy_blkmode_to_reg): Likewise.
3251 (read_complex_part): Likewise.
3252 (store_field): Likewise.
3253 (expand_expr_real_1): Likewise.
3254 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3255 to extract_bit_field.
3256 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3257 call to extract_bit_field.
3258 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3259 call to extract_bit_field.
3260 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3261 and special packedp behavior from -fstrict-volatile-bitfields
3262 documentation.
3263
522d4efc
JBG		 32642013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3265
3266 * lra-eliminations.c (init_elim_table): Guard value_p.
3267
19cdb489
MM		 32682013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3269
3270 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3271 DFmode, DImode, and SFmode in the upper VSX registers based on the
3272 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3273 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3274 of the VSX register handling.
3275
3276 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3277 (f32_sv): Likewise.
3278 (zero_extendsidi2_lfiwzx): Add support for loading into the
3279 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3280 only do VSX memory options on Altivec registers.
3281 (extendsidi2_lfiwax): Likewise.
3282 (extendsfdf2_fpr): Likewise.
3283 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3284 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3285 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3286 (movdi_internal64): Likewise.
3287
d6d11272
XDL		 32882013-09-27 Xinliang David Li <davidxl@google.com>
3289
3290 * opts.c (finish_options): Adjust parameters
3291 according to vect cost model.
3292 (common_handle_option): Set dynamic vect cost
3293 model for FDO.
3294 targhooks.c (default_add_stmt_cost): Compute stmt cost
3295 unconditionally.
3296 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3297 Use helper function.
3298 * tree-vectorizer.h (unlimited_cost_model): New function.
3299 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3300 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3301 function.
3302 (vect_enhance_data_refs_alignment): Ditto.
3303 * flag-types.h: New enum.
3304 * common/config/i386/i386-common.c (ix86_option_init_struct):
3305 No need to initialize vect_cost_model flag.
3306 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3307 unconditionally.
3308
ac1857a3
DN		 33092013-09-27 Diego Novillo <dnovillo@google.com>
3310
3311 * gimple.h (enum ssa_mode): Remove.
3312
0d0e2af6
PM		 33132013-09-27 Paulo Matos <pmatos@broadcom.com>
3314
3315 * cfgloop.h (number_of_loops): Fix typo in check for null.
3316
09dc585e
JJ		 33172013-09-27 Jakub Jelinek <jakub@redhat.com>
3318
3319 PR middle-end/58551
3320 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3321 are children of outermost saved_cfun's loop, and set it up to
3322 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3323 if loop != loop0 and SESE region contains bbs that belong to loop0.
3324
ec5a3504
RS		 33252013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3326
3327 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3328 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3329 (set_address_segment, set_address_base, set_address_index)
3330 (set_address_disp): Accept the argument unconditionally.
3331 (baseness): Remove must_be_base_p and must_be_index_p checks.
3332 (decompose_normal_address): Classify as much as possible in the
3333 main loop.
3334
f91aec98
RS		 33352013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3336
3337 * cse.c (count_reg_usage): Handle INT_LIST.
3338 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3339 * reginfo.c (reg_scan_mark_refs): Likewise.
3340 * reload1.c (eliminate_regs_1): Likewise.
3341
a9195970
IS		 33422013-09-27 Iain Sandoe <iain@codesourcery.com>
3343
3344 PR middle-end/58547
3345 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3346 signed.
3347
e6f1c509
RB		 33482013-09-27 Richard Biener <rguenther@suse.de>
3349
3350 PR tree-optimization/58459
3351 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3352 restriction not propagating into loops.
3353
84f48495
FW		 33542013-09-26 Florian Weimer <fw@deneb.enyo.de>
3355
3356 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3357 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3358 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3359
74fc8b8a
RB		 33602013-09-26 Richard Biener <rguenther@suse.de>
3361
7a5911d3 3362 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3363
b4ada065
RB		 33642013-09-26 Richard Biener <rguenther@suse.de>
3365
3366 * alias.h (component_uses_parent_alias_set): Rename to ...
3367 (component_uses_parent_alias_set_from): ... this.
3368 * alias.c (component_uses_parent_alias_set): Rename to ...
3369 (component_uses_parent_alias_set_from): ... this and return
3370 the desired parent.
3371 (reference_alias_ptr_type_1): Use the result from
3372 component_uses_parent_alias_set_from instead of stripping
3373 components one at a time.
3374 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3375
78bca40d
AM		 33762013-09-26 Andrew MacLeod <amacleod@redhat.com>
3377
7a5911d3
UB		 3378 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3379 Move prototypes to...
78bca40d
AM		 3380 * tree-ssa-ter.h: New File. Move prototypes here.
3381 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3382 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3383 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3384 from is_replaceable_p.
78bca40d
AM		 3385 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3386 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3387 (process_replaceable): Use ter_is_replaceable_p.
3388 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3389 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3390 newly refactored ssa_is_replaceable_p.
3391 * cfgexpand.c: Include tree-outof-ssa.h.
3392 * ssaexpand.h: Delete.
3393
ff2a63a7
AM		 33942013-09-26 Andrew MacLeod <amacleod@redhat.com>
3395
3396 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3397 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3398 tree-ssa.c
3399 (create_gimple_tmp): Delete.
3400 (get_expr_type, build_assign, build_type_cast): Move to...
3401 * gimple-builder.c: New File.
3402 (get_expr_type): Relocate from gimple.c.
3403 (build_assign, build_type_cast): Change to only create ssanames.
3404 * gimple.h: Move prototypes to...
3405 * gimple-builder.h: New File. Here.
3406 * tree-ssa.h: And here.
3407 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3408 count_uses_and_derefs): Relocate from gimple.c.
3409 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3410 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3411 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3412 gimple_replace_ssa_lhs.
3413 * asan.c: Include gimple-builder.h.
3414 * Makefile.in: Add gimple-builder.o.
3415
a2544177
RB		 34162013-09-26 Richard Biener <rguenther@suse.de>
3417
3418 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3419 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3420 (loe_visit_block): Use gcc_checking_assert.
3421 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3422 gimple_assign_ssa_name_copy_p.
3423 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3424 change.
3425
6d1ca9a0
DE		 34262013-09-26 David Edelsohn <dje.gcc@gmail.com>
3427
3428 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3429 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3430
46e950db
RB		 34312013-09-26 Richard Biener <rguenther@suse.de>
3432
3433 PR tree-optimization/58539
3434 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3435 the fact that debug statements are not taking part in loop-closed
3436 SSA construction.
3437
40ada30a
NC		 34382013-09-26 Nick Clifton <nickc@redhat.com>
3439
3440 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3441 time warning message.
3442 (msp430_print_operand_raw): Delete unused letter parameter.
3443 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3444 (msp430_print_operand_address): New function.
3445 (msp430_print_operand): Move address printing code from here to
3446 new function.
3447 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3448 assembler.
3449 (zero_extendpsisi2): Likewise.
3450 (extendpsisi2): New pattern.
3451 (andneghi3): New pattern.
3452
38e8f663
YR		 34532013-09-26 Yvan Roux <yvan.roux@linaro.org>
3454
3455 * config/aarch64/aarch64.opt (mlra): New option.
3456 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3457 (TARGET_LRA_P): Define.
3458
4167e8d4
EB		 34592013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3460
3461 * expr.c (expand_assignment): Remove obsolete comment.
3462
5254eac4
JL		 34632013-09-25 Jeff Law <law@redhat.com>
3464
3465 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3466 tree-ssa-threadupdate.h.
3467 (register_jump_thread): Similarly.
3468 * tree-ssa-threadupdate.h: New header file.
3469 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3470 * tree-vrp.c: Likewise.
3471 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3472 (thread_around_empty_blocks): Change type of path vector argument to
3473 an edge,type pair from just an edge. Initialize both elements when
3474 appending to a jump threading path. Tweak references to elements
3475 appropriately.
3476 (thread_across_edge): Similarly. Release memory for the elements
3477 as needed.
3478 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3479 (dump_jump_thread_path): New function broken out from
3480 register_jump_thread.
3481 (register_jump_thread): Use dump_jump_thread_path. Change type of
3482 path vector entries. Search the path for NULL edges and dump
3483 the path if one is found. Tweak the conversion of path to 3-edge
3484 form to use the block copy type information embedded in the path.
3485
20f114a3
YR		 34862013-09-25 Yvan Roux <yvan.roux@linaro.org>
3487
3488 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3489
3936bafc
YR		 34902013-09-25 Yvan Roux <yvan.roux@linaro.org>
3491 Vladimir Makarov <vmakarov@redhat.com>
3492
3493 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3494 from the least significant bit.
3495 (strip_address_mutations): Add bitfield operations handling.
3496 (must_be_index_p): Add shifting and rotate operations handling.
3497 (set_address_base): Use must_be_base_p predicate.
4167e8d4 3498 (set_address_index): Use must_be_index_p predicate.
3936bafc 3499
b86f6e9e
AI		 35002013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3501 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3502 Sergey Lega <sergey.s.lega@intel.com>
3503 Anna Tikhonova <anna.tikhonova@intel.com>
3504 Ilya Tocar <ilya.tocar@intel.com>
3505 Andrey Turetskiy <andrey.turetskiy@intel.com>
3506 Ilya Verbin <ilya.verbin@intel.com>
3507 Kirill Yukhin <kirill.yukhin@intel.com>
3508 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3509
3510 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3511 Use new names.
3512 (ix86_expand_vector_move_misalign): Support new unaligned load and
3513 stores and use new names.
3514 (CODE_FOR_sse2_storedqu): Rename to ...
3515 (CODE_FOR_sse2_storedquv16qi): ... this.
3516 (CODE_FOR_sse2_loaddqu): Rename to ...
3517 (CODE_FOR_sse2_loaddquv16qi): ... this.
3518 (CODE_FOR_avx_loaddqu256): Rename to ...
3519 (CODE_FOR_avx_loaddquv32qi): ... this.
3520 (CODE_FOR_avx_storedqu256): Rename to ...
3521 (CODE_FOR_avx_storedquv32qi): ... this.
3522 * config/i386/i386.md (fpint_logic): New.
3523 * config/i386/sse.md (VMOVE): Extend for AVX512.
3524 (VF): Ditto.
3525 (VF_128_256): New.
3526 (VF_512): Ditto.
3527 (VI_UNALIGNED_LOADSTORE): Ditto.
3528 (sse2_avx_avx512f): Ditto.
3529 (sse2_avx2): Extend for AVX512.
3530 (sse4_1_avx2): Ditto.
3531 (avx2_avx512f): New.
3532 (sse): Extend for AVX512.
3533 (sse2): Ditto.
3534 (sse4_1): Ditto.
3535 (avxsizesuffix): Ditto.
3536 (sseintvecmode): Ditto.
3537 (ssePSmode): Ditto.
3538 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3539 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3540 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3541 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3542 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3543 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3544 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3545 (STORENT_MODE): Extend for AVX512.
3546 (*absneg<mode>2): Replace constraint "x" with "v".
3547 (*mul<mode>3): Ditto.
3548 (*ieee_smin<mode>3): Ditto.
3549 (*ieee_smax<mode>3): Ditto.
3550 (avx_cmp<mode>3): Replace VF with VF_128_256.
3551 (*<sse>_maskcmp<mode>3_comm): Ditto.
3552 (<sse>_maskcmp<mode>3): Ditto.
3553 (<sse>_andnot<mode>3): Extend for AVX512.
3554 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3555 (<code><mode>3, fpint_logic): New.
3556 (*<code><mode>3): Extend for AVX512.
3557 (avx512flogicsuff): New.
3558 (avx512f_<logic><mode>): Ditto.
3559 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3560 VF_128_256.
3561 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3562 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3563 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3564 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3565 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3566 (xop_vpermil2<mode>3): Ditto.
3567 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3568 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3569 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3570 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3571
4d44d03c
TT		 35722013-09-25 Tom Tromey <tromey@redhat.com>
3573
3574 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3575 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3576 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3577 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3578 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3579 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3580 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3581 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3582 (GRAPHITE_HTAB_H): Remove.
3583
86aaf68d
TT		 35842013-09-25 Tom Tromey <tromey@redhat.com>
3585
3586 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3587
c987deb8
TT		 35882013-09-25 Tom Tromey <tromey@redhat.com>
3589
3590 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3591
9439747e
TT		 35922013-09-25 Tom Tromey <tromey@redhat.com>
3593
3594 * config/i386/t-i386 (i386.o): Remove.
3595 (i386-c.o): Use COMPILE and POSTCOMPILE.
3596
aefc31a1
TT		 35972013-09-25 Tom Tromey <tromey@redhat.com>
3598
3599 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3600
b6541edc
TT		 36012013-09-25 Tom Tromey <tromey@redhat.com>
3602
3603 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3604 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3605 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3606 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3607 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3608 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3609 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3610 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3611 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3612 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3613 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3614 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3615 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3616 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3617 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3618 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3619 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3620 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3621 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3622 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3623 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3624 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3625 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3626 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3627 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3628 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3629 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3630 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3631 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3632 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3633 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3634 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3635 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3636 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3637 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3638 (graphite-clast-to-gimple.o, graphite-dependences.o)
3639 (graphite-interchange.o, graphite-poly.o)
3640 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3641 (graphite-optimize-isl.o, tree-vect-loop.o)
3642 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3643 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3644 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3645 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3646 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3647 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3648 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3649 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3650 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3651 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3652 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3653 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3654 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3655 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3656 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3657 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3658 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3659 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3660 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3661 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3662 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3663 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3664 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3665 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3666 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3667 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3668 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3669 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3670 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3671 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3672 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3673 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3674 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3675 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3676 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3677 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3678 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3679 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3680 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3681 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3682 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3683 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3684 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3685 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3686 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3687 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3688 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3689 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3690 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3691 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3692 (gcov-dump.o): Remove.
3693 (default-c.o): Use COMPILE and POSTCOMPILE.
3694 (CFLAGS-gcc.o): New variable.
3695 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3696
8c796d19
TT		 36972013-09-25 Tom Tromey <tromey@redhat.com>
3698
3699 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3700 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3701 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3702 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3703 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3704 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3705 (c-family/array-notation-common.o, c-family/stub-objc.o)
3706 (c-family/c-ubsan.o): Remove.
3707
0a6c2227
TT		 37082013-09-25 Tom Tromey <tromey@redhat.com>
3709
3710 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3711
f3bc55f0
TT		 37122013-09-25 Tom Tromey <tromey@redhat.com>
3713
3714 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3715 to add -DENABLE_SHARED_LIBGCC.
3716 (gcc.o): Don't use subshell.
3717
c11c795e
TT		 37182013-09-25 Tom Tromey <tromey@redhat.com>
3719
3720 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3721 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3722 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3723 * configure, config.in: Rebuild.
3724
043378c3
TT		 37252013-09-25 Tom Tromey <tromey@redhat.com>
3726
3727 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3728 (COMPILE, POSTCOMPILE): New variables.
3729 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3730 (DEPFILES): New variable.
3731 Include ".Po" files.
3732 * configure.ac: Add checks for dependency checking.
3733 * configure, aclocal.m4: Regenerate.
3734
eab34643
TT		 37352013-09-25 Tom Tromey <tromey@redhat.com>
3736
3737 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 3738 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 3739
d2db5e91
TT		 37402013-09-25 Tom Tromey <tromey@redhat.com>
3741
3742 * Makefile.in (generated_files): Add options.h,
3743 target-hooks-def.h, insn-opinit.h,
3744 common/common-target-hooks-def.h, pass-instances.def,
3745 c-family/c-target-hooks-def.h.
3746
af33629e
JL		 37472013-09-25 Jeff Law <law@redhat.com>
3748
3749 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3750 than foo[foo.length () - 1] to access last member in a vec.
3751 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3752
ed0d3051
RB		 37532013-09-25 Richard Biener <rguenther@suse.de>
3754
3755 PR middle-end/58521
3756 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3757
123485ca
JH		 37582013-09-25 Jan Hubicka <jh@suse.cz>
3759
3760 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3761 test.
3762
d70b0f1f
MP		 37632013-09-25 Marek Polacek <polacek@redhat.com>
3764
3765 PR sanitizer/58420
3766 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3767 when determining the type name.
3768
e3f02798
OE		 37692013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3770
3771 * config/sh/sh.md: Fix formatting.
3772
7a5911d3 37732013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 3774
7a5911d3
UB		 3775 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3776 max peel iterations parameter.
4f17aa0b
XDL		 3777 * param.def: New parameter.
3778 * doc/invoke.texi: Document New parameter.
3779
0498471b
CL		 37802013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3781
3782 * gimple-pretty-print.c: Various whitespace tweaks.
3783 * tree-core.h: Likewise.
3784 * tree-pretty-print.c: Likewise.
3785 * tree-ssa-alias.c: Likewise.
3786 * tree-ssa-copy.c: Likewise.
3787 * tree-ssanames.c: Likewise.
3788 * tree-ssanames.h: Likewise.
3789 * tree-vrp.c: Likewise.
3790
5a8c2b57
AM		 37912013-09-24 Alan Modra <amodra@gmail.com>
3792
3793 PR middle-end/57134
3794 PR middle-end/57586
7a5911d3
UB		 3795 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3796 for output operands that disallow regs. Don't use EXPAND_WRITE on
3797 inout operands.
5a8c2b57 3798
f40333af
RB		 37992013-09-24 Richard Biener <rguenther@suse.de>
3800
3801 PR middle-end/58513
3802 * tree.c (reference_alias_ptr_type): Move ...
3803 * alias.c (reference_alias_ptr_type): ... here and implement
3804 in terms of the new reference_alias_ptr_type_1.
3805 (ref_all_alias_ptr_type_p): New helper.
3806 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3807 use ref_all_alias_ptr_type_p.
3808 (get_deref_alias_set): Add flag_strict_aliasing check here.
3809 (reference_alias_ptr_type_1): New function, split out from ...
3810 (get_alias_set): ... here.
3811 (alias_ptr_types_compatible_p): New function.
3812 * alias.h (reference_alias_ptr_type): Declare.
3813 (alias_ptr_types_compatible_p): Likewise.
3814 * tree.h (reference_alias_ptr_type): Remove.
3815 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3816 to compare MEM_REF alias types.
3817
583e8bf5
RB		 38182013-09-24 Richard Biener <rguenther@suse.de>
3819
3820 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3821
5e8586d7
MM		 38222013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3823
19c3e797
MM		 3824 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3825 reload helper function arrays into a single array reg_addr.
3826 (reload_fpr_gpr): Likewise.
3827 (reload_gpr_vsx): Likewise.
3828 (reload_vsx_gpr): Likewise.
3829 (struct rs6000_reg_addr): Likewise.
3830 (reg_addr): Likewise.
3831 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3832 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3833 (rs6000_init_hard_regno_mode_ok): Likewise.
3834 (rs6000_secondary_reload_direct_move): Likewise.
3835 (rs6000_secondary_reload): Likewise.
3836
5e8586d7
MM		 3837 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3838 constraints: wu, ww, and wy. Repurpose wv constraint added during
3839 power8 changes. Put wg constraint in alphabetical order.
3840
3841 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3842 for future work to add ISA 2.07 VSX single precision support.
3843 (-mvsx-scalar-double): Change default from -1 to 1, update
3844 documentation comment.
3845 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3846 (-mupper-regs-df): New debug switch to control whether DF values
3847 can go in the traditional Altivec registers.
3848 (-mupper-regs-sf): New debug switch to control whether SF values
3849 can go in the traditional Altivec registers.
3850
3851 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3852 and wy constraints.
3853 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3854 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3855 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3856 (rs6000_debug_legitimate_address_p): Print if we are running
3857 before, during, or after reload.
3858 (rs6000_secondary_reload): Add a comment.
3859 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3860
3861 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3862 constraints. Update documentation string.
3863 (wd constraint): Likewise.
3864 (wf constraint): Likewise.
3865 (wg constraint): Likewise.
3866 (wn constraint): Likewise.
3867 (ws constraint): Likewise.
3868 (wt constraint): Likewise.
3869 (wx constraint): Likewise.
3870 (wz constraint): Likewise.
3871 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3872 instructions.
3873 (ww constraint): Likewise.
3874 (wy constraint): Likewise.
3875 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3876 the previous submissions.
3877 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3878
e5af9ddd
RS		 38792013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3880
3881 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3882 (REG_BR_PROB): Say that the probability is stored in an int_list.
3883 * reg-notes.def: Update commentary to mention INT_LIST.
3884 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3885 (INT_LIST): New rtx.
3886 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3887 * rtlanal.c (int_reg_note_p): New function.
3888 (alloc_reg_note): Assert that the note does not have an int argument.
3889 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3890 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3891 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3892 rather than an INSN_LIST. Handle INT_LIST.
3893 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3894 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3895 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3896 Manipulate them as ints rather than rtxes.
3897 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3898 * regmove.c (copy_src_to_dest): Likewise.
3899 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3900
3901 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3902 into the cases that need it.
3903 * config/arm/arm.c (arm_unwind_emit): Likewise.
3904
3905 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3906 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3907 * loop-doloop.c (add_test, doloop_modify): Likewise.
3908 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3909 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3910 * predict.c (combine_predictions_for_insn): Likewise.
3911 * print-rtl.c (print_rtx): Handle INT_LIST.
3912 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3913 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3914 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3915 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3916 (ix86_split_fp_branch, predict_jump): Likewise.
3917 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3918 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3919 * config/spu/spu.c (ea_load_store_inline): Likewise.
3920
3921 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3922 value of a REG_BR_PROB note.
3923 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3924 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3925 * emit-rtl.c (try_split): Likewise.
3926 * predict.c (br_prob_note_reliable_p): Likewise.
3927 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
3928 * reorg.c (mostly_true_jump): Likewise.
3929 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
3930 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
3931 * config/i386/i386.c (ix86_print_operand): Likewise.
3932 * config/ia64/ia64.c (ia64_print_operand): Likewise.
3933 * config/mmix/mmix.c (mmix_print_operand): Likewise.
3934 * config/rs6000/rs6000.c (output_cbranch): Likewise.
3935 * config/s390/s390.c (s390_expand_tbegin): Likewise.
3936 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
3937 * config/sparc/sparc.c (output_cbranch): Likewise.
3938 * config/spu/spu.c (get_branch_target): Likewise.
3939 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
3940 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
3941
450ad0cd
JH		 39422013-09-23 Jan Hubicka <jh@suse.cz>
3943
3944 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
3945 for ipa-devirt.
3946 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
3947 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
3948 of external calls
3949 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
3950 (gimple_fold_call): Dump inconsistent devirtualizations; add
3951 sanity check for type based devirtualizations.
3952 * ipa-prop.c: Include ipa-utils.h
3953 (ipa_intraprocedural_devirtualization): Add sanity check.
3954 (try_make_edge_direct_virtual_call): Likewise.
3955
cefce769
EB		 39562013-09-23 Eric Botcazou <ebotcazou@adacore.com>
3957
3958 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
3959 assignment statements.
3960
a895a2b8
KV		 39612013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
3962
3963 * gimple-pretty-print.c (dump_ssaname_info): New function.
3964 (dump_gimple_phi): Call it.
3965 (pp_gimple_stmt_1): Likewise.
3966 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
3967 (range_info_def): Declare.
3968 * tree-pretty-print.c (pp_double_int): New function.
3969 (dump_generic_node): Call it.
3970 * tree-pretty-print.h (pp_double_int): Declare.
3971 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
3972 * tree-ssanames.h (range_info_def): New structure.
3973 (value_range_type): Move definition here.
3974 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
3975 Declare.
3976 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
3977 initialization.
3978 (set_range_info): New function.
3979 (get_range_info): Likewise.
3980 (duplicate_ssa_name_range_info): Likewise.
3981 (duplicate_ssa_name_fn): Check pointer type and call
3982 duplicate_ssa_name_range_info.
3983 * tree-ssa-copy.c (fini_copy_prop): Likewise.
3984 * tree-vrp.c (value_range_type): Remove definition, now in
3985 tree-ssanames.h.
7a5911d3 3986 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV		 3987 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
3988 (SSA_NAME_RANGE_INFO): New macro.
3989
984af6ac
RB		 39902013-09-23 Richard Biener <rguenther@suse.de>
3991
3992 PR tree-optimization/58464
3993 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 3994 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB		 3995 (phi_translate_1): Remove recursion detection here.
3996 (phi_translate): Pre-seed the cache with NULL to catch
3997 recursion here in a more generic way.
3998 (bitmap_find_leader): Adjust comment.
3999 (get_representative_for): Dump value-numbers.
4000 (create_expression_by_pieces): Likewise.
4001 (insert_into_preds_of_block): Likewise.
4002
0e26cf79
CB		 40032013-09-23 Christian Bruel <christian.bruel@st.com>
4004
4005 PR target/58475
4006 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4007 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4008
4d0e5b54
JG		 40092013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4010
4011 Revert r202780:
4012 2013-09-20 Renlin Li <renlin.li@arm.com>
4013
7a5911d3
UB		 4014 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4015 plus_constant.
4d0e5b54
JG		 4016 (aarch64_expand_epilogue): Likewise.
4017 (aarch64_legitimize_reload_address): Likewise.
4018
ca081cc8
EB		 40192013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4020
4021 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4022 NULL_TREE before pushing them onto the vector. Likewise for labels.
4023
8fc53a5f
EB		 40242013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4025
4026 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4027 comparison operators when -fno-trapping-math is in effect.
4028 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4029 comparison operators in TFmode and assert that unsupported operators
4030 cannot reach here.
4031 (ia64_print_operand): Likewise.
4032
40332013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH		 4034
4035 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4036 (sse_typeless_stores): Enable for core
4037 (sse_load0_by_pxor): Likewise.
4038 (four_jump_limit): Disable for core.
4039 (pad_returns): Likewise.
4040 (avoid_vector_decode): Likewise.
4041 (fuse_cmp_and_branch): Enable for cores.
4042 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4043
578e51e6
JDA		 40442013-09-20 John David Anglin <danglin@gcc.gnu.org>
4045
96ce28a3
JDA		 4046 PR middle-end/56791
4047 * config/pa/pa.c (pa_option_override): Disable auto increment and
4048 decrement instructions until reload is completed.
4049
578e51e6
JDA		 4050 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4051 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4052 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4053
6fcd3a13 40542013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 4055 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD		 4056
4057 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4058 (need_to_save): Save bank 0 on interrupts.
4059 (characterize_address): Strip far address wrappers.
4060 (rl78_as_legitimate_address): Likewise.
4061 (transcode_memory_rtx): Likewise.
4062 (rl78_peep_movhi_p): Disable this peephole after devirt.
4063 (rl78_propogate_register_origins): Forget all origins when a
4064 CLOBBER is seen.
4065 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4066 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4067 peephole2's.
4068 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4069 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4070 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4071 relocating.
4072 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4073
4a08db33
JDA		 40742013-09-20 John David Anglin <danglin@gcc.gnu.org>
4075
4076 * config/pa/pa.md: In "scc" insn patterns, change output template to
4077 handle const0_rtx in reg_or_0_operand operands.
4078
7691132c
JBG		 40792013-09-20 Martin Husemann <martin@NetBSD.org>
4080
4081 PR target/56875
4082 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4083 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4084
a3cc13cc
RB		 40852013-09-20 Richard Biener <rguenther@suse.de>
4086
4087 PR middle-end/58484
4088 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4089 remembering SSA name version and block index.
4090 (new_scev_info_str): Adjust.
4091 (hash_scev_info): Likewise. Also hash the block index.
4092 (eq_scev_info): Adjust.
4093 (find_var_scev_info): Likewise.
4094 (struct instantiate_cache_entry): Remove.
4095 (struct instantiate_cache_type): Use a htab to map name, block
4096 to chrec.
4097 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4098 (get_instantiated_value_entry): Likewise.
4099 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4100 (instantiate_scev_name): Adjust.
4101
925f3871
JL		 41022013-09-20 Jeff Law <law@redhat.com>
4103
4104 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4105
2888c331
YZ		 41062013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4107
4108 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4109 Call aarch64_simd_expand_args to update op[argc].
4110
0a811e96
BS		 41112013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4112
7a5911d3
UB		 4113 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4114 plugin argument.
0a811e96 4115
0078f462
BS		 41162013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4117
7a5911d3
UB		 4118 * gengtype.c (file_rules): Added rule for *.cc files.
4119 (get_output_file_with_visibility): Give fatal message when no
4120 rules found.
0078f462 4121
d20e48b0
RL		 41222013-09-20 Renlin Li <renlin.li@arm.com>
4123
4124 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4125 (aarch64_expand_epilogue): Likewise.
4126 (aarch64_legitimize_reload_address): Likewise.
4127
59bb1544
BE		 41282013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4129
4130 PR middle-end/57748
4131 * expr.c (expand_assignment): Remove misalignp code path.
4132
a5430019
MP		 41332013-09-20 Marek Polacek <polacek@redhat.com>
4134
4135 PR sanitizer/58413
4136 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4137 TYPE_PRECISION. Add asserts.
4138
9fed7f3a
RB		 41392013-09-20 Richard Biener <rguenther@suse.de>
4140
4141 PR tree-optimization/58453
4142 * tree-loop-distribution.c (distribute_loop): Apply the cost
4143 model for -ftree-loop-distribute-patterns, too.
4144
a820c834
RB		 41452013-09-20 Richard Biener <rguenther@suse.de>
4146
4147 PR middle-end/58473
4148 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4149 make type comparison less strict.
4150
3cbe17f7
AM		 41512013-09-20 Alan Modra <amodra@gmail.com>
4152
4153 * configure: Regenerate.
4154 * aclocal.m4: Regenerate.
4155
87fccdbb
MP		 41562013-09-20 Marek Polacek <polacek@redhat.com>
4157
4158 PR other/58467
4159 * doc/extend.texi: Document that attribute used is meant to be used
4160 on variables with static storage duration.
4161
8928eff3
JJ		 41622013-09-19 Jakub Jelinek <jakub@redhat.com>
4163
4164 PR tree-optimization/58472
4165 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4166 simd_lane_access set inv_p = false.
4167 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4168 the simduid magic VAR_DECL.
4169
b7b28c53
JH		 41702013-09-19 Jan Hubicka <jh@suse.cz>
4171
4172 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4173
83ae86f5
JL		 41742013-09-17 Jeff Law <law@redhat.com>
4175
4176 * tree-ssa-dom.c (record_temporary_equivalences): New function
4177 split out of dom_opt_dom_walker::after_dom_children.
4178 (dom_opt_dom_walker::thread_across_edge): Move common code
4179 in here from dom_opt_dom_walker::after_dom_children.
4180 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4181
9d532162
JH		 41822013-09-19 Jan Hubicka <jh@suse.cz>
4183
4184 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4185 (TARGET_GENERIC): Use PROCESOR_GENERIC
4186 (enum processor_type): Unify generic32 and 64.
4187 * i386.md (cpu): Likewise.
4188 * x86-tune.def (use_leave): Enable for generic32.
4189 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4190 * athlon.md: Change generic64 to generic in all occurences.
4191 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4192 (ix86_target_macros_internal): Likewise.
4193 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB		 4194 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4195 to ..
9d532162
JH		 4196 (generic_memcpy, generic_memset, generic_cost): This one.
4197 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4198 (m_GENERIC32, m_GENERIC64): Remove.
4199 (m_GENERIC): Turn into one flag.
4200 (processor_target): Unify generic tunnings.
4201 (ix86_option_override_internal): Replace generic32/64 by generic.
4202 (ix86_issue_rate): Likewise.
4203 (ix86_adjust_cost): Likewise.
4204
188c7d00
JH		 42052013-09-19 Jan Hubicka <jh@suse.cz>
4206
4207 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4208 of speculative flag.
4209
4befd127
JJ		 42102013-09-19 Jakub Jelinek <jakub@redhat.com>
4211
4212 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4213 GOMP_sections_start, even if !exit_reachable.
4214
8a26ad39
VM		 42152013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4216
4217 * lra-constraints.c (need_for_all_save_p): Use macro
4218 HARD_REGNO_CALL_PART_CLOBBERED.
4219 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4220 set up pseudo conflict hard regs.
4221
09487185
MM		 42222013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4223
4224 PR target/58452
4225 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4226 operands.
4227
080cbf9e
VM		 42282013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4229
4230 PR rtl-optimization/58438
4231 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4232 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4233 from previous subpasses.
4234
a0f70fcb
RE		 42352013-09-18 Richard Earnshaw <rearnsha@arm.com>
4236
4237 * arm.c (arm_get_frame_offsets): Validate architecture supports
4238 LDRD/STRD before accepting the tuning preference.
4239 (arm_expand_prologue): Likewise.
4240 (arm_expand_epilogue): Likewise.
4241
0547c9b6
RB		 42422013-09-18 Richard Biener <rguenther@suse.de>
4243
4244 PR tree-optimization/58417
4245 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4246 have chrecs with symbols defined in the loop as operands.
4247 (chrec_fold_multiply): Likewise.
4248 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4249 parameters before folding binary operations.
4250 (struct instantiate_cache_entry_hasher): Remove.
4251 (struct instantiate_cache_type): Use a pointer-map.
4252 (instantiate_cache_type::instantiate_cache_type): New function.
4253 (instantiate_cache_type::get): Likewise.
4254 (instantiate_cache_type::set): Likewise.
4255 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4256 (get_instantiated_value_entry): Likewise.
4257 (global_cache): New global.
4258 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4259 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4260 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4261 (instantiate_scev_name): Adjust.
4262 (instantiate_scev): Construct global instead of local cache.
4263 (resolve_mixers): Likewise.
4264
309dc1aa
DM		 42652013-09-18 Daniel Morris <danielm@ecoscentric.com>
4266 Paolo Carlini <paolo.carlini@oracle.com>
4267
4268 PR c++/58458
4269 * doc/implement-cxx.texi: Fix references to the C++ standards.
4270
995a1b4a
JJ		 42712013-09-18 Jakub Jelinek <jakub@redhat.com>
4272
4273 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4274 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4275 simd_lane_access drs, update also DR_ALIGNED_TO.
4276
ce6923c5
MP		 42772013-09-18 Marek Polacek <polacek@redhat.com>
4278
4279 PR sanitizer/58411
4280 * doc/extend.texi: Document no_sanitize_undefined attribute.
4281 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4282 no_sanitize_undefined attribute.
4283
d30d00a2
NC		 42842013-09-18 Nick Clifton <nickc@redhat.com>
4285
4286 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4287 (ASM_DECLARE_FUNCTION_NAME): Define.
4288
4d9192b5
TS		 42892013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4290
4291 * compare-elim.c (find_comparison_dom_walker): New class
4292 (find_comparisons_in_bb): Rename to
4293 find_comparison_dom_walker::before_dom_children
4294 (find_comparisons): Adjust
4295 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4296 adjust.
4297 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4298 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4299 (init_walk_dominator_tree): Remove declaration.
4300 (fini_walk_dominator_tree): Remove declaration.
4301 * fwprop.c (single_def_use_dom_walker): New class
4302 (single_def_use_enter_block): Convert to
4303 single_def_use_dom_walker::before_dom_children.
4304 (single_def_use_leave_block): Convert to
4305 single_def_use_dom_walker::after_dom_children.
4306 (build_single_def_use_links): Adjust.
4307 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4308 class.
4309 (find_candidates_in_block): Convert to
4310 find_candidates_dom_walker::before_dom_children.
4311 (execute_strength_reduction): Adjust.
4312 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4313 (sese_dom_walker): New class.
4314 (sese_dom_walker::sese_dom_walker): New constructor.
4315 (sese_dom_walker::~sese_dom_walker): New destructor.
4316 (build_sese_conditions_before): Convert to
4317 sese_dom_walker::before_dom_children.
4318 (build_sese_conditions_after): Convert to
4319 sese_dom_walker::after_dom_children.
4320 (build_poly_scop): Adjust
4321 * tree-into-ssa.c (rewrite_dom_walker): New class
4322 (rewrite_enter_block): Convert to
4323 rewrite_dom_walker::before_dom_children.
4324 (rewrite_leave_block): Convert to
4325 rewrite_dom_walker::after_dom_children.
4326 (rewrite_update_dom_walker): New class.
4327 (rewrite_update_enter_block): Convert to
4328 rewrite_update_dom_walker::before_dom_children.
4329 (rewrite_update_leave_block): Convert to
4330 rewrite_update_dom_walker::after_dom_children.
4331 (rewrite_blocks, rewrite_into_ssa): Adjust.
4332 (mark_def_dom_walker): New class.
4333 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4334 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4335 (mark_def_sites_blocks): Convert to
4336 mark_def_dom_walker::before_dom_children.
4337 (mark_def_site_blocks): Remove.
4338 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4339 (tree_ssa_dominator_optimize): Adjust.
4340 (dom_thread_across_edge): Convert to method
4341 dom_opt_dom_walker::thread_across_edge.
4342 (dom_opt_enter_block): Convert to member function
4343 dom_opt_dom_walker::before_dom_children.
4344 (dom_opt_leave_block): Convert to member function
4345 dom_opt_dom_walker::after_dom_children.
4346 * tree-ssa-dse.c (dse_dom_walker): New class.
4347 (dse_enter_block): Convert to member function
4348 dse_dom_walker::before_dom_children.
4349 (tree_ssa_dse): Adjust.
4350 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4351 (determine_invariantness_stmt): Convert to method
4352 invariantness_dom_walker::before_dom_children.
4353 (determine_invariantness): Remove
4354 (move_computations_dom_walker): New class.
4355 (move_computations_stmt): Convert to method
4356 move_computations_dom_walker::before_dom_children.
4357 (move_computations, tree_ssa_lim): Adjust.
4358 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4359 (nt_init_block): Make method
4360 notrappping_dom_walker::before_dom_children.
4361 (nt_fini_block): Make
4362 method nontrapping_dom_walker::after_dom_children.
4363 (get_non_trapping): Adjust.
4364 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4365 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4366 (eliminate_leave_block): Make method.
4367 eliminate_dom_walker::after_dom_children.
4368 (eliminate): Adjust
4369 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4370 (strlen_enter_block): Make method
4371 strlen_dom_walker::before_dom_children.
4372 (strlen_leave_block): Make
4373 method strlen_dom_walker::after_dom_children.
4374 (tree_ssa_strlen): Adjust.
4375 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4376 (tree_ssa_uncprop): Adjust.
4377 (uncprop_leave_block): Make method
4378 uncprop_dom_walker::after_dom_children.
4379 (uncprop_leave_block): Make method
4380 uncprop_dom_walker::before_dom_children.
4381
3b3cc26b
BC		 43822013-09-18 Bin Cheng <bin.cheng@arm.com>
4383
7a5911d3
UB		 4384 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4385 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4386
b5ba3a9b
NC		 43872013-09-17 Nick Clifton <nickc@redhat.com>
4388
4389 * config/rl78/rl78-real.md (bf): New pattern.
4390 (bt): New pattern.
4391 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4392 (rl78_print_operand): Do not put a # before a %B.
4393 * config/rl78/rl78.opt: Tweak doc strings.
4394
09d5094b
DD		 43952013-09-17 DJ Delorie <dj@redhat.com>
4396
4397 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4398 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4399 (rl78_option_override): Likewise, if -mallregs.
4400 (is_virtual_register): Likewise.
4401 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4402 (REGNO_OK_FOR_BASE_P): Likewise.
4403 * config/rl78/rl78.opt (-mallregs): New.
4404
2ba1dca1
NC		 44052013-09-17 Nick Clifton <nickc@redhat.com>
4406
4407 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4408 For interrupt functions: save all call clobbered registers if the
4409 interrupt handler is not a leaf function.
4410 (rl78_expand_prologue): Always recompute the frame information.
4411 For interrupt functions: only select bank 0 if one of the bank 0
4412 registers is going to be psuhed.
4413
2e7c3f21
DD		 44142013-09-17 DJ Delorie <dj@redhat.com>
4415
4416 * config/rl78/constraints.md: For each W* constraint, rename to C*
4417 and create a W* constraint that checks for an optional ES: prefix
4418 pattern also.
4419 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4420 (es_addr): New. Used to wrap far addresses.
4421 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4422 (rl78_es_base): New.
4423 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4424 wrapped far addresses.
4425 (rl78_print_operand_1): Unwrap far addresses before processing.
4426 (rl78_lo16): Wrap far addresses in unspecs.
4427 (rl78_es_addr): New.
4428 (rl78_es_base): New.
4429 (insn_ok_now): Check for not-yet-wrapped far addresses.
4430 (transcode_memory_rtx): Properly re-wrap far addresses.
4431
5c26a69a
SH		 44322013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4433
4434 * config/sparc/t-rtems: Add leon3 multilibs.
4435
181f5f3e
CH		 44362013-09-17 Cong Hou <congh@google.com>
4437
4438 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4439 when checking the dot production pattern. The type of rhs operand
4440 of multiply is now checked correctly.
4441
8d34e421
JL		 44422013-09-17 Jeff Law <law@redhat.com>
4443
4444 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4445 edge implied equivalences into successor phis.
4446 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4447 here from tree-ssa-threadedge.c.
4448 (mark_threaded_blocks): When threading through a joiner, if both
4449 successors of the joiner's clone reach the same block, verify the
4450 PHI arguments are equal. If not, cancel the jump threading request.
4451 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4452 tree-ssa-threadupdate.c
4453 (thread_across_edge): Don't check PHI argument equality when
4454 threading through joiner block here.
7a5911d3 4455
c152901f
AM		 44562013-09-17 Andrew MacLeod <amacleod@redhat.com>
4457
4458 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4459 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB		 4460 (warn_uninit, warn_uninitialized_vars,
4461 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4462 Move to tree-ssa-uninit.c.
78cedfb1 4463 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 4464 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM		 4465 ssa_undefined_value_p.
4466 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB		 4467 (warn_uninit, warn_uninitialized_vars,
4468 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4469 Move from tree-ssa.c.
4470 * tree-ssa.h: Adjust prototypes.
c152901f 4471
b631d45a
JH		 44722013-09-17 Jan Hubicka <jh@suse.cz>
4473
4474 PR middle-end/58332
4475 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4476 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4477 FUNCTION_NOT_OPTIMIZED.
4478 * ipa-inline-analysis.c (compute_inline_parameters): Function
4479 not optimized is not inlinable unless it is alwaysinline.
4480 (inline_analyze_function): Force calls in not optimized
4481 function not inlinable.
4482
5b79657a
JH		 44832013-09-17 Jan Hubicka <jh@suse.cz>
4484
4485 PR middle-end/58329
4486 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4487 to return NULL.
4488 * ipa.c (function_and_variable_visibility): Likewise.
4489 * ipa-profile.c (ipa_profile): Likewise.
4490
27442c24
BE		 44912013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4492
4493 PR ipa/58398
4494 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4495 attribute, and don't inline the resolver in this case.
4496
753b9bc3
TJ		 44972013-09-17 Teresa Johnson <tejohnson@google.com>
4498
4499 * coverage.c (get_coverage_counts): Add missing newline.
4500
e3d3cfb4
KT		 45012013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4502
4503 PR tree-optimization/58088
4504 * fold-const.c (mask_with_trailing_zeros): New function.
4505 (fold_binary_loc): Make sure we don't recurse infinitely
4506 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4507 Use mask_with_trailing_zeros where appropriate.
4508
9a0ac98f
YR		 45092013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4510
4511 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4512 of 'prev' var to get better distance estimation.
4513
78bbd765
EB		 45142013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4515
4516 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4517 * tree-inline.c (is_parameter_of): New predicate.
4518 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4519 a parameter has been remapped.
4520 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4521 (optimize_inline_calls): Initialize transform_parameter.
4522 (copy_gimple_seq_and_replace_locals): Likewise.
4523 (tree_function_versioning): Likewise.
4524 (maybe_inline_call_in_expr): Likewise.
4525
cad055a4
NC		 45262013-09-17 Nick Clifton <nickc@redhat.com>
4527
4528 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4529 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4530 interrupt handlers.
4531 (is_attr_func): New function.
4532 (msp430_is_interrupt_func): New function.
4533 (is_naked_func): New function.
4534 (is_reentrant_func): New function.
4535 (is_critical_func): New function.
4536 (msp430_start_function): Add annotations for function attributes.
4537 (msp430_attr): New function.
4538 (msp430_attribute_table): New.
4539 (msp430_function_section): New function.
4540 (TARGET_ASM_FUNCTION_SECTION): Define.
4541 (msp430_builtin): New enum.
4542 (msp430_init_builtins): New function.
4543 (msp430_builtin_devl): New function.
4544 (msp430_expand_builtin): New function.
4545 (TARGET_INIT_BUILTINS): Define.
4546 (TARGET_EXPAND_BUILTINS): Define.
4547 (TARGET_BUILTIN_DECL): Define.
4548 (msp430_expand_prologue): Add support for naked, interrupt,
4549 critical and reentrant functions.
4550 (msp430_expand_epilogue): Likewise.
4551 (msp430_print_operand): Handle 'O' character.
4552 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4553 NO_TRAMPOLINES.
4554 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4555 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4556 (pushm): Use a 'n' rather than an 'i' constraint.
4557 (msp_return): Add generation of the interrupt return instruction.
4558 (disable_interrupts): New pattern.
4559 (enable_interrupts): New pattern.
4560 (push_intr_state): New pattern.
4561 (pop_intr_state): New pattern.
4562 (bic_SR): New pattern.
4563 (bis_SR): New pattern.
4564 * doc/extend.texi: Document MSP430 function attributes and builtin
4565 functions.
4566
deb6c11a
RB		 45672013-09-17 Richard Biener <rguenther@suse.de>
4568
4569 PR tree-optimization/58432
4570 * tree-loop-distribution.c (tree_loop_distribution): Also
4571 scan PHIs for outside loop uses and seed a partition from them.
4572
c068654b
BC		 45732013-09-17 Bin Cheng <bin.cheng@arm.com>
4574
4575 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4576 (restructure_reference): Call backtrace_base_for_ref.
4577
567969e4
AM		 45782013-09-17 Alan Modra <amodra@gmail.com>
4579
4580 PR target/57589
7a5911d3
UB		 4581 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4582 patch.
567969e4 4583
5c0029de
DD		 45842013-09-16 DJ Delorie <dj@redhat.com>
4585
4586 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4587 vregs location for RL78/G10.
4588 (rl78_expand_prologue): Avoid SEL on G10.
4589 (rl78_expand_epilogue): Likewise.
4590 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4591 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4592 __RL78_G10__ when appropriate.
4593 (ASM_SPEC): Pass -mg10 along to the assembler.
4594 * config/rl78/rl78.md (sel_rb): Disable for G10.
4595 * config/rl78/rl78.opt: Add -mg10 option.
4596 * config/rl78/t-rl78: Add -mg10 multilib.
4597
ea0f3e87
XDL		 45982013-09-16 Xinliang David Li <davidxl@google.com>
4599
4600 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4601 * omp-low.c (omp_max_vf): Ditto.
4602 (expand_omp_simd): Ditto.
4603 * tree-vectorizer.c (vectorize_loops): Ditto.
4604 (gate_vect_slp): Ditto.
4605 (gate_increase_alignment): Ditto.
4606 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4607 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4608 (gate_tree_vectorize): Name change.
4609 (tree_vectorize): Ditto.
4610 (pass_vectorize::gate): Call new function.
4611 (pass_vectorize::execute): Ditto.
7a5911d3 4612 * opts.c: O3 default setting change.
ea0f3e87
XDL		 4613 (finish_options): Check new flag.
4614 * doc/invoke.texi: Document new flags.
4615 * common.opt: New flags.
4616
4f3da2e5
AS		 46172013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4618
4619 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4620 (Condition Code): Don't reference it.
4621 * doc/tm.texi: Regenerate.
4622
aa940f7c
VM		 46232013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4624
4625 PR middle-end/58418
4626 * lra-constraints.c (undo_optional_reloads): Consider all optional
4627 reload even if it did not get a hard reg.
4628
e645e942
TJ		 46292013-09-16 Teresa Johnson <tejohnson@google.com>
4630
4631 * dumpfile.c (dump_loc): Remove newline emission.
4632 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4633 emission to dump_printf_loc calls where missing.
4634 (vect_mark_for_runtime_alias_test): Ditto.
4635 (vect_analyze_data_ref_dependence): Ditto.
4636 (vect_analyze_data_ref_dependences): Ditto.
4637 (vect_slp_analyze_data_ref_dependence): Ditto.
4638 (vect_slp_analyze_data_ref_dependences): Ditto.
4639 (vect_compute_data_ref_alignment): Ditto.
4640 (vect_update_misalignment_for_peel): Ditto.
4641 (vect_verify_datarefs_alignment): Ditto.
4642 (vector_alignment_reachable_p): Ditto.
4643 (vect_get_data_access_cost): Ditto.
4644 (vect_enhance_data_refs_alignment): Ditto.
4645 (vect_find_same_alignment_drs): Ditto.
4646 (vect_analyze_data_refs_alignment): Ditto.
4647 (vect_analyze_group_access): Ditto.
4648 (vect_analyze_data_ref_access): Ditto.
4649 (vect_analyze_data_ref_accesses): Ditto.
4650 (vect_prune_runtime_alias_test_list): Ditto.
4651 (vect_analyze_data_refs): Ditto.
4652 (vect_create_addr_base_for_vector_ref): Ditto.
4653 (vect_create_data_ref_ptr): Ditto.
4654 (vect_grouped_store_supported): Ditto.
4655 (vect_grouped_load_supported): Ditto.
4656 * value-prof.c (check_counter): Ditto.
4657 (check_ic_target): Ditto.
4658 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4659 (vect_recog_widen_mult_pattern): Ditto.
4660 (vect_recog_widen_sum_pattern): Ditto.
4661 (vect_recog_over_widening_pattern): Ditto.
4662 (vect_recog_widen_shift_pattern): Ditto.
4663 (vect_recog_rotate_pattern): Ditto.
4664 (vect_recog_vector_vector_shift_pattern): Ditto.
4665 (vect_recog_divmod_pattern): Ditto.
4666 (vect_recog_mixed_size_cond_pattern): Ditto.
4667 (vect_recog_bool_pattern): Ditto.
4668 (vect_pattern_recog_1): Ditto.
4669 (vect_pattern_recog): Ditto.
4670 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4671 (vect_is_simple_iv_evolution): Ditto.
4672 (vect_analyze_scalar_cycles_1): Ditto.
4673 (vect_get_loop_niters): Ditto.
4674 (vect_analyze_loop_1): Ditto.
4675 (vect_analyze_loop_form): Ditto.
4676 (vect_analyze_loop_operations): Ditto.
4677 (vect_analyze_loop_2): Ditto.
4678 (vect_analyze_loop): Ditto.
4679 (report_vect_op): Ditto.
4680 (vect_is_slp_reduction): Ditto.
4681 (vect_is_simple_reduction_1): Ditto.
4682 (vect_get_known_peeling_cost): Ditto.
4683 (vect_estimate_min_profitable_iters): Ditto.
4684 (vect_model_reduction_cost): Ditto.
4685 (vect_model_induction_cost): Ditto.
4686 (get_initial_def_for_induction): Ditto.
4687 (vect_create_epilog_for_reduction): Ditto.
4688 (vectorizable_reduction): Ditto.
4689 (vectorizable_induction): Ditto.
4690 (vectorizable_live_operation): Ditto.
4691 (vect_loop_kill_debug_uses): Ditto.
4692 (vect_transform_loop): Ditto.
4693 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4694 (vect_stmt_relevant_p): Ditto.
4695 (process_use): Ditto.
4696 (vect_mark_stmts_to_be_vectorized): Ditto.
4697 (vect_model_simple_cost): Ditto.
4698 (vect_model_promotion_demotion_cost): Ditto.
4699 (vect_model_store_cost): Ditto.
4700 (vect_get_store_cost): Ditto.
4701 (vect_model_load_cost): Ditto.
4702 (vect_get_load_cost): Ditto.
4703 (vect_init_vector_1): Ditto.
4704 (vect_get_vec_def_for_operand): Ditto.
4705 (vect_finish_stmt_generation): Ditto.
4706 (vectorizable_call): Ditto.
4707 (vectorizable_conversion): Ditto.
4708 (vectorizable_assignment): Ditto.
4709 (vectorizable_shift): Ditto.
4710 (vectorizable_operation): Ditto.
4711 (vectorizable_store): Ditto.
4712 (vectorizable_load): Ditto.
4713 (vectorizable_condition): Ditto.
4714 (vect_analyze_stmt): Ditto.
4715 (vect_transform_stmt): Ditto.
4716 (vect_is_simple_use): Ditto.
4717 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4718 (vect_can_advance_ivs_p): Ditto.
4719 (vect_update_ivs_after_vectorizer): Ditto.
4720 (vect_do_peeling_for_loop_bound): Ditto.
4721 (vect_gen_niters_for_prolog_loop): Ditto.
4722 (vect_update_inits_of_drs): Ditto.
4723 (vect_create_cond_for_alias_checks): Ditto.
4724 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4725 (vect_build_slp_tree_1): Ditto.
4726 (vect_supported_load_permutation_p): Ditto.
4727 (vect_analyze_slp_instance): Ditto.
4728 (vect_analyze_slp): Ditto.
4729 (vect_make_slp_decision): Ditto.
4730 (vect_detect_hybrid_slp): Ditto.
4731 (vect_bb_vectorization_profitable_p): Ditto.
4732 (vect_slp_analyze_bb_1): Ditto.
4733 (vect_update_slp_costs_according_to_vf): Ditto.
4734 (vect_get_mask_element): Ditto.
4735 (vect_transform_slp_perm_load): Ditto.
4736 (vect_schedule_slp_instance): Ditto.
4737 (vect_schedule_slp): Ditto.
4738 (vect_slp_transform_bb): Ditto.
4739 * profile.c (read_profile_edge_counts): Ditto.
4740 (compute_branch_probabilities): Ditto.
4741 * coverage.c (get_coverage_counts): Ditto.
4742
46f851f3
DN		 47432013-09-16 Diego Novillo <dnovillo@google.com>
4744
7a5911d3 4745 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 4746
1632a686
JH		 47472013-09-16 Jan Hubicka <jh@suse.cz>
4748
4749 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 4750 abstract functions; for static functions check the presence of body.
1632a686 4751
828e70c1
JG		 47522013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4753
4754 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4755 * config/aarch64/aarch64-simd.md
4756 (aarch64_mla_elt<mode>): New.
4757 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4758 (aarch64_mls_elt<mode>): Likewise.
4759 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4760 (aarch64_fma4_elt<mode>): Likewise.
4761 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4762 (aarch64_fma4_elt_to_128v2df): Likewise.
4763 (aarch64_fma4_elt_to_64df): Likewise.
4764 (fnma<mode>4): Likewise.
4765 (aarch64_fnma4_elt<mode>): Likewise.
4766 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4767 (aarch64_fnma4_elt_to_128v2df): Likewise.
4768 (aarch64_fnma4_elt_to_64df): Likewise.
4769 * config/aarch64/iterators.md (VDQSF): New.
4770 * config/aarch64/arm_neon.h
4771 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4772 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4773
779aea46
JG		 47742013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4775
4776 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4777 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4778 (aarch64_mul3_elt_to_128df): Likewise.
4779 (aarch64_mul3_elt_to_64v2df): Likewise.
4780 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4781 (VMUL): New.
4782 (VMUL_CHANGE_NLANES) Likewise.
4783 (h_con): Likewise.
4784 (f): Likewise.
4785 * config/aarch64/arm_neon.h
4786 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4787
a407a750
JG		 47882013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4789
4790 * config/aarch64/arm_neon.h
4791 (vcvtx_high_f32_f64): Fix parameters.
4792
f77e2d2d 47932013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 4794 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 4795
a5bd71fa 4796 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 4797
36875e8f
RB		 47982013-09-16 Richard Biener <rguenther@suse.de>
4799
4800 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4801 (dot_rdg_1): Handle control_dd.
4802 (create_edge_for_control_dependence): New function.
4803 (create_rdg_edges): Add control dependences if asked for.
4804 (build_rdg): Likewise.
4805 (generate_loops_for_partition): If there are not necessary
4806 control stmts remove all their dependencies.
4807 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4808 (distribute_loop): Pass on control dependences.
4809 (tree_loop_distribution): Compute control dependences and remove
4810 restriction on number of loop nodes.
4811
96e24d49
JJ		 48122013-09-16 Jakub Jelinek <jakub@redhat.com>
4813
4814 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4815 for internal calls.
4816
40dbb05c
RS		 48172013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4818
4819 * cse.c (try_const_anchors): Punt on CC modes.
4820
97191ad0
JBG		 48212013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4822
4823 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4824
403a3fb7
JDA		 48252013-09-14 John David Anglin <danglin@gcc.gnu.org>
4826
4827 PR target/58382
4828 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4829 calls to word_mode.
4830
fb5610fb
IS		 48312013-09-14 Iain Sandoe <iain@codesourcery.com>
4832
4833 PR target/48094
4834 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG		 4835 seen.
4836 (darwin_objc1_section): Likewise.
fb5610fb
IS		 4837 (darwin_file_end): Emit Image Info section when required.
4838
afdec9bd
JH		 48392013-09-14 Jan Hubicka <jh@suse.cz>
4840
4841 * tree-into-ssa.c (gate_into_ssa): New.
4842 (pass_data_build_ssa): Use it.
4843 * cgraph.h (expand_thunk): Update prototype.
4844 * cgraphunit.c (analyze_function): Expand thunks early.
4845 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4846 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4847 set lowered flag; do not add new function.
4848 (assemble_thunks_and_aliases): Update.
4849 * tree-ssa.c (gate_init_datastructures): New gate.
4850 (pass_data_init_datastructures): Use it.
4851
492d1e0a
IS		 48522013-09-14 Iain Sandoe <iain@codesourcery.com>
4853
4854 PR target/58269
f77e2d2d 4855 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS		 4856 xmm register set described in the psABI.
4857
d13c6b70
EG		 48582013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4859
f77e2d2d
JBG		 4860 * dwarf2out.c (should_emit_struct_debug): Add check
4861 for type_decl variable is not NULL.
d13c6b70 4862
88b0e79e
JC		 48632013-09-13 Jacek Caban <jacek@codeweavers.com>
4864
4865 * config.gcc: Use new winnt-c.c target hooks
4866 * config/t-winnt: New file
4867 * config/winnt-c.c: New file
4868 * doc/tm.texi.in: Document new hook
4869 * doc/tm.texi: Regenerated
4870
19ba6aab
JH		 48712013-09-13 Jan Hubicka <jh@suse.cz>
4872
4873 PR middle-end/58094
4874 * ipa-inline.c (check_callers): New function.
4875 (check_caller_edge): Remove.
4876 (want_inline_function_to_all_callers_p): Also permit alises that are
4877 called dirrectly.
4878 (inline_to_all_callers): Terminate the walk when devirtualization
4879 introduce new calls.
4880
a93c18c8
JH		 48812013-09-13 Jan Hubicka <jh@suse.cz>
4882
4883 * ipa-inline-analysis.c (struct growth_data): Add node.
4884 (do_estimate_growth_1): Fix detection of recursion.
4885
f3b331d1
JJ		 48862013-09-13 Jakub Jelinek <jakub@redhat.com>
4887
4888 PR tree-optimization/58392
4889 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4890 to avoid shadowing of outer loop variable. If
4891 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4892 replace_by_duplicate_decl simduid of loops that have it set and
4893 set dest_cfun->has_simduid_loops and/or
4894 dest_cfun->has_force_vect_loops.
4895 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4896 instead of maybe_lookup_decl.
4897 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4898 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4899 call. Copy over force_vect and copy and remap simduid. Set
4900 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4901 (copy_cfg_body): Remove blocks_to_copy argument. Use
4902 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4903 caller. Don't set cfun->has_simduid_loops and/or
4904 cfun->has_force_vect_loops here.
4905 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4906 caller.
4907 (expand_call_inline, tree_function_versioning): Adjust copy_body
4908 callers.
4909
ae6d0907
MJ		 49102013-09-13 Martin Jambor <mjambor@suse.cz>
4911
4912 PR bootstrap/58388
4913 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4914 the assert if the edge was a speculative one.
4915
2fd5894f
RB		 49162013-09-13 Richard Biener <rguenther@suse.de>
4917
4918 * tree-data-ref.h (known_dependences_p): Move here ...
4919 * tree-loop-distribution.c (known_dependences_p): ... from here.
4920 (dump_rdg_component, debug_rdg_component): Remove.
4921 (dump_rdg): Adjust.
4922 (generate_loops_for_partition): Use gimple_uid instead of
4923 relying on matching stmt visit order.
4924 (rdg_build_partitions): Take starting stmt vector.
4925 (ldist_gen): Merge into ...
4926 (distribute_loop): ... this function. Do not compute starting
4927 vertices vector.
4928 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
4929
86eb4bd7
KT		 49302013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4931
4932 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
4933 Set type attribute correctly. Set predicable_short_it attribute.
4934 (cmpsi_shiftsi): Remove %? from output template.
4935
83a95546
RB		 49362013-09-13 Richard Biener <rguenther@suse.de>
4937
4938 * tree-loop-distribution.c (struct rdg_component,
4939 rdg_defs_used_in_other_loops_p, free_rdg_components,
4940 rdg_build_components): Remove.
4941 (stmts_from_loop): Do not record virtual PHIs.
4942 (generate_loops_for_partition): Skip virtual PHIs.
4943 (build_rdg_partition_for_component): Rename to ...
4944 (build_rdg_partition_for_vertex): ... this and adjust.
4945 (rdg_build_partitions): Take a vector of starting vertices
4946 instead of components. Remove unnecessary leftover handling.
4947 (ldist_gen): Do not build components or record other stores.
4948 (distribute_loop): Do not distribute loops containing stmts
4949 with side-effects.
4950
c0ecf7f1
CB		 49512013-09-13 Christian Bruel <christian.bruel@st.com>
4952
4953 PR target/58314
4954 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
4955
34f56efc
KT		 49562013-09-13 Kai Tietz <ktietz@redhat.com>
4957
740f2bf3
KT		 4958 * config.gcc: Separate cases for mingw and cygwin targets,
4959 and add 64-bit cygwin target case.
4960
34f56efc
KT		 4961 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
4962 dll-export inline-functions.
4963 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
4964
4965
49662013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL		 4967
4968 PR middle-end/58387
4969 Revert:
4970 2013-09-06 Jeff Law <law@redhat.com>
4971
4972 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4973 edge implied equivalences into successor phis.
4974
aadb52ad
DD		 49752013-09-12 DJ Delorie <dj@redhat.com>
4976
4977 * config/rl78/rl78-virt.md: Change from | to \; for asm line
4978 separators.
4979
90ae7019
BM		 49802013-09-12 Brooks Moses <bmoses@google.com>
4981
4982 PR driver/42955
4983 * Makefile.in: Do not install driver binaries in $(target)/bin.
4984
17b2557c
DD		 49852013-09-12 DJ Delorie <dj@redhat.com>
4986
c5215a9b
DD		 4987 * config/rl78/rl78.opt (mrelax): New.
4988 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
4989 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
4990
17b2557c
DD		 4991 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
4992 between SP and FP.
4993 (rl78_expand_epilogue): Likewise.
4994
ecafcf05
VM		 49952013-09-12 Vladimir Makarov <vmakarov@redhat.com>
4996
4997 PR middle-end/58335
84f6d9fa 4998 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM		 4999 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5000 pointer elimination with using remove_reg_equal_offset_note.
5001
8369f38a
DD		 50022013-09-12 DJ Delorie <dj@redhat.com>
5003
f6a83b4a
DD		 5004 * config/msp430/: New port.
5005 * config.gcc (msp430): Added.
5006 * doc/invoke.texi: Document MSP430 options.
5007 * doc/install.texi: Document msp430-elf
5008 * doc/md.texi: Document msp430-elf
5009 * doc/contrib.texi: Document msp430-elf
5010
8369f38a
DD		 5011 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5012 PSImode.
5013
a854f856
MJ		 50142013-09-12 Martin Jambor <mjambor@suse.cz>
5015
5016 PR ipa/58389
5017 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5018 reference descriptor is NULL.
5019 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5020 edge to NULL.
5021
b12ebd96
AM		 50222013-09-12 Andrew MacLeod <amacleod@redhat.com>
5023
5024 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5025 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5026 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5027 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5028 make_temp_ssa_name): move to tree-ssanames.h
5029 * tree-ssa-alias.h: Move prototype.
5030 * tree-ssa.h: Include tree-ssanames.h.
5031 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5032 * tree-ssanames.h: New. Move items from tree-flow*.h
5033 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5034
e1fb4ad3
RB		 50352013-09-12 Richard Biener <rguenther@suse.de>
5036
5037 PR tree-optimization/58404
5038 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5039 propagate non-invariant addresses into dereferences wrapped
5040 in component references.
5041
154c9188
RB		 50422013-09-12 Richard Biener <rguenther@suse.de>
5043
5044 PR tree-optimization/58402
5045 * passes.def: Move pass_late_warn_uninitialized later.
5046
7a300452
AM		 50472013-09-12 Andrew MacLeod <amacleod@redhat.com>
5048
5049 * tree-ssa.h: New. Move content from tree-flow.h and
5050 tree-flow-inline.h.
5051 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5052 Move prototypes belonging to tree-ssa.c.
f77e2d2d 5053 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM		 5054 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5055 tree-ssa.h.
5056 * gimple.h: Adjust prototypes.
f77e2d2d
JBG		 5057 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5058 to...
5059 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM		 5060 * tree.h: Move prototype to tree-ssa.h.
5061 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5062 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5063 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5064 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5065 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5066 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5067 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5068 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG		 5069 graphite-blocking.c, graphite-clast-to-gimple.c,
5070 graphite-dependences.c, graphite-interchange.c,
5071 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5072 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5073 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5074 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM		 5075 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5076 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5077 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5078 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5079 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5080 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5081 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5082 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5083 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5084 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5085 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5086 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5087 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5088 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5089 value-prof.c, var-tracking.c,
5090 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5091
97463b2b
RB		 50922013-09-12 Richard Biener <rguenther@suse.de>
5093
5094 PR tree-optimization/58396
5095 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5096 (build_rdg): Take a loop-nest parameter, fix memleaks.
5097 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5098
80ecd89d
YR		 50992013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5100
5101 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5102 for SLM.
5103
893e85fa
CM		 51042013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5105
5106 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5107 The FMA instruction names should have a 'v' prefix.
5108
174ec470
RB		 51092013-09-12 Richard Biener <rguenther@suse.de>
5110
5111 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5112 (dot_rdg): Use popen instead of system in optional code.
5113 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5114 (already_processed_vertex_p): Adjust.
5115 (has_anti_or_output_dependence, predecessor_has_mem_write,
5116 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5117 rdg_flag_uses): Remove.
5118 (rdg_flag_vertex): Simplify.
5119 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5120 remove recursion.
5121 (build_rdg_partition_for_component): Process the first vertex
5122 of a component only.
5123 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5124
38c89763
AM		 51252013-09-12 Alan Modra <amodra@gmail.com>
5126
5127 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5128
3e3d9d17
DD		 51292013-09-11 DJ Delorie <dj@redhat.com>
5130 Nick Clifton <nickc@redhat.com>
5131
5132 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5133 (rl78_stack_based_mem): New.
5134 * config/rl78/constraints.md (Iv08): New.
5135 (Iv16): New.
5136 (Iv24): New.
5137 (Is09): New.
5138 (Is17): New.
5139 (Is25): New.
5140 (ISsi): New.
5141 (IShi): New.
5142 (ISqi): New.
5143 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5144 (movhi): Likewise.
5145 (movsi): Change from expand to insn-and-split.
5146 (ashrsi3): Clobber AX.
5147 (lshrsi3): New.
5148 (ashlsi3): New.
5149 (cbranchsi4): New.
5150 * config/rl78/rl78.md (CC_REG): Fix.
5151 (addsi3): Allow memory and immediate operands.
5152 (addsi3_internal): Split into...
5153 (addsi3_internal_virt): ...new, and ...
5154 (addsi3_internal_real): ...new.
5155 (subsi): New.
5156 (subsi3_internal_virt): New.
5157 (subsi3_internal_real): New.
5158 (mulsi3): Add memory operand.
5159 (mulsi3_rl78): Likewise.
5160 (mulsi3_g13): Likewise.
5161 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5162 (cbranchqi4_real): Add more constraint options.
5163 (cbranchhi4_real): Expand pattern.
5164 (cbranchhi4_real_signed): New.
5165 (cbranchhi4_real_inverted): New.
5166 (cbranchsi4_real_lt): New.
5167 (cbranchsi4_real_ge): New.
5168 (cbranchsi4_real_signed): New.
5169 (cbranchsi4_real): New.
5170 (peephole2): New.
893e85fa
CM		 5171 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5172 constant shifts.
3e3d9d17
DD		 5173 (lshrsi3_virt): Likewise.
5174 (ashlsi3_virt): Likewise.
5175 (cbranchqi4_virt_signed): New.
5176 (cbranchhi4_virt_signed): New.
5177 (cbranchsi4_virt): New.
5178 * config/rl78/rl78.c: Whitespace fixes throughout.
5179 (move_elim_pass): New.
5180 (pass_data_rl78_move_elim): New.
5181 (pass_rl78_move_elim): New.
5182 (make_pass_rl78_move_elim): New.
5183 (rl78_devirt_info): Run devirt earlier.
5184 (rl78_move_elim_info): New.
5185 (rl78_asm_file_start): Register it.
5186 (rl78_split_movsi): New.
5187 (rl78_as_legitimate_address): Allow virtual base registers when
5188 appropriate.
5189 (rl78_addr_space_convert): Remove spurious debug stuff.
5190 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5191 (rl78_print_operand): More cases for not printing '#'.
5192 (rl78_expand_compare): Remove most of the logic.
5193 (content_memory): New.
5194 (clear_content_memory): New.
5195 (get_content_index): New.
5196 (get_content_name): New.
5197 (display_content_memory): New.
5198 (update_content): New.
5199 (record_content): New.
5200 (already_contains): New.
5201 (insn_ok_now): Re-recog insns with virtual registers.
5202 (add_postponed_content_update): New.
5203 (process_postponed_content_update): New.
5204 (gen_and_emit_move): New.
893e85fa
CM		 5205 (transcode_memory_rtx): Record new location content.
5206 Use gen_and_emit_move.
3e3d9d17
DD		 5207 (force_into_acc): New.
5208 (move_to_acc): Use gen_and_emit_move.
5209 (move_from_acc): Likewise.
5210 (move_acc_to_reg): Likewise.
5211 (move_to_x): Likewise.
5212 (move_to_hl): Likewise.
5213 (move_to_de): Likewise.
5214 (rl78_alloc_physical_registers_op1): Record location content.
5215 (has_constraint): New.
5216 (rl78_alloc_physical_registers_op2): Record location content.
5217 Optimize use of HL.
5218 (rl78_alloc_physical_registers_ro1): Likewise.
5219 (rl78_alloc_physical_registers_cmp): Likewise.
5220 (rl78_alloc_physical_registers_umul): Likewise.
5221 (rl78_alloc_address_registers_macax): New.
5222 (rl78_alloc_physical_registers): Initialize and set location
5223 content memory as needed.
5224 (rl78_reorg): Make sure split2 is called.
5225 (rl78_rtx_costs): New.
5226
088845a5
RS		 52272013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5228
5229 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5230 for (not (neg ...)) and (neg (not ...)) cases.
5231
5c2961cf
RB		 52322013-09-11 Richard Biener <rguenther@suse.de>
5233
5234 PR middle-end/58377
5235 * passes.def: Split critical edges before late uninit warning passes.
5236 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5237
427e6a14
JJ		 52382013-09-11 Jakub Jelinek <jakub@redhat.com>
5239
5240 PR tree-optimization/58385
5241 * fold-const.c (build_range_check): If both low and high are NULL,
5242 use omit_one_operand_loc to preserve exp side-effects.
5243
df45c0cc
KT		 52442013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5245
5246 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5247
80ab0b19
RB		 52482013-09-11 Richard Biener <rguenther@suse.de>
5249
5250 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5251 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5252 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5253 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5254 stmts_from_loop, known_dependences_p, build_empty_rdg,
5255 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5256 * tree-loop-distribution.c: ... here.
5257 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5258 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5259 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5260 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5261 * tree-loop-distribution.c: ... here.
5262 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5263 (struct partition_s): Add loops member.
5264 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5265 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5266 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5267
85a77221
AI		 52682013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5269 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5270 Sergey Lega <sergey.s.lega@intel.com>
5271 Anna Tikhonova <anna.tikhonova@intel.com>
5272 Ilya Tocar <ilya.tocar@intel.com>
5273 Andrey Turetskiy <andrey.turetskiy@intel.com>
5274 Ilya Verbin <ilya.verbin@intel.com>
5275 Kirill Yukhin <kirill.yukhin@intel.com>
5276 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5277
5278 * config/i386/constraints.md (k): New.
5279 (Yk): Ditto.
5280 * config/i386/i386.c (const regclass_map): Add new mask registers.
5281 (dbx_register_map): Ditto.
5282 (dbx64_register_map): Ditto.
5283 (svr4_dbx_register_map): Ditto.
5284 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5285 disabled.
5286 (ix86_preferred_reload_class): Disable constants for mask registers.
5287 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5288 (ix86_hard_regno_mode_ok): Support new mask registers.
5289 (x86_order_regs_for_local_alloc): Ditto.
5290 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5291 (FIXED_REGISTERS): Add new mask registers.
5292 (CALL_USED_REGISTERS): Ditto.
5293 (REG_ALLOC_ORDER): Ditto.
5294 (VALID_MASK_REG_MODE): New.
5295 (FIRST_MASK_REG): Ditto.
5296 (LAST_MASK_REG): Ditto.
5297 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5298 (MAYBE_MASK_CLASS_P): New.
5299 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5300 (REG_CLASS_CONTENTS): Ditto.
5301 (MASK_REGNO_P): New.
5302 (ANY_MASK_REG_P): Ditto.
5303 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM		 5304 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5305 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5306 mask registers.
85a77221
AI		 5307 (attribute "type"): Add mskmov, msklog.
5308 (attribute "length_immediate"): Support them.
5309 (attribute "memory"): Ditto.
5310 (attribute "prefix_0f"): Ditto.
5311 (*movhi_internal): Support new mask registers.
5312 (*movqi_internal): Ditto.
5313 (define_split): Split out clobber pattern is a logic
5314 insn on mask registers.
5315 (*k<logic><mode>): New.
5316 (*andhi_1): Extend to support mask regs.
5317 (*andqi_1): Extend to support mask regs.
5318 (kandn<mode>): New.
5319 (define_split): Split and-not to and and not if operands
5320 are not mask regs.
5321 (*<code><mode>_1): Separate HI mode to new pattern...
5322 (*<code>hi_1): This.
5323 (*<code>qi_1): Extend to support mask regs.
5324 (kxnor<mode>): New.
5325 (kortestzhi): Ditto.
5326 (kortestchi): Ditto.
5327 (kunpckhi): Ditto.
5328 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5329 (*one_cmplhi2_1): ...Here, now with mask registers support.
5330 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM		 5331 (HI/QImode arithmetics splitter): Don't split if mask registers
5332 are used.
85a77221
AI		 5333 (HI/QImode not splitter): Ditto.
5334 * config/i386/predicated.md (mask_reg_operand): New.
5335 (general_reg_operand): Ditto.
5336
7e570821
AI		 53372013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5338
5339 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5340 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5341
9e1376e9
JL		 53422013-09-10 Jeff Law <law@redhat.com>
5343
5344 PR tree-optimization/58380
5345 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5346 of threading through a buried loop header.
5347
5348 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5349 return value for single successor case.
5350
5bbcb888
JH		 53512013-09-10 Jan Hubicka <jh@suse.cz>
5352
5353 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5354
d742ff4b
RE		 53552013-09-10 Richard Earnshaw <rearnsha@arm.com>
5356
5357 PR target/58361
5358 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5359 support conditional execution.
5360 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5361
b0681c9e
VM		 53622013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5363
5364 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5365 constraint pass.
f77e2d2d
JBG		 5366 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5367 Check destination too to check move insn.
b0681c9e 5368 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG		 5369 changed its allocation and the optional reload was inherited on last
5370 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM		 5371 (lra_undo_inheritance): Add check that inherited pseudo still in
5372 memory.
5373
c9b1c957
JG		 53742013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5375
5376 * config/aarch64/aarch64.md (generic_sched): New.
5377 * config/aarch64/aarch64-generic.md (load): Make conditional
5378 on generic_sched attribute.
5379 (nonload): Likewise.
5380
82d618d3
JH		 53812013-09-10 Jan Hubicka <jh@suse.cz>
5382
5383 * lto-cgraph.c: Include ipa-utils.h.
5384 (compute_ltrans_boundary): Also add possible targets into the boundary.
5385
668b6894
JH		 53862013-09-10 Jan Hubicka <jh@suse.cz>
5387
5388 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5389 VAR_DECL of vtable rather than full expression.
5390
4f90d3e0
JH		 53912013-09-10 Jan Hubicka <jh@suse.cz>
5392 Paolo Carlini <paolo.carlini@oracle.com>
5393
5394 * cgraphunit.c (analyze_functions): Save input_location, set it
5395 to UNKNOWN_LOCATION and restore it at the end.
5396
39e87baf
MJ		 53972013-09-10 Martin Jambor <mjambor@suse.cz>
5398
5399 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5400 represented by a thunk.
5401
581aedec
JL		 54022013-09-10 Jeff Law <law@redhat.com>
5403
5404 PR tree-optimization/58343
5405 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5406 jump threading requests through loop headers buried in the middle
5407 of a jump threading path.
5408
5409 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5410 in return value/type.
5411
84cf4ab6
JJ		 54122013-09-10 Jakub Jelinek <jakub@redhat.com>
5413
5414 PR rtl-optimization/58365
5415 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5416 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5417 it differs.
5418
447dd906
RB		 54192013-09-10 Richard Biener <rguenther@suse.de>
5420
5421 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5422 * tree-data-ref.c (create_rdg_vertices): Collect all data
5423 references, signal failure to the caller, use data-ref API.
893e85fa
CM		 5424 (build_rdg): Compute data references only once. Maintain lifetime
5425 of data references and data dependences from within RDG.
447dd906
RB		 5426 (free_rdg): Free dependence relations.
5427 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5428 inventing extra dependences.
5429 (distribute_loop): Update for RDG API changes.
5430
f14a7324
KT		 54312013-09-10 Kai Tietz <ktietz@redhat.com>
5432
5433 * doc/invoke.texi (fms-extensions): Document changed
5434 behavior for ms-abi targets.
5435 * config/i386/i386.c (ix86_option_override_internal):
5436 Set default value of option -fms-extension for ms-abi targets.
5437
ff131f26
MZ		 54382013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5439
5440 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5441
33115a2c
AM		 54422013-09-10 Alan Modra <amodra@gmail.com>
5443
5444 PR target/58330
5445 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5446
b6872a13
AM		 54472013-09-10 Alan Modra <amodra@gmail.com>
5448
5449 * config/rs6000/predicates.md (add_cint_operand): New.
5450 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5451 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5452 using add_cint_operand.
5453 (largetoc_high_plus_aix): Likewise.
5454
28fd0ba2
JJ		 54552013-09-09 Jakub Jelinek <jakub@redhat.com>
5456
5457 PR tree-optimization/58364
5458 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5459 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5460 the current range can't be an unconditional true or false.
5461
552d99fe
JG		 54622013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5463
5464 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5465
f43245d1
UB		 54662013-09-09 Uros Bizjak <ubizjak@gmail.com>
5467
5468 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5469
c9b0866a
PC		 54702013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5471
5472 PR c++/43452
5473 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5474
78d8b9f0
IB		 54752013-09-09 Ian Bolton <ian.bolton@arm.com>
5476
5477 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5478 NO_REGS for immediate that can't be moved directly into FP_REGS.
5479
274b2532
KT		 54802013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5481
5482 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5483 comparison with negated operand.
c9b0866a
PC		 5484 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5485 RTL form.
274b2532 5486
467a3558
RB		 54872013-09-09 Richard Biener <rguenther@suse.de>
5488
5489 PR middle-end/58326
5490 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5491 of a subloop record all its block as affecting loop-closed SSA form.
5492
088c5368
RS		 54932013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5494
5495 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5496 of an rtx/bitpos pair.
5497 (store_fixed_bit_field): Update accordingly.
5498
69db2d57
RS		 54992013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5500
5501 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5502 GEN_INT.
5503 * builtins.c (expand_errno_check): Likewise.
5504 * dwarf2cfi.c (init_return_column_size): Likewise.
5505 * except.c (sjlj_mark_call_sites): Likewise.
5506 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5507 * lra-constraints.c (emit_inc): Likewise.
5508 * ree.c (combine_set_extension): Likewise.
5509 * regmove.c (fixup_match_2): Likewise.
5510 * reload1.c (inc_for_reload): Likewise.
5511
69a59f0f
RS		 55122013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5513
5514 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5515 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5516 Use gen_int_mode with the mode of the associated simplify_* call.
5517 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5518 * expmed.c (expand_shift_1): Likewise.
5519 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5520 * loop-iv.c (iv_number_of_iterations): Likewise.
5521 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5522 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5523
2f1cd2eb
RS		 55242013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5525
5526 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5527 of the associated expand_* call.
5528 (asan_emit_stack_protection): Likewise.
5529 * builtins.c (round_trampoline_addr): Likewise.
5530 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5531 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5532 (emit_store_flag): Likewise.
5533 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5534 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5535 Likewise.
5536 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5537 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5538 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5539 * modulo-sched.c (generate_prolog_epilog): Likewise.
5540 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5541 (expand_ctz, expand_ffs, expand_unop): Likewise.
5542
4789c0ce
RS		 55432013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5544
5545 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5546 of the associated gen_rtx_* call.
5547 * caller-save.c (init_caller_save): Likewise.
5548 * combine.c (find_split_point, make_extraction): Likewise.
5549 (make_compound_operation): Likewise.
5550 * dwarf2out.c (mem_loc_descriptor): Likewise.
5551 * explow.c (plus_constant, probe_stack_range): Likewise.
5552 * expmed.c (expand_mult_const): Likewise.
5553 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5554 * reload1.c (init_reload): Likewise.
5555 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5556 * var-tracking.c (adjust_mems): Likewise.
5557 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5558 rather than gen_rtx_fmt_ee.
5559
a811f7cb
JH		 55602013-09-09 Jan Hubicka <jh@suse.cz>
5561
5562 PR middle-end/58294
5563 * value-prof.c (gimple_ic): Copy also abnormal edges.
5564
dcad1dd3
RS		 55652013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5566
5567 * asan.c (asan_shadow_cst): Use gen_int_mode.
5568
daf5c770
JH		 55692013-09-08 Jan Hubicka <jh@suse.cz>
5570
5571 * ipa-profile.c: Add toplevel comment.
5572 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5573 (contains_hot_call_p): New function.
5574 (ipa_propagate_frequency): Set frequencies based on counts when
5575 profile is read.
5576 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5577 profile; do not tamper with profile after inlining if it is read.
5578
a12cd2db
JH		 55792013-09-08 Jan Hubicka <jh@suse.cz>
5580
5581 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5582 speculative edges.
5583
856a1588
JH		 55842013-09-08 Jan Hubicka <jh@suse.cz>
5585
5586 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5587 summary generation.
5588
4f4b0b73
JL		 55892013-09-08 Jeff Law <law@redhat.com>
5590
5591 PR bootstrap/58340
5592 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5593 of 'found'.
5594
cc4029ee
AK		 55952013-09-08 Andi Kleen <ak@linux.intel.com>
5596
5597 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5598
7b395ddd
JH		 55992013-09-08 Jan Hubicka <jh@suse.cz>
5600
5601 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5602
c34c46dd
RS		 56032013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5604
5605 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5606 for non-debug insns.
5607 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5608 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5609 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5610 accordingly.
5611
3462aa02
JH		 56122013-09-08 Jan Hubicka <jh@suse.cz>
5613
5614 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5615 targets and devirtualize to BUILT_IN_UNREACHABLE.
5616 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5617 * ipa.c (walk_polymorphic_call_targets): New function.
5618 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5619 functions; use the new timevar.
5620 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5621 was removed from the program.
5622 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5623 not consider it in the walk when its vtable is dead.
5624 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5625 record_binfo.
5626 (devirt_variable_node_removal_hook): New function.
5627 (possible_polymorphic_call_targets): Also register
5628 devirt_variable_node_removal_hook.
5629 (ipa_devirt): Do not do non-speculative devirtualization.
5630 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5631
26e5b0fd
JH		 56322013-09-08 Jan Hubicka <jh@suse.cz>
5633
5634 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5635 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5636 varpool_remove_variable_insertion_hook): Declare.
5637 * varpool.c (varpool_node_hook_list): New structure.
5638 (first_varpool_node_removal_hook,
5639 first_varpool_variable_insertion_hook): New variables.
5640 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5641 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5642 varpool_remove_variable_insertion_hook,
5643 varpool_call_variable_insertion_hooks): New functions.
5644 (varpool_remove_node): Use it.
5645
97f1f314 56462013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 5647
97f1f314
PC		 5648 PR c++/54941
5649 * diagnostic.c (diagnostic_build_prefix): When s.file is
5650 "<built-in>" don't output line and column numbers.
6df2ca09 5651
bcb650cb
JH		 56522013-09-06 Jan Hubicka <jh@suse.cz>
5653
5654 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5655 * lto-streamer-out.c: Stream thunks, too.
5656 * lto-streamer-in.c (input_function): Pop cfun here
5657 (lto_read_body): Instead of here.
5658
87f3fea8
CT		 56592013-09-06 Caroline Tice <cmtice@google.com>
5660
5661 * doc/install.texi: Add documentation for the --enable-vtable-verify
5662 and the --disable-libvtv configure options.
5663
5e8d9b45
JL		 56642013-09-06 Jeff Law <law@redhat.com>
5665
5666 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5667 edge implied equivalences into successor phis.
5668
3a001aff
JR		 56692013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5670
5671 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5672
56732013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 5674
5675 * resource.c (mark_target_live_regs): Compute resources taking
5676 into account if a call is predicated or not.
5677
2c641f8b
EB		 56782013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5679
5680 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5681 the compiler in the function names.
5682
a81b0a3d
JH		 56832013-09-06 Jan Hubicka <jh@suse.cz>
5684
5685 PR middle-end/58094
5686 * ipa-inline.c (has_caller_p): New function.
5687 (want_inline_function_to_all_callers_p): Use it.
5688 (sum_callers, inline_to_all_callers): Break out from ...
5689 (ipa_inline): ... here.
5690
68e4ca7e
JH		 56912013-09-06 Jan Hubicka <jh@suse.cz>
5692
2c641f8b
EB		 5693 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5694 only when AVX is enabled.
68e4ca7e 5695
3c73e2fe
JG		 56962013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5697
5698 * config/aarch64/aarch64.md
5699 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5700 is fpsimd_<load/store>2.
5701 (load_pair<mode>): Likewise.
5702 (store_pair<mode>): Likewise.
5703
8c48eecd
JG		 57042013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5705
5706 * config/arm/types.md (type): Add "mrs" type.
5707 * config/aarch64/aarch64.md
5708 (aarch64_load_tp_hard): Make type "mrs".
5709 * config/arm/arm.md
5710 (load_tp_hard): Make type "mrs".
5711 * config/arm/cortex-a15.md: Update with new attributes.
5712 * config/arm/cortex-a5.md: Update with new attributes.
5713 * config/arm/cortex-a53.md: Update with new attributes.
5714 * config/arm/cortex-a7.md: Update with new attributes.
5715 * config/arm/cortex-a8.md: Update with new attributes.
5716 * config/arm/cortex-a9.md: Update with new attributes.
5717 * config/arm/cortex-m4.md: Update with new attributes.
5718 * config/arm/cortex-r4.md: Update with new attributes.
5719 * config/arm/fa526.md: Update with new attributes.
5720 * config/arm/fa606te.md: Update with new attributes.
5721 * config/arm/fa626te.md: Update with new attributes.
5722 * config/arm/fa726te.md: Update with new attributes.
5723
9ea187a5
JG		 57242013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5725
5726 * config/aarch64/aarch64.md
5727 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5728 (*movtf_aarch64): Likewise.
5729 * config/arm/arm.md
5730 (thumb1_movdi_insn): Use "multiple" for type where more than one
5731 instruction is used for a move.
5732 (*arm32_movhf): Likewise.
5733 (*thumb_movdf_insn): Likewise.
5734
292b89b3
JG		 57352013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5736
5737 * config/arm/types.md (type): Rename fcpys to fmov.
5738 * config/arm/vfp.md
5739 (*arm_movsi_vfp): Rename type fcpys as fmov.
5740 (*thumb2_movsi_vfp): Likewise
5741 (*movhf_vfp_neon): Likewise
5742 (*movhf_vfp): Likewise
5743 (*movsf_vfp): Likewise
5744 (*thumb2_movsf_vfp): Likewise
5745 (*movsfcc_vfp): Likewise
5746 (*thumb2_movsfcc_vfp): Likewise
5747 * config/aarch64/aarch64-simd.md
5748 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5749 * config/aarch64/aarch64.md
5750 (*movsi_aarch64): Replace type mov_reg with fmovs.
5751 (*movdi_aarch64): Likewise
5752 (*movsf_aarch64): Likewise
5753 (*movdf_aarch64): Likewise
5754 * config/arm/arm.c
5755 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5756 * config/arm/iwmmxt.md
5757 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5758 * config/arm/arm1020e.md: Update with new attributes.
5759 * config/arm/cortex-a15-neon.md: Update with new attributes.
5760 * config/arm/cortex-a5.md: Update with new attributes.
5761 * config/arm/cortex-a53.md: Update with new attributes.
5762 * config/arm/cortex-a7.md: Update with new attributes.
5763 * config/arm/cortex-a8-neon.md: Update with new attributes.
5764 * config/arm/cortex-a9.md: Update with new attributes.
5765 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5766 * config/arm/cortex-r4f.md: Update with new attributes.
5767 * config/arm/marvell-pj4.md: Update with new attributes.
5768 * config/arm/vfp11.md: Update with new attributes.
5769
6a24a83e
JG		 57702013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5771
5772 * config/aarch64/aarch64.md
5773 (*madd<mode>): Fix type attribute.
5774 (*maddsi_uxtw): Likewise.
5775 (*msub<mode>): Likewise.
5776 (*msubsi_uxtw): Likewise.
5777 (<su_optab>maddsidi4): Likewise.
5778 (<su_optab>msubsidi4): Likewise.
5779
b86923f0
JG		 57802013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5781
5782 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5783 * config/arm/arm.md (core_cycles): Remove fdiv.
5784 * config/arm/vfp.md:
5785 (*sqrtsf2_vfp): Update for attribute changes.
5786 (*sqrtdf2_vfp): Likewise.
5787 * config/aarch64/aarch64.md:
5788 (sqrt<mode>2): Update for attribute changes.
5789 * config/arm/arm1020e.md: Update with new attributes.
5790 * config/arm/cortex-a15-neon.md: Update with new attributes.
5791 * config/arm/cortex-a5.md: Update with new attributes.
5792 * config/arm/cortex-a53.md: Update with new attributes.
5793 * config/arm/cortex-a7.md: Update with new attributes.
5794 * config/arm/cortex-a8-neon.md: Update with new attributes.
5795 * config/arm/cortex-a9.md: Update with new attributes.
5796 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5797 * config/arm/cortex-r4f.md: Update with new attributes.
5798 * config/arm/marvell-pj4.md: Update with new attributes.
5799 * config/arm/vfp11.md: Update with new attributes.
5800
7b49c9e1
JG		 58012013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5802
5803 * config/arm/types.md
5804 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5805 * config/aarch64/aarch64.md
5806 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5807 new attributes.
5808 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5809 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5810 (float<GPI:mode><GPF:mode>2): Likewise.
5811 * config/arm/vfp.md
5812 (*truncsisf2_vfp): Update with new attributes.
5813 (*truncsidf2_vfp): Likewise.
5814 (fixuns_truncsfsi2): Likewise.
5815 (fixuns_truncdfsi2): Likewise.
5816 (*floatsisf2_vfp): Likewise.
5817 (*floatsidf2_vfp): Likewise.
5818 (floatunssisf2): Likewise.
5819 (floatunssidf2): Likewise.
5820 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5821 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5822 * config/arm/arm1020e.md: Update with new attributes.
5823 * config/arm/cortex-a15-neon.md: Update with new attributes.
5824 * config/arm/cortex-a5.md: Update with new attributes.
5825 * config/arm/cortex-a53.md: Update with new attributes.
5826 * config/arm/cortex-a7.md: Update with new attributes.
5827 * config/arm/cortex-a8-neon.md: Update with new attributes.
5828 * config/arm/cortex-a9.md: Update with new attributes.
5829 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5830 * config/arm/cortex-r4f.md: Update with new attributes.
5831 * config/arm/marvell-pj4.md: Update with new attributes.
5832 * config/arm/vfp11.md: Update with new attributes.
5833
e7df8af8
JG		 58342013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5835
5836 * config/aarch64/arm_neon.h
5837 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5838 (vqtbx<1,2,3,4><q>_s8): Likewise.
5839
594726e4
JG		 58402013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5841
f43245d1
UB		 5842 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5843 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG		 5844 (add<mode>3): Add type attribute.
5845 (add<mode>3): Likewise.
5846 (usadd<mode>3): Likewise.
5847 (ssadd<mode>3): Likewise.
5848 (sub<mode>3): Likewise.
5849 (sub<mode>3): Likewise.
5850 (ussub<mode>3): Likewise.
5851 (sssub<mode>3): Likewise.
5852 (ssmulsa3): Likewise.
5853 (usmulusa3): Likewise.
5854 (arm_usatsihi): Likewise.
5855 * config/arm/vfp.md
5856 (*movdi_vfp): Add types for all instructions.
5857 (*movdi_vfp_cortexa8): Likewise.
5858 (*movhf_vfp_neon): Likewise.
5859 (*movhf_vfp): Likewise.
5860 (*movdf_vfp): Likewise.
5861 (*thumb2_movdf_vfp): Likewise.
5862 (*thumb2_movdfcc_vfp): Likewise.
5863 * config/arm/arm.md: Add type attribute to all insn patterns.
5864 (*thumb1_adddi3): Add type attribute.
5865 (*arm_adddi3): Likewise.
5866 (*adddi_sesidi_di): Likewise.
5867 (*adddi_zesidi_di): Likewise.
5868 (*thumb1_addsi3): Likewise.
5869 (addsi3_compare0): Likewise.
5870 (*addsi3_compare0_scratch): Likewise.
5871 (*compare_negsi_si): Likewise.
5872 (cmpsi2_addneg): Likewise.
5873 (*addsi3_carryin_<optab>): Likewise.
5874 (*addsi3_carryin_alt2_<optab>): Likewise.
5875 (*addsi3_carryin_clobercc_<optab>): Likewise.
5876 (*subsi3_carryin): Likewise.
5877 (*subsi3_carryin_const): Likewise.
5878 (*subsi3_carryin_compare): Likewise.
5879 (*subsi3_carryin_compare_const): Likewise.
5880 (*arm_subdi3): Likewise.
5881 (*thumb_subdi3): Likewise.
5882 (*subdi_di_zesidi): Likewise.
5883 (*subdi_di_sesidi): Likewise.
5884 (*subdi_zesidi_di): Likewise.
5885 (*subdi_sesidi_di): Likewise.
5886 (*subdi_zesidi_ze): Likewise.
5887 (thumb1_subsi3_insn): Likewise.
5888 (*arm_subsi3_insn): Likewise.
5889 (*anddi3_insn): Likewise.
5890 (*anddi_zesidi_di): Likewise.
5891 (*anddi_sesdi_di): Likewise.
5892 (*ne_zeroextracts): Likewise.
5893 (*ne_zeroextracts): Likewise.
5894 (*ite_ne_zeroextr): Likewise.
5895 (*ite_ne_zeroextr): Likewise.
5896 (*anddi_notdi_di): Likewise.
5897 (*anddi_notzesidi): Likewise.
5898 (*anddi_notsesidi): Likewise.
5899 (andsi_notsi_si): Likewise.
5900 (thumb1_bicsi3): Likewise.
5901 (*iordi3_insn): Likewise.
5902 (*iordi_zesidi_di): Likewise.
5903 (*iordi_sesidi_di): Likewise.
5904 (*thumb1_iorsi3_insn): Likewise.
5905 (*xordi3_insn): Likewise.
5906 (*xordi_zesidi_di): Likewise.
5907 (*xordi_sesidi_di): Likewise.
5908 (*arm_xorsi3): Likewise.
5909 (*andsi_iorsi3_no): Likewise.
5910 (*smax_0): Likewise.
5911 (*smax_m1): Likewise.
5912 (*arm_smax_insn): Likewise.
5913 (*smin_0): Likewise.
5914 (*arm_smin_insn): Likewise.
5915 (*arm_umaxsi3): Likewise.
5916 (*arm_uminsi3): Likewise.
5917 (*minmax_arithsi): Likewise.
5918 (*minmax_arithsi_): Likewise.
5919 (*satsi_<SAT:code>): Likewise.
5920 (arm_ashldi3_1bit): Likewise.
5921 (arm_ashrdi3_1bit): Likewise.
5922 (arm_lshrdi3_1bit): Likewise.
5923 (*arm_negdi2): Likewise.
5924 (*thumb1_negdi2): Likewise.
5925 (*arm_negsi2): Likewise.
5926 (*thumb1_negsi2): Likewise.
5927 (*negdi_extendsid): Likewise.
5928 (*negdi_zero_extend): Likewise.
5929 (*arm_abssi2): Likewise.
5930 (*thumb1_abssi2): Likewise.
5931 (*arm_neg_abssi2): Likewise.
5932 (*thumb1_neg_abss): Likewise.
5933 (one_cmpldi2): Likewise.
5934 (extend<mode>di2): Likewise.
5935 (*compareqi_eq0): Likewise.
5936 (*arm_extendhisi2addsi): Likewise.
5937 (*arm_movdi): Likewise.
5938 (*thumb1_movdi_insn): Likewise.
5939 (*arm_movt): Likewise.
5940 (*thumb1_movsi_insn): Likewise.
5941 (pic_add_dot_plus_four): Likewise.
5942 (pic_add_dot_plus_eight): Likewise.
5943 (tls_load_dot_plus_eight): Likewise.
5944 (*thumb1_movhi_insn): Likewise.
5945 (*thumb1_movsf_insn): Likewise.
5946 (*movdf_soft_insn): Likewise.
5947 (*thumb_movdf_insn): Likewise.
5948 (cbranchsi4_insn): Likewise.
5949 (cbranchsi4_scratch): Likewise.
5950 (*negated_cbranchsi4): Likewise.
5951 (*tbit_cbranch): Likewise.
5952 (*tlobits_cbranch): Likewise.
5953 (*tstsi3_cbranch): Likewise.
5954 (*cbranchne_decr1): Likewise.
5955 (*addsi3_cbranch): Likewise.
5956 (*addsi3_cbranch_scratch): Likewise.
5957 (*arm_cmpdi_insn): Likewise.
5958 (*arm_cmpdi_unsig): Likewise.
5959 (*arm_cmpdi_zero): Likewise.
5960 (*thumb_cmpdi_zero): Likewise.
5961 (*deleted_compare): Likewise.
5962 (*mov_scc): Likewise.
5963 (*mov_negscc): Likewise.
5964 (*mov_notscc): Likewise.
5965 (*cstoresi_eq0_thumb1_insn): Likewise.
5966 (cstoresi_nltu_thumb1): Likewise.
5967 (cstoresi_ltu_thu): Likewise.
5968 (thumb1_addsi3_addgeu): Likewise.
5969 (*arm_jump): Likewise.
5970 (*thumb_jump): Likewise.
5971 (*check_arch2): Likewise.
5972 (arm_casesi_internal): Likewise.
5973 (thumb1_casesi_dispatch): Likewise.
5974 (*arm_indirect_jump): Likewise.
5975 (*thumb1_indirect_jump): Likewise.
5976 (nop): Likewise.
5977 (*and_scc): Likewise.
5978 (*ior_scc): Likewise.
5979 (*compare_scc): Likewise.
5980 (*cond_move): Likewise.
5981 (*cond_arith): Likewise.
5982 (*cond_sub): Likewise.
5983 (*cmp_ite0): Likewise.
5984 (*cmp_ite1): Likewise.
5985 (*cmp_and): Likewise.
5986 (*cmp_ior): Likewise.
5987 (*ior_scc_scc): Likewise.
5988 (*ior_scc_scc_cmp): Likewise.
5989 (*and_scc_scc): Likewise.
5990 (*and_scc_scc_cmp): Likewise.
5991 (*and_scc_scc_nod): Likewise.
5992 (*negscc): Likewise.
5993 (movcond_addsi): Likewise.
5994 (movcond): Likewise.
5995 (*ifcompare_plus_move): Likewise.
5996 (*if_plus_move): Likewise.
5997 (*ifcompare_move_plus): Likewise.
5998 (*if_move_plus): Likewise.
5999 (*ifcompare_arith_arith): Likewise.
6000 (*if_arith_arith): Likewise.
6001 (*ifcompare_arith_move): Likewise.
6002 (*if_arith_move): Likewise.
6003 (*ifcompare_move_arith): Likewise.
6004 (*if_move_arith): Likewise.
6005 (*ifcompare_move_not): Likewise.
6006 (*if_move_not): Likewise.
6007 (*ifcompare_not_move): Likewise.
6008 (*if_not_move): Likewise.
6009 (*ifcompare_shift_move): Likewise.
6010 (*if_shift_move): Likewise.
6011 (*ifcompare_move_shift): Likewise.
6012 (*if_move_shift): Likewise.
6013 (*ifcompare_shift_shift): Likewise.
6014 (*ifcompare_not_arith): Likewise.
6015 (*ifcompare_arith_not): Likewise.
6016 (*if_arith_not): Likewise.
6017 (*ifcompare_neg_move): Likewise.
6018 (*if_neg_move): Likewise.
6019 (*ifcompare_move_neg): Likewise.
6020 (*if_move_neg): Likewise.
6021 (prologue_thumb1_interwork): Likewise.
6022 (*cond_move_not): Likewise.
6023 (*sign_extract_onebit): Likewise.
6024 (*not_signextract_onebit): Likewise.
6025 (stack_tie): Likewise.
6026 (align_4): Likewise.
6027 (align_8): Likewise.
6028 (consttable_end): Likewise.
6029 (consttable_1): Likewise.
6030 (consttable_2): Likewise.
6031 (consttable_4): Likewise.
6032 (consttable_8): Likewise.
6033 (consttable_16): Likewise.
6034 (*thumb1_tablejump): Likewise.
6035 (prefetch): Likewise.
6036 (force_register_use): Likewise.
6037 (thumb_eh_return): Likewise.
6038 (load_tp_hard): Likewise.
6039 (load_tp_soft): Likewise.
6040 (tlscall): Likewise.
6041 (*arm_movtas_ze): Likewise.
6042 (*arm_rev): Likewise.
6043 (*arm_revsh): Likewise.
6044 (*arm_rev16): Likewise.
6045 * config/arm/thumb2.md
6046 (*thumb2_smaxsi3): Likewise.
6047 (*thumb2_sminsi3): Likewise.
6048 (*thumb32_umaxsi3): Likewise.
6049 (*thumb2_uminsi3): Likewise.
6050 (*thumb2_negdi2): Likewise.
6051 (*thumb2_abssi2): Likewise.
6052 (*thumb2_neg_abss): Likewise.
6053 (*thumb2_movsi_insn): Likewise.
6054 (tls_load_dot_plus_four): Likewise.
6055 (*thumb2_movhi_insn): Likewise.
6056 (*thumb2_mov_scc): Likewise.
6057 (*thumb2_mov_negs): Likewise.
6058 (*thumb2_mov_negs): Likewise.
6059 (*thumb2_mov_nots): Likewise.
6060 (*thumb2_mov_nots): Likewise.
6061 (*thumb2_movsicc_): Likewise.
6062 (*thumb2_movsfcc_soft_insn): Likewise.
6063 (*thumb2_indirect_jump): Likewise.
6064 (*thumb2_and_scc): Likewise.
6065 (*thumb2_ior_scc): Likewise.
6066 (*thumb2_ior_scc_strict_it): Likewise.
6067 (*thumb2_cond_move): Likewise.
6068 (*thumb2_cond_arith): Likewise.
6069 (*thumb2_cond_ari): Likewise.
6070 (*thumb2_cond_sub): Likewise.
6071 (*thumb2_negscc): Likewise.
6072 (*thumb2_movcond): Likewise.
6073 (thumb2_casesi_internal): Likewise.
6074 (thumb2_casesi_internal_pic): Likewise.
6075 (*thumb2_alusi3_short): Likewise.
6076 (*thumb2_mov<mode>_shortim): Likewise.
6077 (*thumb2_addsi_short): Likewise.
6078 (*thumb2_subsi_short): Likewise.
6079 (thumb2_addsi3_compare0): Likewise.
6080 (*thumb2_cbz): Likewise.
6081 (*thumb2_cbnz): Likewise.
6082 (*thumb2_one_cmplsi2_short): Likewise.
6083 (*thumb2_negsi2_short): Likewise.
6084 (*orsi_notsi_si): Likewise.
6085 * config/arm/arm1020e.md: Update with new attributes.
6086 * config/arm/arm1026ejs.md: Update with new attributes.
6087 * config/arm/arm1136jfs.md: Update with new attributes.
6088 * config/arm/arm926ejs.md: Update with new attributes.
6089 * config/arm/cortex-a15.md: Update with new attributes.
6090 * config/arm/cortex-a5.md: Update with new attributes.
6091 * config/arm/cortex-a53.md: Update with new attributes.
6092 * config/arm/cortex-a7.md: Update with new attributes.
6093 * config/arm/cortex-a8.md: Update with new attributes.
6094 * config/arm/cortex-a9.md: Update with new attributes.
6095 * config/arm/cortex-m4.md: Update with new attributes.
6096 * config/arm/cortex-r4.md: Update with new attributes.
6097 * config/arm/fa526.md: Update with new attributes.
6098 * config/arm/fa606te.md: Update with new attributes.
6099 * config/arm/fa626te.md: Update with new attributes.
6100 * config/arm/fa726te.md: Update with new attributes.
6101
1c83b673
JG		 61022013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6103
6104 * config/aarch64/aarch64-simd.md
6105 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6106 <vwx> iterator to ensure correct register choice.
6107 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6108 (aarch64_sqdmull_n<mode>): Likewise.
6109 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6110 * config/aarch64/arm_neon.h
6111 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6112 (vml<as><q>_n_<su>16): Likewise.
6113 (vml<as>l_high_lane<q>_<su>16): Likewise.
6114 (vml<as>l_high_n_<su>16): Likewise.
6115 (vml<as>l_lane<q>_<su>16): Likewise.
6116 (vml<as>l_n_<su>16): Likewise.
6117 (vmul<q>_lane<q>_<su>16): Likewise.
6118 (vmul<q>_n_<su>16): Likewise.
6119 (vmull_lane<q>_<su>16): Likewise.
6120 (vmull_n_<su>16): Likewise.
6121 (vmull_high_lane<q>_<su>16): Likewise.
6122 (vmull_high_n_<su>16): Likewise.
6123 (vqrdmulh<q>_n_s16): Likewise.
6124
bb1ae543
TB		 61252013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6126
6127 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6128 have the correct lane parameter.
6129
f23c0742
RB		 61302013-09-06 Richard Biener <rguenther@suse.de>
6131
6132 * cfganal.c (control_dependences::~control_dependences):
6133 Properly free all of the vector.
6134
c1d49770
KY		 61352013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6136
6137 PR target/58269
6138 * config/i386/i386.c (ix86_conditional_register_usage):
6139 Proper initialize extended SSE registers.
6140
6da47f52
JH		 61412013-09-06 Jan Hubicka <jh@suse.cz>
6142
6143 PR tree-optimization/58311
6144 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6145
61462013-09-06 Jan Hubicka <jh@suse.cz>
6147
6148 * Makefile.in (tree-sra.o): Update dependencies.
6149 * tree-sra.c: Include ipa-utils.h
6150 (scan_function): Use recursive_call_p.
6151 (has_caller_p): New function.
6152 (cgraph_for_node_and_aliases): Count also callers of aliases.
6153
fc11f321
JH		 61542013-09-06 Jan Hubicka <jh@suse.cz>
6155
6156 PR middle-end/58094
6157 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6158 * tree-tailcall.c: Include ipa-utils.h.
6159 (find_tail_calls): Use it.
6160 * ipa-pure-const.c (check_call): Likewise.
6161 * ipa-utils.c (recursive_call_p): New function.
6162 * ipa-utils.h (recursive_call_p): Dclare.
6163 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6164 (symtab_semantically_equivalent_p): New function.
6165 * Makefile.in (tree-tailcall.o): Update dependencies.
6166
9a6af450
EB		 61672013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6168
6169 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6170 non-inlinable part.
6171
35f32ad4
RB		 61722013-09-06 Richard Biener <rguenther@suse.de>
6173
6174 * lto-streamer.h (lto_global_var_decls): Remove.
6175 * Makefile.in (OBJS): Remove lto-symtab.o.
6176 (lto-symtab.o): Remove.
6177 (GTFILES): Remove lto-symtab.c
6178 * lto-symtab.c: Move to lto/
6179
d12a76f3
AK		 61802013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6181
6182 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6183 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6184 (UNSPEC_FPINT_RINT): New constant definitions.
6185 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6186 definition with 2 attributes.
6187 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6188 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6189 definitions.
6190
61912013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6192
f43245d1 6193 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK		 6194 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6195 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6196 attribute to "z196_alone".
6197 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6198 "zEC12_simple".
6199
c8e9d8c3
RB		 62002013-09-06 Richard Biener <rguenther@suse.de>
6201
6202 * basic-block.h (class control_dependences): New.
6203 * tree-ssa-dce.c (control_dependence_map): Remove.
6204 (cd): New global.
6205 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6206 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6207 find_pdom, find_control_dependence, find_all_control_dependences):
6208 Move to cfganal.c.
f43245d1
UB		 6209 (mark_control_dependent_edges_necessary,
6210 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6211 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB		 6212 * cfganal.c (set_control_dependence_map_bit,
6213 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6214 find_all_control_dependences): Move from tree-ssa-dce.c and
6215 implement as methods of control_dependences class.
6216 (control_dependences::control_dependences): New.
6217 (control_dependences::~control_dependences): Likewise.
6218 (control_dependences::get_edges_dependent_on): Likewise.
6219 (control_dependences::get_edge): Likewise.
6220
7d2ba471
JH		 62212013-09-04 Jan Hubicka <jh@suse.cz>
6222
6223 * tree.c (types_same_for_odr): Drop overactive check.
6224 * ipa-devirt.c (hash_type_name): Likewise.
6225
e18412fc
JH		 62262013-09-04 Jan Hubicka <jh@suse.cz>
6227
6228 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6229 (analyze_functions): ... here.
6230
d352b245
JH		 62312013-09-04 Jan Hubicka <jh@suse.cz>
6232
6233 PR middle-end/58201
6234 * cgraphunit.c (analyze_functions): Clear AUX fields
6235 after processing; initialize assembler name has.
6236
770da076
JL		 62372013-09-05 Jeff Law <law@redhat.com>
6238
6239 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6240 from thread_around_empty_block. Record threading path into PATH.
6241 Recurse if threading through the initial block is successful.
6242 (thread_across_edge): Corresponding changes to slightly simplify.
6243
fe6f68e2
JG		 62442013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6245
6246 * config/aarch64/aarch64.md
6247 (type): Remove frecpe, frecps, frecpx.
6248 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6249 fix to be a TARGET_SIMD instruction.
6250 (aarch64_frecps): Remove.
6251 * config/aarch64/aarch64-simd.md
6252 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6253 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6254
6e4150e1
JG		 62552013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6256 Sofiane Naci <sofiane.naci@arm.com>
6257
f43245d1 6258 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6259 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB		 6260 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6261 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6262 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6263 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6264 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6265 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6266 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6267 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6268 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG		 6269 shift* types.
6270 * config/arm/arm-fixed.md: Update for attribute changes
6271 all occurrences of arlo_* types.
6272 * config/arm/thumb2.md: Update for attribute changes all occurrences
6273 of arlo_* types.
6274 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6275 (cortexa7_older_only): Likewise.
6276 (cortexa7_younger): Likewise.
6277 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6278 (1020alu_shift_op): Likewise.
6279 (1020alu_shift_reg_op): Likewise.
6280 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6281 (alu_shift_op): Likewise.
6282 (alu_shift_reg_op): Likewise.
f43245d1 6283 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG		 6284 (11_alu_shift_op): Likewise.
6285 (11_alu_shift_reg_op): Likewise.
6286 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6287 (9_alu_shift_reg_op): Likewise.
6288 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6289 attribute changes.
6290 (cortex_a15_alu_shift): Likewise.
6291 (cortex_a15_alu_shift_reg): Likewise.
6292 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6293 attribute changes.
6294 (cortex_a5_alu_shift): Likewise.
f43245d1
UB		 6295 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6296 attribute changes.
6e4150e1 6297 (cortex_a53_alu_shift): Likewise.
f43245d1
UB		 6298 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6299 attribute changes.
6e4150e1
JG		 6300 (cortex_a7_alu_reg): Likewise.
6301 (cortex_a7_alu_shift): Likewise.
f43245d1
UB		 6302 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6303 attribute changes.
6e4150e1
JG		 6304 (cortex_a8_alu_shift): Likewise.
6305 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6306 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6307 (cortex_a9_dp_shift): Likewise.
f43245d1
UB		 6308 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6309 attribute changes.
6e4150e1
JG		 6310 * config/arm/cortex-r4.md
6311 (cortex_r4_alu): Update for attribute changes.
6312 (cortex_r4_mov): Likewise.
6313 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6314 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6315 (526_alu_shift_op): Likewise.
f43245d1
UB		 6316 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6317 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6318 (626te_alu_shift_op): Likewise.
f43245d1 6319 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG		 6320 (726te_alu_shift_op): Likewise.
6321 (726te_alu_shift_reg_op): Likewise.
6322 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6323 (mp626_alu_shift_op): Likewise.
6324 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6325 (pj4_alu_conds): Likewise.
6326 (pj4_shift): Likewise.
6327 (pj4_shift_conds): Likewise.
6328 (pj4_alu_shift): Likewise.
6329 (pj4_alu_shift_conds): Likewise.
6330 * config/aarch64/aarch64.md: Update for attribute change
6331 all occurrences of arlo_* and shift* types.
6332
e6523306
MS		 63332013-09-05 Mike Stump <mikestump@comcast.net>
6334
6335 * tree.h: Move documentation for tree_function_decl to tree-core.h
6336 with the declaration.
6337
80a18161
PB		 63382013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6339
6340 PR target/58139
6341 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6342 looking for widest mode.
6343
6994430a
EB		 63442013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6345
6346 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6347
e67f39f7
RB		 63482013-09-05 Richard Biener <rguenther@suse.de>
6349
6350 PR tree-optimization/58137
6351 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6352 Do not create vectors of pointers.
6353 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6354 types for the components of the vector initializer.
6355 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6356 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6357
568cda29
MJ		 63582013-09-05 Martin Jambor <mjambor@suse.cz>
6359
6360 * ipa-prop.c (remove_described_reference): Accept missing references,
6361 return false if that hppens, otherwise return true.
6362 (cgraph_node_for_jfunc): New function.
6363 (try_decrement_rdesc_refcount): Likewise.
6364 (try_make_edge_direct_simple_call): Use them.
6365 (ipa_edge_removal_hook): Remove references from rdescs.
6366 (ipa_edge_duplication_hook): Clone rdescs and their references
6367 when the new edge has the same caller as the old one.
6368 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6369 reference before removing any edges.
6370
5a200acb
RE		 63712013-09-05 Richard Earnshaw <rearnsha@arm.com>
6372
6373 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6374 initial store.
6375 * thumb2.md (thumb2_storewb_parisi): New pattern.
6376
5922847b
YZ		 63772013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6378
6379 * config/aarch64/aarch64-option-extensions.def: Add
6380 AARCH64_OPT_EXTENSION of 'crc'.
6381 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6382 (AARCH64_ISA_CRC): Ditto.
6383 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6384 description of the CRC extension.
6385
996746aa
AI		 63862013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6387
6388 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6389 * config/rs6000/linux.h: Ditto.
6390 * alpha/linux.h: Ditto.
6391 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6392 no_c99_libc_has_function.
6393 * config/c6x/uclinux-elf.h: Ditto.
6394 * config/lm32/uclinux-elf.h: Ditto.
6395 * config/m68k/uclinux.h: Ditto.
6396 * config/moxie/uclinux.h: Ditto.
6397 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6398 (crisv32-*-linux*, cris-*-linux*): Ditto.
6399 * config/bfin/bfin.c: Include "tm_p.h".
6400
c370d99f
RB		 64012013-09-05 Richard Biener <rguenther@suse.de>
6402
6403 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6404 check for a definition without a basic-block.
6405
003bb7f3
JG		 64062013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6407 Sofiane Naci <sofiane.naci@arm.com>
6408
6409 * config/aarch64/aarch64.md
6410 (*movti_aarch64): Rename r_2_f and f_2_r.
6411 (*movsf_aarch64): Likewise.
6412 (*movdf_aarch64): Likewise.
6413 (*movtf_aarch64): Likewise.
6414 (aarch64_movdi_<mode>low): Likewise.
6415 (aarch64_movdi_<mode>high): Likewise.
6416 (aarch64_mov<mode>high_di): Likewise.
6417 (aarch64_mov<mode>low_di): Likewise.
6418 (aarch64_movtilow_tilow): Likewise.
6419 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6420 values to config/arm/types.md
6421 (attribute "conds"): Update for attribute change.
6422 (anddi3_insn): Likewise.
6423 (iordi3_insn): Likewise.
6424 (xordi3_insn): Likewise.
6425 (one_cmpldi2): Likewise.
6426 * config/arm/types.md (type): Add Neon types.
6427 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6428 use "type" attribute.
6429 (movmisalign<mode>_neon_store): Likewise.
6430 (movmisalign<mode>_neon_load): Likewise.
6431 (vec_set<mode>_internal): Likewise.
6432 (vec_setv2di_internal): Likewise.
6433 (vec_extract<mode>): Likewise.
6434 (vec_extractv2di): Likewise.
6435 (add<mode>3_neon): Likewise.
6436 (adddi3_neon): Likewise.
6437 (sub<mode>3_neon): Likewise.
6438 (subdi3_neon): Likewise.
6439 (mul<mode>3_neon): Likewise.
6440 (mul<mode>3add<mode>_neon): Likewise.
6441 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6442 (fma<VCVTF:mode>4)): Likewise.
6443 (fma<VCVTF:mode>4_intrinsic): Likewise.
6444 (fmsub<VCVTF:mode>4)): Likewise.
6445 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6446 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6447 (ior<mode>3): Likewise.
6448 (and<mode>3): Likewise.
6449 (anddi3_neon): Likewise.
6450 (orn<mode>3_neon): Likewise.
6451 (orndi3_neon): Likewise.
6452 (bic<mode>3_neon): Likewise.
6453 (bicdi3_neon): Likewise.
6454 (xor<mode>3): Likewise.
6455 (one_cmpl<mode>2): Likewise.
6456 (abs<mode>2): Likewise.
6457 (neg<mode>2): Likewise.
6458 (umin<mode>3_neon): Likewise.
6459 (umax<mode>3_neon): Likewise.
6460 (smin<mode>3_neon): Likewise.
6461 (smax<mode>3_neon): Likewise.
6462 (vashl<mode>3): Likewise.
6463 (vashr<mode>3_imm): Likewise.
6464 (vlshr<mode>3_imm): Likewise.
6465 (ashl<mode>3_signed): Likewise.
6466 (ashl<mode>3_unsigned): Likewise.
6467 (neon_load_count): Likewise.
6468 (ashldi3_neon_noclobber): Likewise.
6469 (signed_shift_di3_neon): Likewise.
6470 (unsigned_shift_di3_neon): Likewise.
6471 (ashrdi3_neon_imm_noclobber): Likewise.
6472 (lshrdi3_neon_imm_noclobber): Likewise.
6473 (widen_ssum<mode>3): Likewise.
6474 (widen_usum<mode>3): Likewise.
6475 (quad_halves_<code>v4si): Likewise.
6476 (quad_halves_<code>v4sf): Likewise.
6477 (quad_halves_<code>v8hi): Likewise.
6478 (quad_halves_<code>v16qi): Likewise.
6479 (reduc_splus_v2di): Likewise.
6480 (neon_vpadd_internal<mode>): Likewise.
6481 (neon_vpsmin<mode>): Likewise.
6482 (neon_vpsmax<mode>): Likewise.
6483 (neon_vpumin<mode>): Likewise.
6484 (neon_vpumax<mode>): Likewise.
6485 (ss_add<mode>_neon): Likewise.
6486 (us_add<mode>_neon): Likewise.
6487 (ss_sub<mode>_neon): Likewise.
6488 (us_sub<mode>_neon): Likewise.
6489 (neon_vadd<mode>_unspec): Likewise.
6490 (neon_vaddl<mode>): Likewise.
6491 (neon_vaddw<mode>): Likewise.
6492 (neon_vhadd<mode>): Likewise.
6493 (neon_vqadd<mode>): Likewise.
6494 (neon_vaddhn<mode>): Likewise.
6495 (neon_vmul<mode>): Likewise.
6496 (neon_vmla<mode>): Likewise.
6497 (neon_vmlal<mode>): Likewise.
6498 (neon_vmls<mode>): Likewise.
6499 (neon_vmlsl<mode>): Likewise.
6500 (neon_vqdmulh<mode>): Likewise.
6501 (neon_vqdmlal<mode>): Likewise.
6502 (neon_vqdmlsl<mode>): Likewise.
6503 (neon_vmull<mode>): Likewise.
6504 (neon_vqdmull<mode>): Likewise.
6505 (neon_vsub<mode>_unspec): Likewise.
6506 (neon_vsubl<mode>): Likewise.
6507 (neon_vsubw<mode>): Likewise.
6508 (neon_vqsub<mode>): Likewise.
6509 (neon_vhsub<mode>): Likewise.
6510 (neon_vsubhn<mode>): Likewise.
6511 (neon_vceq<mode>): Likewise.
6512 (neon_vcge<mode>): Likewise.
6513 (neon_vcgeu<mode>): Likewise.
6514 (neon_vcgt<mode>): Likewise.
6515 (neon_vcgtu<mode>): Likewise.
6516 (neon_vcle<mode>): Likewise.
6517 (neon_vclt<mode>): Likewise.
6518 (neon_vcage<mode>): Likewise.
6519 (neon_vcagt<mode>): Likewise.
6520 (neon_vtst<mode>): Likewise.
6521 (neon_vabd<mode>): Likewise.
6522 (neon_vabdl<mode>): Likewise.
6523 (neon_vaba<mode>): Likewise.
6524 (neon_vabal<mode>): Likewise.
6525 (neon_vmax<mode>): Likewise.
6526 (neon_vmin<mode>): Likewise.
6527 (neon_vpaddl<mode>): Likewise.
6528 (neon_vpadal<mode>): Likewise.
6529 (neon_vpmax<mode>): Likewise.
6530 (neon_vpmin<mode>): Likewise.
6531 (neon_vrecps<mode>): Likewise.
6532 (neon_vrsqrts<mode>): Likewise.
6533 (neon_vqabs<mode>): Likewise.
6534 (neon_vqneg<mode>): Likewise.
6535 (neon_vcls<mode>): Likewise.
6536 (clz<mode>2): Likewise.
6537 (popcount<mode>2): Likewise.
6538 (neon_vrecpe): Likewise.
6539 (neon_vrsqrte): Likewise.
6540 (neon_vget_lane<mode>_sext_internal): Likewise.
6541 (neon_vget_lane<mode>_zext_internal): Likewise.
6542 (neon_vdup_n<mode>): Likewise.
6543 (neon_vdup_nv2di): Likewise.
6544 (neon_vdpu_lane<mode>_internal): Likewise.
6545 (neon_vswp<mode>): Likewise.
6546 (float<mode><V_cvtto>2): Likewise.
6547 (floatuns<mode><V_cvtto>2): Likewise.
6548 (fix_trunc<mode><V_cvtto>)2): Likewise
6549 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6550 (neon_vcvt<mode>): Likewise.
6551 (neon_vcvtv4sfv4hf): Likewise.
6552 (neon_vcvtv4hfv4sf): Likewise.
6553 (neon_vcvt_n<mode>): Likewise.
6554 (neon_vmovn<mode>): Likewise.
6555 (neon_vqmovn<mode>): Likewise.
6556 (neon_vqmovun<mode>): Likewise.
6557 (neon_vmovl<mode>): Likewise.
6558 (neon_vmul_lane<mode>): Likewise.
6559 (neon_vmull_lane<mode>): Likewise.
6560 (neon_vqdmull_lane<mode>): Likewise.
6561 (neon_vqdmulh_lane<mode>): Likewise.
6562 (neon_vmla_lane<mode>): Likewise.
6563 (neon_vmlal_lane<mode>): Likewise.
6564 (neon_vqdmlal_lane<mode>): Likewise.
6565 (neon_vmls_lane<mode>): Likewise.
6566 (neon_vmlsl_lane<mode>): Likewise.
6567 (neon_vqdmlsl_lane<mode>): Likewise.
6568 (neon_vext<mode>): Likewise.
6569 (neon_vrev64<mode>): Likewise.
6570 (neon_vrev32<mode>): Likewise.
6571 (neon_vrev16<mode>): Likewise.
6572 (neon_vbsl<mode>_internal): Likewise.
6573 (neon_vshl<mode>): Likewise.
6574 (neon_vqshl<mode>): Likewise.
6575 (neon_vshr_n<mode>): Likewise.
6576 (neon_vshrn_n<mode>): Likewise.
6577 (neon_vqshrn_n<mode>): Likewise.
6578 (neon_vqshrun_n<mode>): Likewise.
6579 (neon_vshl_n<mode>): Likewise.
6580 (neon_vqshl_n<mode>): Likewise.
6581 (neon_vqshlu_n<mode>): Likewise.
6582 (neon_vshll_n<mode>): Likewise.
6583 (neon_vsra_n<mode>): Likewise.
6584 (neon_vsri_n<mode>): Likewise.
6585 (neon_vsli_n<mode>): Likewise.
6586 (neon_vtbl1v8qi): Likewise.
6587 (neon_vtbl2v8qi): Likewise.
6588 (neon_vtbl3v8qi): Likewise.
6589 (neon_vtbl4v8qi): Likewise.
6590 (neon_vtbx1v8qi): Likewise.
6591 (neon_vtbx2v8qi): Likewise.
6592 (neon_vtbx3v8qi): Likewise.
6593 (neon_vtbx4v8qi): Likewise.
6594 (neon_vtrn<mode>_internal): Likewise.
6595 (neon_vzip<mode>_internal): Likewise.
6596 (neon_vuzp<mode>_internal): Likewise.
6597 (neon_vld1<mode>): Likewise.
6598 (neon_vld1_lane<mode>): Likewise.
6599 (neon_vld1_dup<mode>): Likewise.
6600 (neon_vld1_dupv2di): Likewise.
6601 (neon_vst1<mode>): Likewise.
6602 (neon_vst1_lane<mode>): Likewise.
6603 (neon_vld2<mode>): Likewise.
6604 (neon_vld2_lane<mode>): Likewise.
6605 (neon_vld2_dup<mode>): Likewise.
6606 (neon_vst2<mode>): Likewise.
6607 (neon_vst2_lane<mode>): Likewise.
6608 (neon_vld3<mode>): Likewise.
6609 (neon_vld3qa<mode>): Likewise.
6610 (neon_vld3qb<mode>): Likewise.
6611 (neon_vld3_lane<mode>): Likewise.
6612 (neon_vld3_dup<mode>): Likewise.
6613 (neon_vst3<mode>): Likewise.
6614 (neon_vst3qa<mode>): Likewise.
6615 (neon_vst3qb<mode>): Likewise.
6616 (neon_vst3_lane<mode>): Likewise.
6617 (neon_vld4<mode>): Likewise.
6618 (neon_vld4qa<mode>): Likewise.
6619 (neon_vld4qb<mode>): Likewise.
6620 (neon_vld4_lane<mode>): Likewise.
6621 (neon_vld4_dup<mode>): Likewise.
6622 (neon_vst4<mode>): Likewise.
6623 (neon_vst4qa<mode>): Likewise.
6624 (neon_vst4qb<mode>): Likewise.
6625 (neon_vst4_lane<mode>): Likewise.
6626 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6627 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6628 (neon_vec_<US>mult_lo_<mode>): Likewise.
6629 (neon_vec_<US>mult_hi_<mode>): Likewise.
6630 (neon_vec_<US>shiftl_<mode>): Likewise.
6631 (neon_unpack<US>_<mode>): Likewise.
6632 (neon_vec_<US>mult_<mode>): Likewise.
6633 (vec_pack_trunc_<mode>): Likewise.
6634 (neon_vec_pack_trunk_<mode>): Likewise.
6635 (neon_vabd<mode>_2): Likewise.
6636 (neon_vabd<mode>_3): Likewise.
6637 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6638 (thumb2_movsi_vfp): Likewise.
6639 (movdi_vfp): Likewise.
6640 (movdi_vfp_cortexa8): Likewise.
6641 (movhf_vfp_neon): Likewise.
6642 (movhf_vfp): Likewiwse.
6643 (movsf_vfp): Likewiwse.
6644 (thumb2_movsf_vfp): Likewiwse.
6645 (movdf_vfp): Likewise.
6646 (thumb2_movdf_vfp): Likewise.
6647 (movsfcc_vfp): Likewise.
6648 (thumb2_movsfcc_vfp): Likewise.
6649 (movdfcc_vfp): Likewise.
6650 (thumb2_movdfcc_vfp): Likewise.
6651 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6652 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6653 (v10_v2c): Likewise.
6654 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6655 attribute change.
6656 (cortex_a15_neon_int_2): Likewise.
6657 (cortex_a15_neon_int_3): Likewise.
6658 (cortex_a15_neon_int_4): Likewise.
6659 (cortex_a15_neon_int_5): Likewise.
6660 (cortex_a15_neon_vqneg_vqabs): Likewise.
6661 (cortex_a15_neon_vmov): Likewise.
6662 (cortex_a15_neon_vaba): Likewise.
6663 (cortex_a15_neon_vaba_qqq): Likewise.
6664 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6665 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6666 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6667 scalar_64_32_long_scalar): Likewise.
6668 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6669 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6670 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6671 lotype_qdd_64_32_long): Likewise.
6672 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6673 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6674 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6675 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6676 (cortex_a15_neon_shift_1): Likewise.
6677 (cortex_a15_neon_shift_2): Likewise.
6678 (cortex_a15_neon_shift_3): Likewise.
6679 (cortex_a15_neon_vshl_ddd): Likewise.
6680 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6681 (cortex_a15_neon_vsra_vrsra): Likewise.
6682 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6683 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6684 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6685 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6686 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6687 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6688 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6689 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6690 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6691 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6692 (cortex_a15_neon_bp_simple): Likewise.
6693 (cortex_a15_neon_bp_2cycle): Likewise.
6694 (cortex_a15_neon_bp_3cycle): Likewise.
6695 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6696 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6697 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6698 (cortex_a15_neon_vld2_4_regs): Likewise.
6699 (cortex_a15_neon_vld3_vld4): Likewise.
6700 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6701 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6702 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6703 (cortex_a15_neon_vst3_vst4): Likewise.
6704 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6705 (cortex_a15_neon_vld3_vld4_lane" 10
6706 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6707 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6708 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6709 (cortex_a15_neon_ldm_2): Likewise.0
6710 (cortex_a15_neon_stm_2): Likewise.
6711 (cortex_a15_neon_mcr): Likewise.
6712 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6713 (cortex_a15_neon_mrc): Likewise.
6714 (cortex_a15_neon_mrrc): Likewise.
6715 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6716 change.
6717 (cortex_a15_alu_shift): Likewise.
6718 (cortex_a15_alu_shift_reg): Likewise.
6719 (cortex_a15_mult32): Likewise.
6720 (cortex_a15_mult64): Likewise.
6721 (cortex_a15_block): Likewise.
6722 (cortex_a15_branch): Likewise.
6723 (cortex_a15_load1): Likewise.
6724 (cortex_a15_load3): Likewise.
6725 (cortex_a15_store1): Likewise.
6726 (cortex_a15_store3): Likewise.
6727 (cortex_a15_call): Likewise.
f43245d1 6728 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG		 6729 (cortex_a5_f2r): Likewise.
6730 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6731 change.
6732 (cortex_a53_f2r): Likewise.
6733 * config/arm/cortex-a7.md
6734 (cortex_a7_branch): Update for attribute change.
6735 (cortex_a7_call): Likewise.
6736 (cortex_a7_alu_imm): Likewise.
6737 (cortex_a7_alu_reg): Likewise.
6738 (cortex_a7_alu_shift): Likewise.
6739 (cortex_a7_mul): Likewise.
6740 (cortex_a7_load1): Likewise.
6741 (cortex_a7_store1): Likewise.
6742 (cortex_a7_load2): Likewise.
6743 (cortex_a7_store2): Likewise.
6744 (cortex_a7_load3): Likewise.
6745 (cortex_a7_store3): Likewise.
6746 (cortex_a7_load4): Likewise.
6747 (cortex_a7_store4): Likewise.
6748 (cortex_a7_fpalu): Likewise.
6749 (cortex_a7_fconst): Likewise.
6750 (cortex_a7_fpmuls): Likewise.
6751 (cortex_a7_neon_mul): Likewise.
6752 (cortex_a7_fpmacs): Likewise.
6753 (cortex_a7_neon_mla: Likewise.
6754 (cortex_a7_fpmuld: Likewise.
6755 (cortex_a7_fpmacd: Likewise.
6756 (cortex_a7_fpfmad: Likewise.
6757 (cortex_a7_fdivs: Likewise.
6758 (cortex_a7_fdivd: Likewise.
6759 (cortex_a7_r2f: Likewise.
6760 (cortex_a7_f2r: Likewise.
6761 (cortex_a7_f_flags: Likewise.
6762 (cortex_a7_f_loads: Likewise.
6763 (cortex_a7_f_loadd: Likewise.
6764 (cortex_a7_f_stores: Likewise.
6765 (cortex_a7_f_stored: Likewise.
6766 (cortex_a7_neon): Likewise.
6767 * config/arm/cortex-a8-neon.md
6768 (cortex_a8_neon_mrc): Update for attribute change.
6769 (cortex_a8_neon_mrrc): Likewise.
6770 (cortex_a8_neon_int_1): Likewise.
6771 (cortex_a8_neon_int_2): Likewise.
6772 (cortex_a8_neon_int_3): Likewise.
6773 (cortex_a8_neon_int_4): Likewise.
6774 (cortex_a8_neon_int_5): Likewise.
6775 (cortex_a8_neon_vqneg_vqabs): Likewise.
6776 (cortex_a8_neon_vmov): Likewise.
6777 (cortex_a8_neon_vaba): Likewise.
6778 (cortex_a8_neon_vaba_qqq): Likewise.
6779 (cortex_a8_neon_vsma): Likewise.
6780 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6781 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 6782 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6783 long_scalar): Likewise.
003bb7f3
JG		 6784 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6785 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6786 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6787 long_scalar_qdd_64_32_long): Likewise.
6788 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6789 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6790 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6791 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6792 (cortex_a8_neon_shift_1): Likewise.
6793 (cortex_a8_neon_shift_2): Likewise.
6794 (cortex_a8_neon_shift_3): Likewise.
6795 (cortex_a8_neon_vshl_ddd): Likewise.
6796 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6797 (cortex_a8_neon_vsra_vrsra): Likewise.
6798 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6799 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6800 (cortex_a8_neon_fp_vsum): Likewise.
6801 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6802 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6803 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6804 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6805 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6806 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6807 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6808 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6809 (cortex_a8_neon_bp_simple): Likewise.
6810 (cortex_a8_neon_bp_2cycle): Likewise.
6811 (cortex_a8_neon_bp_3cycle): Likewise.
6812 (cortex_a8_neon_ldr): Likewise.
6813 (cortex_a8_neon_str): Likewise.
6814 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6815 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6816 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6817 (cortex_a8_neon_vld2_4_regs): Likewise.
6818 (cortex_a8_neon_vld3_vld4): Likewise.
6819 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6820 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6821 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6822 (cortex_a8_neon_vst3_vst4): Likewise.
6823 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6824 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6825 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6826 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6827 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6828 (cortex_a8_neon_mcr): Likewise.
6829 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 6830 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG		 6831 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6832 change.
6833 (ca9_neon_mrrc): Likewise.
6834 (cortex_a9_neon_int_1): Likewise.
6835 (cortex_a9_neon_int_2): Likewise.
6836 (cortex_a9_neon_int_3): Likewise.
6837 (cortex_a9_neon_int_4): Likewise.
6838 (cortex_a9_neon_int_5): Likewise.
6839 (cortex_a9_neon_vqneg_vqabs): Likewise.
6840 (cortex_a9_neon_vmov): Likewise.
6841 (cortex_a9_neon_vaba): Likewise.
6842 (cortex_a9_neon_vaba_qqq): Likewise.
6843 (cortex_a9_neon_vsma): Likewise.
6844 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6845 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 6846 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6847 long_scalar): Likewise.
003bb7f3
JG		 6848 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6849 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6850 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6851 long_scalar_qdd_64_32_long): Likewise.
6852 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6853 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6854 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6855 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6856 (cortex_a9_neon_shift_1): Likewise.
6857 (cortex_a9_neon_shift_2): Likewise.
6858 (cortex_a9_neon_shift_3): Likewise.
6859 (cortex_a9_neon_vshl_ddd): Likewise.
6860 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6861 (cortex_a9_neon_vsra_vrsra): Likewise.
6862 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6863 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6864 (cortex_a9_neon_fp_vsum): Likewise.
6865 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6866 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6867 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6868 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6869 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6870 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6871 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6872 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6873 (cortex_a9_neon_bp_simple): Likewise.
6874 (cortex_a9_neon_bp_2cycle): Likewise.
6875 (cortex_a9_neon_bp_3cycle): Likewise.
6876 (cortex_a9_neon_ldr): Likewise.
6877 (cortex_a9_neon_str): Likewise.
6878 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6879 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6880 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6881 (cortex_a9_neon_vld2_4_regs): Likewise.
6882 (cortex_a9_neon_vld3_vld4): Likewise.
6883 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6884 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6885 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6886 (cortex_a9_neon_vst3_vst4): Likewise.
6887 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6888 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6889 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6890 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6891 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6892 (cortex_a9_neon_mcr): Likewise.
6893 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6894 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6895 (cortex_a9_fps): Likewise.
6896 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6897 change.
6898 (cortex_m4_fmuls): Likewise.
6899 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6900 change.
6901 (cortex_r4_mrc): Likewise.
6902 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 6903 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 6904 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB		 6905 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6906 attribute change.
003bb7f3
JG		 6907 (pj4_core_to_vfp): Likewise.
6908 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6909 attribute change.
6910 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6911 (vfp_fstore): Likewise.
6912 * doc/md.texi: Change references to neon_type to refer to type.
6913
fbe0dc78
DS		 69142013-09-04 Dodji Seketeli <dodji@redhat.com>
6915
6916 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6917
c167bc5b
DE		 69182013-09-04 David Edelsohn <dje.gcc@gmail.com>
6919
6920 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6921 lglobl if not weak.
6922
035cb59f
ER		 69232013-09-04 Easwaran Raman <eraman@google.com>
6924
6925 PR middle-end/57370
6926 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6927 (build_and_add_sum): Use it.
6928 (appears_later_in_bb): Simplify code.
6929
71191083
TJ		 69302013-09-04 Teresa Johnson <tejohnson@google.com>
6931
6932 * dumpfile.c (dump_finish): Don't close stderr/stdout.
6933
0c488213
JG		 69342013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
6935
6936 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
6937
61a74079
JH		 69382013-09-04 Jan Hubicka <jh@suse.cz>
6939
6940 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
6941 * ipa-devirt.c: Include diganostic.h
6942 (odr_type_d): Add types and types_set.
6943 (hash_type_name): Work for types with vtables during LTO.
6944 (odr_hasher::remove): Fix comment; destroy types_set.
6945 (add_type_duplicate): New function,
6946 (get_odr_type): Use it.
6947 (dump_type_inheritance_graph): Dump type duplicates.
6948 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
6949 graph.
6950 * tree.c (types_same_for_odr): Give exact answers on types with
6951 virtual tables.
6952
24ba8195
DS		 69532013-09-04 Dodji Seketeli <dodji@redhat.com>
6954
6955 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 6956 explaining their differences.
24ba8195 6957
0a04869e
SKS		 69582013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
6959
6960 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
6961
3b18bc42
JL		 69622013-09-03 Jeff Law <law@redhat.com>
6963
6964 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
6965 when not threading through a joiner block. Pass joiner/no joiner
6966 state to register_jump_thread.
6967 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
6968 state from argument rather than implying on path length.
6969 Dump the entire jump thread path into debugging dump.
6970 * tree-flow.h (register_jump_thread): Update prototype.
6971
ae84e151
XDL		 69722013-08-29 Xinliang David Li <davidxl@google.com>
6973
6974 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6975 Remove a trivial gcc_assert.
6976
c716e67f
XDL		 69772013-08-29 Xinliang David Li <davidxl@google.com>
6978
6979 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
6980 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
6981 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6982 Delay base decl alignment adjustment.
6983 * tree-vectorizer.c (vect_destroy_datarefs): New function.
6984 * tree-vectorizer.h: New data structure.
6985 (set_dr_misalignment): New function.
6986 (dr_misalignment): Ditto.
6987 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
6988 (vectorizable_load): Ditto.
6989 (ensure_base_align): New function.
6990 (vectorize_loops): Add dbg_cnt support.
6991 (execute_vect_slp): Ditto.
6992 * dbgcnt.def: New debug counter.
6993 * Makefile: New dependency.
6994
db8800bc
MI		 69952013-09-03 Meador Inge <meadori@codesourcery.com>
6996
6997 Revert:
6998
6999 2013-08-30 Meador Inge <meadori@codesourcery.com>
7000
7001 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7002
0d9b4f5c
DE		 70032013-09-03 David Edelsohn <dje.gcc@gmail.com>
7004
7005 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7006 function descriptor.
7007
29e810b9
RB		 70082013-09-03 Richard Biener <rguenther@suse.de>
7009
7010 * tree-affine.c (add_elt_to_tree): Fix association issue,
7011 avoid useless converts and make sure to always return a
7012 properly typed result.
7013
dbfc2894
RB		 70142013-09-03 Richard Biener <rguenther@suse.de>
7015
7016 PR middle-end/57656
7017 * fold-const.c (negate_expr_p): Fix division case.
7018 (negate_expr): Likewise.
7019
fa3bf4ce
RB		 70202013-09-03 Richard Biener <rguenther@suse.de>
7021
7022 PR lto/58285
7023 * tree-streamer-out.c: Include tm.h.
7024 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7025
a64bbb3f
JH		 70262013-09-03 Jan Hubicka <jh@suse.cz>
7027
7028 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7029
93ed07e2
AM		 70302013-09-03 Alan Modra <amodra@gmail.com>
7031
7032 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7033 multilibs to big-endian target and vice versa.
7034 * config/rs6000/t-linux64: Use := assignment on all vars.
7035 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7036 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7037 * config/rs6000/t-linux64le: New file.
7038 * config/rs6000/t-linux64bele: New file.
7039 * config/rs6000/t-linux64lebe: New file.
7040
07d6cd64
JH		 70412013-09-02 Jan Hubicka <jh@suse.cz>
7042
7043 * ipa-inline-transform.c (inline_transform): Do not
7044 optimize_inline_calls when not optimizing.
7045
8e4c9a10
JH		 70462013-09-02 Jan Hubicka <jh@suse.cz>
7047
7048 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7049 duplicated nodes for assembler names.
7050 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7051 hard registers from assembler name hash.
7052
c91061e6
JH		 70532013-09-02 Jan Hubicka <jh@suse.cz>
7054
7055 * ipa-split.c (execute_split_functions): Split externally visible
7056 functions called once.
7057
2fd0985c
MJ		 70582013-09-02 Martin Jambor <mjambor@suse.cz>
7059
7060 PR ipa/58106
7061 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7062 linked list. When finding the correct duplicate, also consider also
7063 the caller in additon to its inlined_to node.
7064
91bd4114
JG		 70652013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7066
7067 * config/aarch64/aarch64-simd-builtins.def
7068 (dup_lane_scalar): Remove.
7069 * config/aarch64/aarch64-simd.md
7070 (aarch64_simd_dup): Add 'w->w' alternative.
7071 (aarch64_dup_lane<mode>): Allow for VALL.
7072 (aarch64_dup_lane_scalar<mode>): Remove.
7073 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7074 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7075 (aarch64_get_lane_unsigned<mode>): Likewise.
7076 (aarch64_get_lane<mode>): Likewise.
7077 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7078 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7079 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7080 (VCON): Change container of V2SF.
7081 (vswap_width_name): Likewise.
7082 * config/aarch64/arm_neon.h
7083 (__aarch64_vdup_lane_any): New.
7084 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7085 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7086 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7087
d617d2d8
EB		 70882013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7089
7090 PR middle-end/56382
7091 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7092 the source or the destination is a single hard register.
7093
63fdb7be
RB		 70942013-09-02 Richard Biener <rguenther@suse.de>
7095
7096 PR middle-end/57511
7097 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7098 non-linear SCEVs.
7099
78de2333
RB		 71002013-09-02 Richard Biener <rguenther@suse.de>
7101
7102 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7103 arithmetic to sizetype.
7104
85ff4ec6
BC		 71052013-09-02 Bin Cheng <bin.cheng@arm.com>
7106
7107 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7108 Find auto-increment use both before and after candidate.
7109
576016fe
MP		 71102013-09-02 Marek Polacek <polacek@redhat.com>
7111
7112 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7113
08f835dc
JH		 71142013-09-01 Jan Hubicka <jh@suse.cz>
7115
7116 * Makefile.in: Add ipa-profile.o
7117 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7118 * cgraph.c (struct cgraph_propagate_frequency_data,
7119 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7120 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7121 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB		 7122 * ipa-inline-analysis.c: Include ipa-utils.h;
7123 drop duplicated cfgloop.h.
08f835dc
JH		 7124 (inline_update_callee_summaries): Update.
7125 * ipa-profile.c: New file.
7126 * ipa-utils.h (ipa_propagate_frequency): Declare.
7127 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 7128 data-streamer.h, value-prof.h.
08f835dc
JH		 7129 (symtab_remove_unreachable_nodes): Update profile.
7130 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7131 account_time_size, cmp_counts, dump_histogram,
7132 ipa_profile_generate_summary, ipa_profile_write_summary,
7133 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7134 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 7135 Move to ipa-profile.c.
08f835dc 7136
cec8583c
JDA		 71372013-09-01 John David Anglin <danglin@gcc.gnu.org>
7138
7139 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7140
bbc9396b
JH		 71412013-09-01 Jan Hubicka <jh@suse.cz>
7142
7143 * common.opt (fdevirtualize-speculatively): New function.
7144 * invoke.texi (fdevirtualize-speculatively): Document.
7145 * ipa-devirt.c: Include ipa-inline.h
7146 (likely_target_p): New function.
7147 (ipa_devirt): New function.
7148 (gate_ipa_devirt): New function.
7149 (pass_data_ipa_devirt): New static var.
7150 (pass_ipa_devirt): Likewise.
7151 (make_pass_ipa_devirt): New function.
7152 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7153 (common_handle_option): Disable devirtualization when
7154 value range profiling is available.
7155 * passes.def (pass_ipa_devirt): Add.
7156 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 7157 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 7158
0cea211e
IS		 71592013-09-01 Iain Sandoe <iain@codesourcery.com>
7160
7161 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7162 include sanitize(undefined).
7163
5e351e96
DN		 71642013-08-31 Diego Novillo <dnovillo@google.com>
7165
7166 * Makefile.in (TREE_CORE_H): Define.
7167 (TREE_H): Use.
7168 (GTFILES): Add tree-core.h.
7169 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7170 size the array.
7171 * tree-core.h: New file.
7172 Move all data structures, enum, typedefs, global
7173 declarations and constants from ...
7174 * tree.h: ... here.
7175
63bf9a90
JH		 71762013-08-31 Jan Hubicka <jh@suse.cz>
7177
7178 * bulitins.c (expand_builtin): Do not early exit for gcov
7179 instrumented functions.
7180
5c0abd6a
MP		 71812013-08-31 Marek Polacek <polacek@redhat.com>
7182
7183 * ubsan.c: Include tm_p.h.
7184
4843f032
JH		 71852013-08-31 Jan Hubicka <jh@suse.cz>
7186
8356c89c
JH		 7187 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7188 warning.
7189
4843f032
JH		 7190 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7191 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7192 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 7193 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7194 (ipa_merge_profiles): New function.
7195 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7196 (lto_input_function_body): Likewise.
5e581212
JH		 7197 * ipa-utils.h (ipa_merge_profiles): Declare.
7198 * lto-streamer.h (lto_input_function_body): Update prototype.
7199 (emit_label_in_global_context_p): Remove.
7200 * lto-symtab.c: Include ipa-utils.h
7201 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7202
57292ce9
JH		 72032013-08-31 Jan Hubicka <jh@suse.cz>
7204
7205 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7206
b9cbfeeb
JH		 72072013-08-31 Jan Hubicka <jh@suse.cz>
7208
7209 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7210
41dedebd
UB		 72112013-08-31 Uros Bizjak <ubizjak@gmail.com>
7212
7213 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7214 "cmp" RTX before signed_comparison_operator check to account
7215 for "code" changes.
7216
06d65050
JH		 72172013-08-30 Jan Hubicka <jh@suse.cz>
7218
41dedebd 7219 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 7220 (detect_type_change_1): Rename to ...
7221 (detect_type_change): ... this one; early return on non-polymorphic
7222 types.
41dedebd 7223 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 7224 use of detect_type_change.
7225 (compute_complex_assign_jump_func): Add param_type parameter;
7226 update use of detect_type_change_ssa.
7227 (compute_complex_ancestor_jump_func): Likewise.
7228 (ipa_get_callee_param_type): New function.
7229 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7230 update calls to the jump function computation functions.
7231
600b5b1d
TJ		 72322013-08-30 Teresa Johnson <tejohnson@google.com>
7233 Steven Bosscher <steven@gcc.gnu.org>
7234
7235 * cfgrtl.c (fixup_new_cold_bb): New routine.
7236 (commit_edge_insertions): Invoke fixup_partitions.
7237 (find_partition_fixes): New routine.
7238 (fixup_partitions): Ditto.
7239 (verify_hot_cold_block_grouping): Update comments.
7240 (rtl_verify_edges): Invoke find_partition_fixes.
7241 (rtl_verify_bb_pointers): Update comments.
7242 (rtl_verify_bb_layout): Ditto.
7243 * basic-block.h (probably_never_executed_edge_p): Declare.
7244 (fixup_partitions): Ditto.
7245 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7246 * bb-reorder.c (sanitize_hot_paths): New function.
7247 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7248 sanitize_hot_paths.
7249 * predict.c (probably_never_executed_edge_p): New routine.
7250 * cfg.c (check_bb_profile): Add partition insanity warnings.
7251
0a6ea5c9
MI		 72522013-08-30 Meador Inge <meadori@codesourcery.com>
7253
7254 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7255
de5a5fa1
MP		 72562013-08-30 Marek Polacek <polacek@redhat.com>
7257
7258 * Makefile.in (ubsan.o): Add.
7259 (c-family/c-ubsan.o): Add.
7260 (builtins.o): Add ubsan.h dependency.
7261 * ubsan.h: New file.
7262 * ubsan.c: New file.
7263 * common.opt: Add -fsanitize=undefined option.
7264 (flag_sanitize): Add variable.
7265 (fsanitize=): Add option. Add Driver.
7266 (fsanitize=thread): Remove option.
7267 (fsanitize=address): Likewise.
7268 (static-libubsan): New option.
7269 * doc/invoke.texi: Document the new flag and -static-libubsan.
7270 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7271 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7272 * builtin-attrs.def (ATTR_COLD): Define.
7273 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7274 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7275 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7276 * flag-types.h (sanitize_code): New enum.
7277 * opts.c (common_handle_option): Parse command line arguments
7278 of -fsanitize=. Add -fsanitize=unreachable option.
7279 * varasm.c (get_variable_section): Adjust.
7280 (assemble_noswitch_variable): Likewise.
7281 (assemble_variable): Likewise.
7282 (output_constant_def_contents): Likewise.
7283 (categorize_decl_for_section): Likewise.
7284 (place_block_symbol): Likewise.
7285 (output_object_block): Likewise.
7286 * builtins.def: Likewise.
7287 * toplev.c (compile_file): Likewise.
7288 (process_options): Likewise.
7289 * cppbuiltin.c: Likewise.
7290 * tsan.c (tsan_pass): Likewise.
7291 (tsan_gate): Likewise.
7292 (tsan_gate_O0): Likewise.
7293 * cfgexpand.c (partition_stack_vars): Likewise.
7294 (expand_stack_vars): Likewise.
7295 (defer_stack_allocation): Likewise.
7296 (expand_used_vars): Likewise.
7297 * cfgcleanup.c (old_insns_match_p): Likewise.
7298 * asan.c (asan_finish_file): Likewise.
7299 (asan_instrument): Likewise.
7300 (gate_asan): Likewise.
7301 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7302 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7303 (asan_global_struct): Use pointer_sized_int_node instead
7304 calling build_nonstandard_integer_type.
7305 (initialize_sanitizer_builtins): Likewise.
7306 (asan_finish_file): Likewise.
7307 * gcc.c: Document %{%:function(args):X}.
7308 (static_spec_functions): Add sanitize.
7309 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7310 store funcval != NULL there.
7311 (do_spec_1): Adjust handle_spec_function caller.
7312 (handle_braces): Allow %:function(args) as condition.
7313 (sanitize_spec_function): New function.
7314 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7315 (LIBUBSAN_SPEC): Likewise.
7316 (LIBUBSAN_EARLY_SPEC): Likewise.
7317 (SANITIZER_SPEC): Handle libubsan.
7318 (SANITIZER_EARLY_SPEC): Likewise.
7319 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7320 instead of fsanitize=address.
7321 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7322 instead of fsanitize=address*.
7323 * builtins.c: Include ubsan.h.
7324 (fold_builtin_0): Instrument __builtin_unreachable.
7325 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7326 instead of flag_asan.
7327 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7328 (pointer_sized_int_node): Define.
41dedebd 7329 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7330
f07f30cf
MS		 73312013-08-30 Mike Stump <mikestump@comcast.net>
7332
7333 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7334 with RE patterns.
7335
8a41354f
JH		 73362013-08-29 Jan Hubicka <jh@suse.cz>
7337
7338 * cgraph.c (cgraph_function_body_availability): Handle weakref
7339 correctly.
7340 * passes.def: Remove pass_fixup_cfg.
7341 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7342 track when we need to remove functions.
7343 (gate_ipa_inline): Execute inlining always; add comment why.
7344 (pass_data_ipa_inline): Remove TODO_remove_functions.
7345 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7346 do not produce summaries.
7347 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7348 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7349 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7350 constant pool and vtable.
7351
dc8d7a0f
TB		 73522013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7353
41dedebd
UB		 7354 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7355 New arm_neon.h's internal macros to specify 64-bit constants.
7356 Avoid using stdint.h's macros.
dc8d7a0f 7357
ac7eacd2
JR		 73582013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7359
7360 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7361 before applying REG_P.
7362
8b29fd4e
JJ		 73632013-08-30 Jakub Jelinek <jakub@redhat.com>
7364
7365 PR tree-optimization/58277
7366 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7367 after seeing too many stmts with vdef in between dombb and current
7368 bb, invalidate everything.
7369
26d75703
RB		 73702013-08-30 Richard Biener <rguenther@suse.de>
7371
7372 * fold-const.c (fold_single_bit_test): Fix overflow test.
7373
f7e088e7
EB		 73742013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7375
7376 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7377 and which can live in a register, always retrieve the value on entry.
7378 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7379 passed by invisible reference specially.
7380 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7381 (vt_add_function_parameter): Correctly deal with a parameter passed by
7382 invisible reference.
7383
73842013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 7385
7386 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7387
6e8dad05
RB		 73882013-08-30 Richard Biener <rguenther@suse.de>
7389
7390 PR tree-optimization/58228
7391 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7392 allow invariant loads in nested loop vectorization.
7393
062ef2c8
RB		 73942013-08-30 Richard Biener <rguenther@suse.de>
7395
7396 PR tree-optimization/58223
7397 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7398 (has_anti_or_output_dependence): ... this and adjust to also
7399 look for output dependences.
7400 (mark_nodes_having_upstream_mem_writes): Adjust.
7401 (rdg_flag_uses): Likewise.
7402
7a764c60
RB		 74032013-08-30 Richard Biener <rguenther@suse.de>
7404
7405 PR tree-optimization/58010
7406 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7407 assert that we have a loop-closed PHI.
7408
82e9d642
JH		 74092013-08-29 Jan Hubicka <jh@suse.cz>
7410
7411 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7412 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7413 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7414 (lto_free_function_in_decl_state_for_node): New function.
7415
9cc1fb4b
XDL		 74162013-08-29 Xinliang David Li <davidxl@google.com>
7417
41dedebd 7418 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 7419 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7420 Emit alignment peeling message with default -fopt-info.
7421 (vect_loop_versioning): Emit loop version info message.
41dedebd 7422 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 7423 (execute_vect_slp): Ditto.
7424
440a5082
EB		 74252013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7426
7427 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7428 of the clone from the DECL_NAME of the original function.
7429
3fa3690d
OE		 74302013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7431
7432 * passes.c (register_pass): Add overload.
7433 * tree-pass.h (register_pass): Forward declare it. Add comment.
7434
0170f33c
JH		 74352013-08-29 Jan Hubicka <jh@suse.cz>
7436
41dedebd
UB		 7437 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7438 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7439 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7440 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7441 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7442 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 7443 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7444 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7445
103ff0d6
TJ		 74462013-08-29 Teresa Johnson <tejohnson@google.com>
7447
7448 * dumpfile.c (dump_loc): Output column number.
7449 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7450 * doc/invoke.texi: Document optall -fopt-info flag.
7451 * profile.c (read_profile_edge_counts): Use new dump framework.
7452 (compute_branch_probabilities): Ditto.
7453 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7454 when pass not in any opt group.
7455 * pass_manager.h (pass_manager::get_pass_profile): New method.
7456 * value-prof.c (check_counter): Use new dump framework.
7457 (check_ic_target): Ditto.
7458 * coverage.c (get_coverage_counts): Ditto.
7459 (coverage_init): Setup new dump framework.
7460
301bbc16
RB		 74612013-08-29 Richard Biener <rguenther@suse.de>
7462
7463 PR tree-optimization/58246
7464 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7465 handle the dominance check inside a basic-block.
7466
1b275000
RB		 74672013-08-29 Richard Biener <rguenther@suse.de>
7468
7469 PR middle-end/57287
7470 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7471 of default defs that appear in abnormal PHI nodes.
7472
6b1184ba
RB		 74732013-08-29 Richard Biener <rguenther@suse.de>
7474
7475 PR tree-optimization/57685
7476 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7477 single-use operands to avoid exponential complexity.
7478
1b08b734
DC		 74792013-08-28 Dehao Chen <dehao@google.com>
7480
7481 * ipa-inline.c (edge_badness): Fix integer underflow.
7482
48a3fa69
UB		 74832013-08-28 Uros Bizjak <ubizjak@gmail.com>
7484
7485 * gtm-builtins.def (_ITM_free): Declare leaf.
7486
215f73e6
JJ		 74872013-08-28 Jakub Jelinek <jakub@redhat.com>
7488
7489 PR target/58067
7490 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7491 (*tls_local_dynamic_base_64_largepic): Likewise.
7492 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7493 Remove predicate from call operand.
7494 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7495 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7496
34554d1a
JL		 74972013-08-28 Jeff Law <law@redhat.com>
7498
7499 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7500 checks for the number of predecessors and successors allowed.
7501 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7502 which require copying a joiner block if there is a request which
7503 is a subpath that requires no joiner block copying.
7504
c01c111b
JH		 75052013-08-28 Jan Hubicka <jh@suse.cz>
7506
7507 * lto-streamer-out.c (DFS_write_tree_body): Drop
7508 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7509 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7510 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7511 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7512 Do not read DECL_ERROR_ISSUED.
7513 (unpack_ts_decl_with_vis_value_fields): Do not read
7514 DECL_DEFER_OUTPUT.
48a3fa69
UB		 7515 (lto_input_ts_binfo_tree_pointers): Do not read
7516 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 7517 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7518 write DECL_ERROR_ISSUED..
7519 (pack_ts_decl_with_vis_value_fields): Do not write
7520 DECL_DEFER_OUTPUT.
7521 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 7522 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 7523 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7524 * tree.h (tree_decl_common): Update comment.
7525 (DECL_ERROR_ISSUED): Remove.
7526
4b128ece
JJ		 75272013-08-28 Jakub Jelinek <jakub@redhat.com>
7528
7529 PR middle-end/58257
7530 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7531
12211b99 75322013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 7533
7534 * builtins.def (free): Declare leaf.
7535
8dce4dbc
DM		 75362013-08-27 David Malcolm <dmalcolm@redhat.com>
7537
7538 * gdbhooks.py: New.
7539 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7540 * configure: Regenerate.
7541
b8f6e610
MJ		 75422013-08-27 Martin Jambor <mjambor@suse.cz>
7543
7544 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7545 (ipa_ancestor_jf_data): Likewise.
7546 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7547 (ipa_get_jf_pass_through_type_preserved): New function.
7548 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7549 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 7550 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 7551 (ipa_get_jf_ancestor_result): Likewise.
7552 (propagate_vals_accross_pass_through): Use
7553 ipa_get_jf_pass_through_result to do all the value mappings.
7554 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7555 type_preserved flag.
7556 (ipa_set_jf_cst_copy): New function.
7557 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7558 (ipa_set_jf_arith_pass_through): Likewise.
7559 (ipa_set_ancestor_jf): Likewise.
7560 (compute_complex_assign_jump_func): Set type_preserved instead of
7561 punting.
7562 (ipa_compute_jump_functions_for_edge): Likewise.
7563 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7564 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 7565 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 7566 (ipa_write_jump_function): Stream the type_preserved flags.
7567 (ipa_read_jump_function): Likewise.
7568
74bf76ed
JJ		 75692013-08-27 Jakub Jelinek <jakub@redhat.com>
7570 Aldy Hernandez <aldyh@redhat.com>
7571
7572 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7573 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7574 * function.h (struct function): Add has_force_vect_loops and
7575 has_simduid_loops.
12211b99 7576 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 7577 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7578 handle it.
7579 * gimple.def: Update CLAUSES comments.
7580 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7581 (gimple_build_omp_for): Add argument to prototype.
7582 (gimple_omp_for_kind): New.
7583 (gimple_omp_for_set_kind): New.
7584 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7585 GOVD_DATA_SHARE_CLASS.
7586 (enum omp_region_type): Add ORT_SIMD.
7587 (gimple_add_tmp_var): Handle ORT_SIMD.
7588 (gimplify_var_or_parm_decl): Same.
7589 (is_gimple_stmt): Same.
7590 (omp_firstprivatize_variable): Same.
7591 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7592 (omp_notice_variable): Handle ORT_SIMD.
12211b99 7593 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 7594 (omp_check_private): Handle ORT_SIMD.
7595 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7596 OMP_CLAUSE_SAFELEN.
7597 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7598 Handle OMP_CLAUSE_LASTPRIVATE.
7599 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7600 OMP_CLAUSE_SAFELEN.
7601 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7602 (gimplify_expr): Handle OMP_SIMD.
7603 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7604 (expand_GOMP_SIMD_VF): New.
7605 (expand_GOMP_SIMD_LAST_LANE): New.
7606 * internal-fn.def (GOMP_SIMD_LANE): New.
7607 (GOMP_SIMD_VF): New.
7608 (GOMP_SIMD_LAST_LANE): New.
7609 * omp-low.c: Include target.h.
7610 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7611 OMP_CLAUSE_SAFELEN.
7612 (check_omp_nesting_restrictions): Same.
7613 (omp_max_vf): New.
7614 (lower_rec_simd_input_clauses): New.
7615 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7616 OMP_CLAUSE_LINEAR.
7617 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7618 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7619 (expand_omp_build_assign): New.
7620 (expand_omp_for_init_counts): New.
7621 (expand_omp_for_init_vars): New.
7622 (extract_omp_for_update_vars): New.
7623 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7624 and rewrite accordingly.
7625 (expand_omp_simd): New.
7626 (expand_omp_for): Use expand_omp_simd.
7627 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7628 (lower_omp_for): Do not lower the body.
7629 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7630 in their own loops.
7631 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 7632 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7633 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 7634 (gate_tree_if_conversion): Similarly.
7635 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7636 gimple_build_omp_for.
7637 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7638 * tree-parloops (create_parallel_loop): Pass kind argument to
7639 gimple_build_omp_for.
7640 * tree-pretty-print.c (dump_omp_clause): Add cases for
7641 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7642 OMP_CLAUSE__SIMDUID_.
7643 (dump_generic_node): Handle OMP_SIMD.
7644 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7645 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7646 unroll OMP_SIMD loops here.
12211b99 7647 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 7648 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 7649 loop->safelen.
74bf76ed
JJ		 7650 (vect_analyze_data_refs): Handle simd loops.
7651 * tree-vect-loop.c (vectorizable_live_operation): Handle
7652 IFN_GOMP_SIMD*.
12211b99 7653 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 7654 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7655 (vectorizable_load): Same.
12211b99 7656 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 7657 (struct simduid_to_vf): New.
7658 (simduid_to_vf::hash): New.
7659 (simduid_to-vf::equal): New.
7660 (struct simd_array_to_simduid): New.
7661 (simd_array_to_simduid::hash): New.
7662 (simd_array_to_simduid::equal): New.
7663 (adjust_simduid_builtins): New.
7664 (struct note_simd_array_uses_struct): New.
7665 (note_simd_array_uses_cb): New.
7666 (note_simd_array_uses): New.
7667 (vectorize_loops): Handle simd hints and adjust simd builtins
7668 accordingly.
7669 * tree-vectorizer.h (struct _stmt_vec_info): Add
7670 simd_lane_access_p field.
7671 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7672 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7673 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7674 (omp_clause_code_name): Same.
7675 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7676 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7677 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 7678 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7679 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 7680 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7681 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7682 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7683 (OMP_CLAUSE_LINEAR_STEP): New.
7684 (OMP_CLAUSE_SAFELEN_EXPR): New.
7685 (OMP_CLAUSE__SIMDUID__DECL): New.
7686 (find_omp_clause): New prototype.
7687
d0cf4e84
L		 76882013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7689
7690 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7691 Haswell processor detection.
7692
992592ec
CW		 76932013-08-27 Christian Widmer <shadow@umbrox.de>
7694
7695 PR target/57927
7696 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7697 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7698 AVX2 capable processors.
7699
9116eb22
TB		 77002013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7701
7702 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7703 of vget_low_* with implementations in terms of other intrinsics.
7704
1c5abb53
MG		 77052013-08-27 Marc Glisse <marc.glisse@inria.fr>
7706
7707 PR middle-end/57219
7708 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7709 values to -1, 0 and 1.
7710
2e100703
VP		 77112013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7712
7713 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7714 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7715 (<optab><mode>3_insn): Remove.
7716 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7717 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7718 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7719 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7720 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7721 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7722 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7723 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7724 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7725 (ror<mode>3_insn): Likewise.
7726 * config/aarch64/predicates.md (aarch64_simd_register): New.
7727
4ded8276
RB		 77282013-08-27 Richard Biener <rguenther@suse.de>
7729
7730 PR tree-optimization/57521
7731 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7732 one edge is non-critical.
7733 (find_phi_replacement_condition): Make sure to use a non-critical
7734 edge. Cleanup and remove old bug workarounds.
7735 (bb_postdominates_preds): Remove.
7736 (if_convertible_loop_p_1): Do not compute post-dominators.
7737 (combine_blocks): Do not free post-dominators.
7738 (main_tree_if_conversion): Likewise.
7739 (pass_data_if_conversion): Add TODO_verify_ssa.
7740
5aa11061
DD		 77412013-08-27 DJ Delorie <dj@redhat.com>
7742
7743 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7744
fd91cfe3
YZ		 77452013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7746
7747 * function.c (assign_parm_find_data_types): Set passed_mode and
7748 nominal_mode to the TYPE_MODE of nominal_type for the built
7749 pointer type in case of the struct-pass-by-reference.
7750
907555ce
JR		 77512013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7752
7753 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7754 (UINT16_TYPE): Change default to "unsigned int".
7755
67518c93
JR		 7756 * config/avr/avr.opt (mfract-convert-truncate): New option.
7757 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7758 is set, round negative fractional integers according to n1169
7759 when converting to integer types.
7760
d6d989fc
JH		 77612013-08-26 Jan Hubicka <jh@suse.cz>
7762
7763 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7764 methods can not be called indirectly when their address is not taken.
7765
2aa3da06
JH		 77662013-08-26 Jan Hubicka <jh@suse.cz>
7767
12211b99
UB		 7768 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7769 ctor_for_folding.
2aa3da06 7770
0987ffe7
JH		 77712013-08-26 Jan Hubicka <jh@suse.cz>
7772
7773 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7774 can be unshared.
7775
befe8647
JR		 77762013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7777
7778 * reload.c (find_valid_class): Allow classes that do not include
7779 FIRST_PSEUDO_REGISTER - 1.
7780
a21e735e
JH		 77812013-08-26 Jan Hubicka <jh@suse.cz>
7782
7783 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7784 fix edge count/frequency when speculation failed; fix type check
7785 for the direct call.
7786
e067bd43
JH		 77872013-08-26 Jan Hubicka <jh@suse.cz>
7788
7789 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7790
e8aec975
JH		 77912013-08-26 Jan Hubicka <jh@suse.cz>
7792
7793 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7794 to be changed by edge redirection.
7795
d0b66480
JH		 77962013-08-26 Jan Hubicka <jh@suse.cz>
7797
12211b99
UB		 7798 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7799 formating; add sanity check.
d0b66480
JH		 7800 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7801 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7802 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7803 (resolve_noninline_speculation): Update callee keys, too.
7804
0f9aaac7
JH		 78052013-08-26 Jan Hubicka <jh@suse.cz>
7806
7807 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7808 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7809
4b37444e
JR		 78102013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7811
7812 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7813 into proper place.
7814
5339fc0c
UB		 78152013-08-26 Uros Bizjak <ubizjak@gmail.com>
7816
7817 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7818 (x86_64_elf_select_section): Ditto.
7819 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7820 arguments.
7821 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 7822 (output_set_got): Ditto.
7823 (ix86_unary_operator_ok): Ditto.
7824 (ix86_expand_builtin): Ditto.
7825
d5c3d3ef
JH		 78262013-08-23 Jan Hubicka <jh@suse.cz>
7827
5339fc0c 7828 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 7829
5b1e7435
JH		 78302013-08-23 Jan Hubicka <jh@suse.cz>
7831
7832 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7833 (tree_decl_with_vis): Add FINAL field.
7834
ead69dac
JL		 78352013-08-23 Jeff Law <law@redhat.com>
7836
7837 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7838 the debugging dump when the expression is fully redundant.
7839
025311c4
GDR		 78402013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7841
7842 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7843 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7844 * pretty-print.c (pp_formatted_text_data): Likewise.
7845 (pp_write_text_to_stream): Likewise.
7846 (pp_write_text_as_dot_label_to_stream): Likewise.
7847 (pp_append_r): Likewise.
7848 (pp_format): Likewise.
7849 (pp_flush): Likewise.
7850 (pp_clear_output_area): Likewise.
7851 (pp_append_text): Likewise.
7852 (pp_formatted_text): Likewise.
7853 (pp_remaining_character_count_for_line): Likewise.
7854 (pp_newline): Likewise.
7855 (pp_character): Likewise.
7856 (output_buffer::~output_buffer): Define.
7857 (pretty_printer::~pretty_printer): Destruct output buffer.
7858 * pretty-print.h (output_buffer::~output_buffer): Declare.
7859 (pretty_printer::~pretty_printer): Declare virtual.
7860
0cadbfaa
MG		 78612013-08-24 Marc Glisse <marc.glisse@inria.fr>
7862
7863 PR other/57324
7864 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7865 HOST_WIDE_INT_M1U): New macros.
7866 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7867 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7868 unsigned -1 for lshift.
7869 * cse.c (cse_insn): Likewise.
7870 * double-int.c (rshift_double, lshift_double): Likewise.
7871 * builtins.c (fold_builtin_bitop): Likewise.
7872 * combine.c (force_to_mode): Likewise.
7873 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7874 * simplify-rtx.c (simplify_const_unary_operation,
7875 simplify_const_binary_operation): Likewise.
7876 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7877 check_va_list_escapes): Likewise.
7878 * rtlanal.c (nonzero_bits1): Likewise.
7879 * expmed.c (expand_smod_pow2): Likewise.
7880 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7881
009e5353
JH		 78822013-08-23 Jan Hubicka <jh@suse.cz>
7883
7884 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7885 as having address taken.
7886
c4be6568
JH		 78872013-08-23 Jan Hubicka <jh@suse.cz>
7888
64cbf23d
JH		 7889 * ipa-utils.h (method_class_type): Declare.
7890 * ipa-devirt.c (method_class_type): Export.
7891
c4be6568
JH		 7892 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7893 do not walk base classes of anonymous types.
7894
97aba8e9
KK		 78952013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7896
7897 PR rtl-optimization/58220
7898 PR regression/58221
7899 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7900 handle SEQUENCE insns properly.
7901
0fc80001
GDR		 79022013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7903
7904 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7905 definition.
7906 (pp_newline_and_indent): Likewise.
7907 (pp_separate_with): Likewise.
7908 * pretty-print.c (pp_newline_and_flush): Define.
7909 (pp_newline_and_indent): Likewise.
7910 (pp_separate_with): Likewise.
7911
520a5868
JJ		 79122013-08-23 Jakub Jelinek <jakub@redhat.com>
7913
7914 PR target/58218
7915 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7916 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7917
e500c62a
KY		 79182013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7919
f43245d1
UB		 7920 * config/i386/predicates.md (ext_sse_reg_operand): New.
7921 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY		 7922 predicate to determine if EVEX is needed.
7923 (*movsi_internal): Ditto.
7924 (*movdf_internal): Ditto.
7925 (*movsf_internal): Ditto.
f43245d1 7926 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 7927
6b00d7dd
JJ		 79282013-08-23 Jakub Jelinek <jakub@redhat.com>
7929
7930 PR tree-optimization/58209
7931 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
7932 (find_tail_calls): Give up for pointer result types if m is non-NULL.
7933 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
7934 emit POINTER_PLUS_EXPR.
7935 (create_tailcall_accumulator): For pointer result type accumulate in
7936 sizetype type.
7937
4f2a9f90
PC		 79382013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
7939
7940 * configure.ac: Add backslashes missing from the last change.
7941 * configure: Regenerate.
7942
7bf4274e
JH		 79432013-08-22 Jan Hubicka <jh@susue.cz>
7944
4f2a9f90
PC		 7945 * ipa.c (function_and_variable_visibility): First remember function
7946 was global and then make it local.
7bf4274e 7947
1bea243d
JB		 79482013-08-22 Julian Brown <julian@codesourcery.com>
7949
7950 * configure.ac: Add aarch64 to list of arches which use "nop" in
7951 debug_line test.
7952 * configure: Regenerate.
7953
bcba7ecc
AK		 79542013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7955
7956 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
7957 gnu_libc_has_function.
7958 * config/s390/tpf.h: Likewise.
7959
4c495b0d
JH		 79602013-08-22 Jan Hubicka <jh@susue.cz>
7961
7962 * timevar.c (validate_phases): Add cast.
7963
5e302bca
JH		 79642013-08-22 Jan Hubicka <jh@susue.cz>
7965
7966 * timevar.c (validate_phases): Use size_t for memory.
7967 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
7968
da6ca2b5
GDR		 79692013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
7970
7971 * pretty-print.h (output_buffer::output_buffer): Declare.
7972 (pretty_printer::pretty_printer): Likewise.
7973 (pp_construct): Remove.
7974 * pretty-print.c (output_buffer::output_buffer): Define.
7975 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
7976 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
7977 (print_gimple_expr): Likewise.
7978 (print_gimple_seq): Likewise.
7979 (gimple_dump_bb): Likewise.
7980 * sched-vis.c (dump_value_slim): Likewise.
7981 (dump_insn_slim): Likewise.
7982 (dump_rtl_slim): Likewise.
7983 (str_pattern_slim): Likewise.
7984 * tree-mudflap.c (mf_varname_tree): Likewise.
7985 * graph.c (print_graph_cfg): Likewise.
7986 (start_graph_dump): Likewise.
7987 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
7988 placement-new.
7989 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
7990 pretty printer initialization.
7991 * coretypes.h (diagnostic_context): Remove superflous type alias
7992 declaration.
7993 (pretty_printer): Likewise. Declare directly as a class.
7994 (pretty_print_info): Remove declaration as class.
7995 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
7996 and pp_clear_output_area.
7997 (asan_add_global): Likewise.
7998
0e1474e5
JH		 79992013-08-22 Jan Hubicka <jh@suse.cz>
8000
8001 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8002 * ipa-utils.h (update_type_inheritance_graph): Declare.
8003 (possible_polymorphic_call_target_p): Declare.
8004 (possible_polymorphic_call_target_p): New.
8005 * ipa-devirt.c: Update toplevel comments.
8006 (cached_polymorphic_call_targets): Move up.
8007 (odr_type_d): Move ID down.
8008 (polymorphic_type_binfo_p): Update comment.
8009 (odr_hasher::remove): Likewise;
8010 (get_odr_type): Set anonymous_namespace.
8011 (dump_odr_type): Dump it.
8012 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8013 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8014 (record_binfo): Add comment.
5339fc0c
UB		 8015 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8016 built.
0e1474e5
JH		 8017 (devirt_node_removal_hook): Do not iCE when cache is freed.
8018 (possible_polymorphic_call_target_p): New predicate.
8019 (update_type_inheritance_graph): New function.
8020
3f97cb0b
AI		 80212013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8022 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8023 Sergey Lega <sergey.s.lega@intel.com>
8024 Anna Tikhonova <anna.tikhonova@intel.com>
8025 Ilya Tocar <ilya.tocar@intel.com>
8026 Andrey Turetskiy <andrey.turetskiy@intel.com>
8027 Ilya Verbin <ilya.verbin@intel.com>
8028 Kirill Yukhin <kirill.yukhin@intel.com>
8029 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8030
8031 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8032 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8033 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8034 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8035 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8036 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8037 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8038 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8039 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8040 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8041 OPT_mavx512pf, OPT_mavx512er cases.
8042 * config/i386/constraints.md (v): New constraint.
8043 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8044 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8045 (bit_AVX512CD): New.
8046 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8047 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8048 * config/i386/i386-c.c (ix86_target_macros_internal):
8049 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8050 __AVX512PF__.
8051 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8052 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8053 * config/i386/i386.c (regclass_map, dbx_register_map)
8054 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8055 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8056 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8057 -mavx512pf options.
8058 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8059 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8060 -mavx512pf options. Fix formatting.
8061 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8062 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8063 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8064 -mavx512cd, -mavx512pf options.
8065 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8066 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8067 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8068 ALL_SSE_REGS.
8069 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8070 (ix86_set_reg_reg_cost): Ditto.
8071 (x86_order_regs_for_local_alloc): Ditto.
8072 (MAX_VECT_LEN): Extend to 64-byte.
8073 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8074 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8075 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8076 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8077 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8078 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8079 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8080 (SSE_REG_MODE_P): Support new modes.
8081 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8082 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8083 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8084 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8085 (REG_CLASS_CONTENTS): Add new registers.
8086 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8087 (EXT_REX_SSE_REGNO_P): New.
8088 (HI_REGISTER_NAMES): Add new registers.
8089 * config/i386/i386.md: Define constants for new registers.
8090 (mode): Add new 512-bit modes.
8091 (prefix): Support evex prefix.
8092 (isa): Support avx512f, noavx512f, fma_avx512f.
8093 (ssemodesuffix): Add new 512-bit modes.
8094 (movxi): New.
8095 (*movxi_internal_avx512f): Ditto.
8096 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8097 Support MODE_XI.
8098 (*movsi_internal): Likewise.
8099 (*movdf_internal): Likewise.
8100 (*movsf_internal): Likewise.
8101 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8102 (<code><mode>3): Likewise.
5339fc0c
UB		 8103 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8104 New.
3f97cb0b
AI		 8105 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8106 with the new constraint "v".
8107 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8108 modes.
8109 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8110 with the new constraint "v".
8111 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8112 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8113 (*<plusminus_insn><mode>3): Likewise.
8114 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8115 (*mul<mode>3): Likewise.
8116 (<sse>_vmmul<mode>3): Likewise.
8117 (<sse>_div<mode>3): Likewise.
8118 (<sse>_vmdiv<mode>3): Likewise.
8119 (<sse>_sqrt<mode>2): Likewise.
8120 (<sse>_vmsqrt<mode>2): Likewise.
8121 (*<code><mode>3_finite): Likewise.
8122 (*<code><mode>3) <smaxmin>: Likewise.
8123 (<sse>_vm<code><mode>3): Likewise.
8124 (*<code><mode>3) <any_logic>: Likewise.
8125 (*fma_fmadd_<mode>): Likewise.
8126 (*fma_fmsub_<mode>): Likewise.
8127 (*fma_fnmadd_<mode>): Likewise.
8128 (*fma_fnmsub_<mode>): Likewise.
8129 (*fma_fmaddsub_<mode>): Likewise.
8130 (*fma_fmsubadd_<mode>): Likewise.
8131 (*fmai_fmadd_<mode>): Likewise.
8132 (*fmai_fmsub_<mode>): Likewise.
8133 (*fmai_fnmadd_<mode>): Likewise.
8134 (*fmai_fnmsub_<mode>): Likewise.
8135 (sse_cvtsi2ss): Likewise.
8136 (sse_cvtsi2ssq): Likewise.
8137 (sse_cvtss2si): Likewise.
8138 (sse_cvtss2si_2): Likewise.
8139 (sse_cvtss2siq): Likewise.
8140 (sse_cvtss2siq_2): Likewise.
8141 (sse_cvttss2si): Likewise.
8142 (sse_cvtss2siq_2): Likewise.
8143 (float<sseintvecmodelower><mode>2): Likewise.
8144 (sse2_cvtsd2si_2): Likewise.
8145 (sse2_cvtsd2siq_2): Likewise.
8146 (*<plusminus_insn><mode>3): Likewise.
8147 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8148 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8149 (ashr<mode>3): Likewise.
8150 (<shift_insn><mode>3): Likewise.
8151 (avx2_<code><mode>3): Likewise.
8152 (*avx2_<code><mode>3): Likewise.
8153 (*andnot<mode>3): Likewise.
8154 (*<code><mode>3) <any_logic>: Likewise.
8155 (abs<mode>2): Likewise.
8156 (avx2_permvar<mode>): Likewise.
8157 (avx2_perm<mode>_1): Likewise.
8158 (*avx_vpermilp<mode>): Likewise.
8159 (avx_vpermilvar<mode>3): Likewise.
8160 (avx2_ashrv<mode>): Likewise.
8161 (avx2_<shift_insn>v<mode>): Likewise.
8162 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8163 -mavx512cd.
8164 * doc/rtl.texi: Document XImode.
8165
dd5e8423
JL		 81662013-08-21 Jeff Law <law@redhat.com>
8167
b9ebee5d
JL		 8168 * tree-flow.h (register_jump_thread): Pass vector of edges
8169 instead of each important edge.
8170 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8171 thread path into a vector and pass that to register_jump_thread.
8172 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8173 passed in edge vector to the current 3-edge form.
8174
dd5e8423
JL		 8175 Revert:
8176 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8177
9bb6628e 8178 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 8179 delete_unmarked_insns removed anything.
8180
445dc8df
JR		 81812013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8182
8183 * reload.h (struct reg_equivs): Rename to ..
8184 (struct reg_equivs_s): .. this.
8185
e0df53dd
ML		 81862013-08-20 Martin Liska <marxin.liska@gmail.com>
8187
8188 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8189
e7d1d3eb
RO		 81902013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8191
5339fc0c 8192 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8193
5562e26e
JL		 81942013-08-21 Jeff Law <law@redhat.com>
8195
8196 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8197 simplify assignments too. If the RHS collapses to a singleton
8198 range, then return the value for the range.
8199
c7ecdec6
KY		 82002013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8201
7a5911d3
UB		 8202 * config/i386/sse.md (V16): Rename to...
8203 (VMOVE): this.
8204 (mov<mode>): Update iterator name.
8205 (*mov<mode>_internal): Ditto.
8206 (push<mode>1): Ditto.
8207 (movmisalign<mode>): Ditto.
c7ecdec6 8208
bfa3b50a
JH		 82092013-08-20 Jan Hubicka <jh@suse.cz>
8210
8211 PR bootstrap/58186
8212 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8213 entry for direct edges.
8214 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8215
6cd4d135
DM		 82162013-08-20 David Malcolm <dmalcolm@redhat.com>
8217
8218 Revert my last two changes, r201865 and r201864:
8219
8220 Revert r201865:
8221 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8222
8223 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8224 instances can own GC refs.
8225
8226 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8227 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8228 (gcc::context::gt_pch_nx): Likewise.
8229 (gcc::context::gt_pch_nx): Likewise.
8230 * ggc.h (gt_ggc_mx <T>): New.
8231 (gt_pch_nx_with_op <T>): New.
8232 (gt_pch_nx <T>): New.
8233 * passes.c (opt_pass::gt_ggc_mx): New.
8234 (opt_pass::gt_pch_nx): New.
8235 (opt_pass::gt_pch_nx_with_op): New.
8236 (pass_manager::gt_ggc_mx): New.
8237 (pass_manager::gt_pch_nx): New.
8238 (pass_manager::gt_pch_nx_with_op): New.
8239 (pass_manager::operator new): Use
8240 ggc_internal_cleared_alloc_stat rather than xcalloc.
8241 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8242 (pass_manager::gt_ggc_mx): New.
8243 (pass_manager::gt_pch_nx): New.
8244 (pass_manager::gt_pch_nx_with_op): New.
8245 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8246 (opt_pass::operator new): New.
8247 (opt_pass::gt_ggc_mx): New.
8248 (opt_pass::gt_pch_nx): New.
8249 (opt_pass::gt_pch_nx_with_op): New.
8250
8251 Revert r201864:
8252 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8253
8254 * Makefile.in (GTFILES): Add context.h.
8255 * context.c (gcc::context::operator new): New.
8256 (gcc::context::gt_ggc_mx): New.
8257 (gcc::context::gt_pch_nx): New.
8258 (gcc::context::gt_pch_nx): New.
8259 * context.h (gcc::context): Add GTY((user)) marking.
8260 (gcc::context::operator new): New.
8261 (gcc::context::gt_ggc_mx): New.
8262 (gcc::context::gt_pch_nx): New.
8263 (gcc::context::gt_pch_nx): New.
8264 (g): Add GTY marking.
8265 (gt_ggc_mx (gcc::context *)): New.
8266 (gt_pch_nx (gcc::context *)): New.
8267 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8268 void *cookie)): New.
8269 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8270
a046136a
AM		 82712013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8272
8273 * dce.c (fini_dce): Call df_analyze again just in case
8274 delete_unmarked_insns removed anything.
8275
67598720
TJ		 82762013-08-20 Teresa Johnson <tejohnson@google.com>
8277
8278 PR rtl-optimizations/57451
8279 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8280 from crossing split section boundaries.
8281
f8693faf
MGD		 82822013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8283
8284 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8285 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8286 with MULTLIB_DEFAULTS.
8287
7d0b9a9c
NC		 82882013-08-20 Nick Clifton <nickc@redhat.com>
8289
8290 * target.def (narrow_volatile_bitfield): Note that the default
8291 value is false, not !TARGET_STRICT_ALIGN.
8292 * doc/tm.texi: Regenerate.
8293
8edb8dc8
PC		 82942013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8295
8296 Fix LIB_SPEC for systems without libpthread.
8297
8298 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8299 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8300 for Android.
8301 * config/i386/linux-common.h: Likewise.
8302 * config/mips/linux-common.h: Likewise.
8303
12211b99 83042013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 8305
8306 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8307 checks.
8308
e42c64cb
DM		 83092013-08-20 David Malcolm <dmalcolm@redhat.com>
8310
8311 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8312 instances can own GC refs.
8313
8314 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8315 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8316 (gcc::context::gt_pch_nx): Likewise.
8317 (gcc::context::gt_pch_nx): Likewise.
8318 * ggc.h (gt_ggc_mx <T>): New.
8319 (gt_pch_nx_with_op <T>): New.
8320 (gt_pch_nx <T>): New.
8321 * passes.c (opt_pass::gt_ggc_mx): New.
8322 (opt_pass::gt_pch_nx): New.
8323 (opt_pass::gt_pch_nx_with_op): New.
8324 (pass_manager::gt_ggc_mx): New.
8325 (pass_manager::gt_pch_nx): New.
8326 (pass_manager::gt_pch_nx_with_op): New.
8327 (pass_manager::operator new): Use
8328 ggc_internal_cleared_alloc_stat rather than xcalloc.
8329 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8330 (pass_manager::gt_ggc_mx): New.
8331 (pass_manager::gt_pch_nx): New.
8332 (pass_manager::gt_pch_nx_with_op): New.
8333 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8334 (opt_pass::operator new): New.
8335 (opt_pass::gt_ggc_mx): New.
8336 (opt_pass::gt_pch_nx): New.
8337 (opt_pass::gt_pch_nx_with_op): New.
8338
910c02a0
DM		 83392013-08-20 David Malcolm <dmalcolm@redhat.com>
8340
8341 * Makefile.in (GTFILES): Add context.h.
8342 * context.c (gcc::context::operator new): New.
8343 (gcc::context::gt_ggc_mx): New.
8344 (gcc::context::gt_pch_nx): New.
8345 (gcc::context::gt_pch_nx): New.
8346 * context.h (gcc::context): Add GTY((user)) marking.
8347 (gcc::context::operator new): New.
8348 (gcc::context::gt_ggc_mx): New.
8349 (gcc::context::gt_pch_nx): New.
8350 (gcc::context::gt_pch_nx): New.
8351 (g): Add GTY marking.
8352 (gt_ggc_mx (gcc::context *)): New.
8353 (gt_pch_nx (gcc::context *)): New.
8354 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8355 void *cookie)): New.
8356 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8357
af4b4236
AM		 83582013-08-20 Alan Modra <amodra@gmail.com>
8359
8360 PR target/57865
8361 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8362 (rs6000_emit_epilogue): Likewise.
8363
12211b99 83642013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 8365
8366 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8367
e2323f5b
PB		 83682013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8369 Jakub Jelinek <jakub@redhat.com>
8370
8371 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8372 (BUILT_IN_FABSD64): Likewise.
8373 (BUILT_IN_FABSD128): Likewise.
8374 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8375 (fold_builtin_1): Likewise.
5339fc0c
UB		 8376 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8377 destination and source operands.
e2323f5b
PB		 8378 (*abstd2_fpr): Likewise.
8379 (*nabstd2_fpr): Likewise.
8380
cfbf3ee8
RS		 83812013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8382
8383 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8384 JUMP_P and INSN_P.
8385
d3136aeb
AH		 83862013-08-19 Aldy Hernandez <aldyh@redhat.com>
8387
8388 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8389 incomplete.
8390
d33d9e47
AI		 83912013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8392
8393 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8394 * builtins.c (default_libc_has_function): New.
8395 (gnu_libc_has_function): Ditto.
8396 (no_c99_libc_has_function): Ditto.
8397 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8398 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8399 (fold_builtin_sincos): Likewise.
8400 (fold_builtin_cexp): Likewise.
8401 * builtins.def (DEF_C94_BUILTIN): Likewise.
8402 (DEF_C99_BUILTIN): Likewise.
8403 (DEF_C99_C90RES_BUILTIN): Likewise.
8404 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8405 definitions to using this define.
8406 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 8407 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 8408 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8409 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8410 * config/darwin.h: Ditto.
8411 * config/elfos.h: Ditto.
8412 * config/freebsd.h: Ditto.
8413 * config/i386/cygming.h: Ditto.
8414 * config/i386/djgpp.h: Ditto.
8415 * config/i386/i386-interix.h: Ditto.
8416 * config/microblaze/microblaze.h: Ditto.
8417 * config/mmix/mmix.h: Ditto.
8418 * config/gnu-user.h: Ditto.
8419 * config/ia64/hpux.h: Ditto.
8420 * config/pa/pa-hpux.h: Ditto.
8421 * config/pdp11/pdp11.h: Ditto.
8422 * config/picochip/picochip.h: Ditto.
8423 * config/linux.h: Ditto.
8424 * config/netbsd.h: Ditto.
8425 * config/openbsd.h: Ditto.
8426 * config/rs6000/aix43.h: Ditto.
8427 * config/rs6000/aix51.h: Ditto.
8428 * config/rs6000/aix52.h: Ditto.
8429 * config/rs6000/aix53.h: Ditto.
8430 * config/rs6000/aix61.h: Ditto.
8431 * config/rs6000/darwin.h: Ditto.
8432 * config/rs6000/linux.h: Ditto.
8433 * config/rs6000/linux64.h: Ditto.
8434 * config/s390/tpf.h: Ditto.
8435 * config/sol2-10.h: Ditto.
8436 * config/sol2.h: Ditto.
8437 * config/vms/vms.h: Ditto.
8438 * config/vxworks.h: Ditto.
8439 * config/linux-android.c (linux_android_libc_has_function):
8440 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8441 * config/linux-protos.h (linux_android_libc_has_function):
8442 New declaration.
8443 * config/i386/i386.c (ix86_libc_has_function): New.
8444 * config/i386/i386-protos.h
8445 (ix86_libc_has_function): New declaration.
8446 * config/i386/i386.md
8447 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8448 ("isinf<mode>2): Likewise.
8449 * convert.c (convert_to_integer): Using new target hook
8450 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8451 TARGET_C99_FUNCTIONS.
8452 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8453 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8454 * coretypes.h (function_class): New enum for different
8455 classes of functions.
8456 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8457 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8458 (TARGET_HAS_SINCOS): Likewise.
8459 (TARGET_LIBC_HAS_FUNCTION): New.
8460 * doc/tm.texi: Regenerated.
8461 * targhooks.h (default_libc_has_function): New declaration.
8462 (no_c99_libc_has_function): Ditto.
8463 (gnu_libc_has_function): Ditto.
8464 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8465 and TARGET_HAS_SINCOS.
8466
eefe9a99
JH		 84672013-08-18 Jan Hubicka <jh@suse.cz>
8468
8469 * Makeifle-in (ipa-devirt.o): New.
8470 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8471 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8472 (analyze_functions): Look into possible targets of polymorphic call.
8473 * dumpfile.c (dump_files): Add type-inheritance dump.
8474 * dumpfile.h (TDI_inheritance): New.
8475 * ipa-devirt.c: New file.
8476 * ipa-utils.h (odr_type_d): Forward declare.
8477 (odr_type): New type.
8478 (build_type_inheritance_graph): Declare.
8479 (possible_polymorphic_call_targets): Declare and introduce inline
8480 variant when only edge is pased.
8481 (dump_possible_polymorphic_call_targets): Likewise.
8482 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8483 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8484 (types_same_for_odr): ... here.
8485 * tree.h (type_in_anonymous_namespace_p): Declare.
8486
29b89442
JJ		 84872013-08-18 Jakub Jelinek <jakub@redhat.com>
8488
8489 PR tree-optimization/58006
8490 * tree-parloops.c (take_address_of): Don't ICE if get_name
8491 returns NULL.
8492 (eliminate_local_variables_stmt): Remove clobber stmts.
8493
4f219961
EB		 84942013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8495
8496 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8497 error message is issued for an alias to undefined symbol.
8498
1d5755ef
JH		 84992013-08-18 Jan Hubicka <jh@suse.cz>
8500
8501 * cgraph.c (cgraph_create_indirect_edge): Discover
8502 polymorphic calls and record basic info into indirect_info.
8503 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8504 devirtualization, ignore objc function calls.
8505 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8506 call with no parm index info.
8507 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8508 * tree.c (virtual_method_call_p): New function.
8509 * tree.h (virtual_method_call_p): Declare.
8510
5945bebf
JH		 85112013-08-16 Jan Hubicka <jh@suse.cz>
8512
8513 PR middle-end/58179
8514 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8515
47fa5d23
DE		 85162013-08-16 David Edelsohn <dje.gcc@gmail.com>
8517
8518 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8519 attribute.
8520
c74f54a0
DM		 85212013-08-16 David Malcolm <dmalcolm@redhat.com>
8522
8523 * gengtype.c (type_for_name): Add special-case support for
8524 locating types within the "gcc::" namespace.
8525 (open_base_files): Emit a "using namespace gcc" directive.
8526
f758f299
MM		 85272013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8528
8529 PR target/58160
8530 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8531 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8532
8533 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8534 array instead of each individual operand as a separate argument.
8535 (emit_fusion_gpr_load): Likewise.
8536 (expand_fusion_gpr_load): Add new function declaration.
8537
8538 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8539 signature to have the operands passed as an array, instead of as
8540 separate arguments. Allow ZERO_EXTEND to be in the memory
8541 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8542 depend on the register live/dead flags when peepholes are run.
8543 (expand_fusion_gpr_load): New function to be called from the
8544 peephole2 pass, to change the register that addis sets to be the
8545 target register.
8546 (emit_fusion_gpr_load): Change the calling signature to have the
8547 operands passed as an array, instead of as separate arguments.
8548 Allow ZERO_EXTEND to be in the memory address, and also
8549 SIGN_EXTEND if -mpower8-fusion-sign.
8550
8551 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8552 unspec enumeration.
8553 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8554 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 8555 depend on the register live/dead state when the peephole pass is done.
f758f299 8556
158f4e4f
DM		 85572013-08-16 David Malcolm <dmalcolm@redhat.com>
8558
8559 * gengtype.c (create_user_defined_type): Ensure that the kind
8560 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8561 declaration is seen before the GTY((user)) marking.
8562
22f8cea5
BE		 85632013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8564
8565 PR target/58105
8566 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8567
c49bdb2e
JH		 85682013-08-16 Jan Hubicka <jh@suse.cz>
8569
8570 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8571 arugment expected_type.
8572 (gimple_fold_call): Use it.
8573 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8574 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 8575 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 8576 (try_make_edge_direct_virtual_call): Likewise.
8577 * tree.c (obj_type_ref_class): New.
8578 * tree.h (obj_type_ref_class): Use it.
8579
4042dca9
GDR		 85802013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8581
8582 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8583 (rtl_slim_pp): Likewise.
8584 (init_rtl_slim_pretty_print): Likewise.
8585 (dump_value_slim): Don't call it. Use local pretty printer.
8586 (dump_insn_slim): Likewise.
8587 (dump_rtl_slim): Likewise.
8588 (str_pattern_slim): Likewise.
8589 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8590 Simplify.
8591
7a460594
JJ		 85922013-08-16 Jakub Jelinek <jakub@redhat.com>
8593
639dc669
JJ		 8594 PR tree-optimization/58164
8595 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8596 walk gimple_goto_dest of GIMPLE_GOTO.
8597
7a460594
JJ		 8598 PR tree-optimization/58165
8599 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8600 bi_call must be the last stmt in a bb, don't split_block, instead
8601 use fallthru edge from it and give up if there is none.
8602 Release conds vector when returning early.
8603
e3c02cc7
XDL		 86042013-08-14 Xinliang David Li <davidxl@google.com>
8605
8606 * config/i386/i386.c (ix86_option_override_internal):
8607 Remove unused variable and field.
8608
6469da2c
BS		 86092013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8610
8611 PR target/57949
22f8cea5 8612 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 8613 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8614 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8615 and Linux, correct BLKmode alignment when 128-bit alignment is
8616 required and compatibility flag is not set.
22f8cea5
BE		 8617 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8618 for zero-size arguments when compatibility flag is not set.
6469da2c 8619
4ca890e2
JJ		 86202013-08-14 Jakub Jelinek <jakub@redhat.com>
8621
8622 PR tree-optimization/58145
8623 * tree-sra.c (build_ref_for_offset): If prev_base has
8624 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8625
daec1759 86262013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 8627
daec1759 8628 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 8629 Fix uninitialized variable error.
daec1759 8630
3ad20bd4
XDL		 86312013-08-14 Xinliang David Li <davidxl@google.com>
8632
8633 * config/i386/i386.opt: Define two new options.
8634 * config/i386/x86-tune.def: Add arch selector field in macros.
8635 * config/i386/i386.h: Adjust macro definition.
8636 * config/i386/i386.c (ix86_option_override_internal):
8637 Refactor the code.
8638 (parse_mtune_ctrl_str): New function.
8639 (set_ix86_tune_features): New function.
8640 (ix86_function_specific_restore): Call the new helper function.
8641
e839e2a9
AB		 86422013-08-14 Andrey Belevantsev <abel@ispras.ru>
8643
8644 PR rtl-optimization/57662
8645 * sel-sched.c (code_motion_process_successors): When the current insn
8646 is removed after the recursive traversal, break from the loop.
8647 Add comments and debug printouts.
8648
27ed665f
JJ		 86492013-08-14 Jakub Jelinek <jakub@redhat.com>
8650 Alexandre Oliva <aoliva@redhat.com>
8651
8652 PR target/58067
8653 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8654 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8655 there also UNSPEC_PLTOFF.
8656
61c1a609
MP		 86572013-08-14 Marek Polacek <polacek@redhat.com>
8658
8659 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8660 NUM_CONDITIONS bit positions.
8661
0e901c67
CC		 86622013-08-13 Cary Coutant <ccoutant@google.com>
8663
8664 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8665 (attr_checksum): Hash vector contents instead of pointer.
8666 (attr_checksum_ordered): Likewise.
8667
061eff6d
UB		 86682013-08-13 Uros Bizjak <ubizjak@gmail.com>
8669
8670 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8671 when Pmode != word_mode. Add length_address attribute.
8672 (sse3_monitor_<mode>): Merge from sse3_monitor and
8673 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8674 Pmode != word_mode. Update insn length attribute.
8675 * config/i386/i386.c (ix86_option_override_internal): Update
8676 ix86_gen_monitor selection for merged sse3_monitor insn.
8677
8ec9249b
JB		 86782013-08-13 Julian Brown <julian@codesourcery.com>
8679
8680 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8681 perform invalid legitimization on greater-than-word-size modes for
8682 TARGET_E500_DOUBLE.
8683
761a8eb7
VM		 86842013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8685
8686 * ira.c (setup_class_translate_array): Use aclass instead of cl
8687 for classes not fully covered by allocno classes.
8688
99c2bd54
JJ		 86892013-08-13 Jakub Jelinek <jakub@redhat.com>
8690
4029a5e0
JJ		 8691 PR tree-optimization/57661
8692 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8693 * tree-inline.c (tree_function_versioning): Initialize it.
8694 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8695 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8696 that is not being copied.
8697
99c2bd54
JJ		 8698 PR sanitizer/56417
8699 * asan.c (instrument_strlen_call): Fix typo in comment.
8700 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8701
b5f8f063
SE		 87022013-08-13 Steve Ellcey <sellcey@mips.com>
8703
8704 * config/mips/mips.md (prefetch): Use lw instead of ld on
8705 loongson in 32bit mode.
8706
4de80584
NC		 87072013-08-13 Nick Clifton <nickc@redhat.com>
8708
8709 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8710
09ce3660
JH		 87112013-08-13 Jan Hubicka <jh@suse.cz>
8712
8713 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8714 introduced edge; fix typo in sanity check.
8715 (cgraph_resolve_speculation): Export; improve diagnostic.
8716 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8717 speculation at type mismatch.
8718 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8719 (cgraph_resolve_speculation): Declare.
8720 (symtab_can_be_discarded): New function.
8721 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8722 * ipa-inline-transform.c (speculation_removed): New global var.
8723 (clone_inlined_nodes): See if speculation can be removed.
8724 (inline_call): If speculations was removed, we growths may not match.
8725 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8726 (speculation_useful_p): New function.
8727 (resolve_noninline_speculation): New function.
8728 (inline_small_functions): Resolve useless speculations.
8729 * ipa-inline.h (speculation_useful_p): Declare
8730 * ipa.c (can_replace_by_local_alias): Simplify.
8731 (ipa_profile): Produce speculative calls in non-lto, too;
8732 add simple cost model; produce local aliases.
8733
f971dc24
DM		 87342013-08-13 David Malcolm <dmalcolm@redhat.com>
8735
8736 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 8737 PASS_MANAGER_H.
f971dc24 8738
2e507b9a
PC		 87392013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8740
8741 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8742
c97fd4b8
UB		 87432013-08-12 Uros Bizjak <ubizjak@gmail.com>
8744
8745 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8746 iterator instead of X87MODEF.
8747
87482013-08-12 Perez Read <netfirewall@gmail.com>
8749
8750 PR target/58132
8751 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8752 operand 0 for intel asm alternative.
8753 (*movabs<mode>_2): Ditto for operand 1.
8754
f4f4204c
JG		 87552013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8756
8757 * config/aarch64/arm_none.h
8758 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8759
133b1a8e
NC		 87602013-08-12 Nick Clifton <nickc@redhat.com>
8761
8762 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8763
15dd8b3a
YR		 87642013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8765
8766 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8767 expand for QI/HImode operand to produce more effictive code for
8768 unsigned char(short) --> float(double) conversion.
8769
94109a6a
AM		 87702013-08-12 Alexander Monakov <amonakov@ispras.ru>
8771
8772 * doc/invoke.texi: Mention that -ftls-model does not force the final
8773 model.
8774
4e115102
MP		 87752013-08-12 Marek Polacek <polacek@redhat.com>
8776 Marc Glisse <marc.glisse@inria.fr>
8777
8778 PR tree-optimization/57980
8779 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8780 when creating -1 constant.
8781
748d71f3
JH		 87822013-08-10 Jan Hubicka <jh@suse.cz>
8783
1f97e193 8784 Workaround binutils PR14342.
748d71f3
JH		 8785 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8786 (gimple_init_edge_profiler): Likewise.
8787 (gimple_gen_ic_func_profiler): Likewise.
8788
5979aa54
JH		 87892013-08-09 Jan Hubicka <jh@suse.cz>
8790
8791 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8792
ad83025e
XDL		 87932013-08-09 Xinliang David Li <davidxl@google.com>
8794
8795 * config/i386/stringop.def: New file.
8796 * config/i386/stringop.opt: New file.
8797 * config/i386/i386-opts.h: Include stringopt.def.
8798 * config/i386/i386.opt: Include stringopt.opt.
8799 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 8800 Override default size based stringop inline strategies with options.
ad83025e
XDL		 8801 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8802 New function.
8803
e59df5fd
JH		 88042013-08-09 Jan Hubicka <jh@suse.cz>
8805
c97fd4b8 8806 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 8807
634ab819
JH		 88082013-08-09 Jan Hubicka <jh@suse.cz>
8809
8810 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8811 CGRAPH_FREQ_MAX.
8812 (dump_cgraph_node): Dump profile-id.
8813 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8814 and common_target_probability.
8815 * lto-cgraph.c (lto_output_edge): Stream common targets.
8816 (lto_output_node): Stream profile ids.
8817 (input_node): Stream profile ids.
8818 (input_edge): Stream common targets.
8819 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8820 * ipa.c: Include value-prof.h
8821 (ipa_profile_generate_summary): Turn indirect call statement histograms
8822 into common targets.
8823 (ipa_profile): Turn common targets into speculative edges.
8824
2fa3d31b
JH		 88252013-08-09 Jan Hubicka <jh@suse.cz>
8826
8827 * cgraph.h (cgraph_node): Add profile_id.
8828 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 8829 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 8830 (del_node_map): Update.
8831 (find_func_by_funcdef_no): Replace by ...
8832 (find_func_by_profile_id): ... this one.
8833 (gimple_ic_transform): Do not remove useful histograms when
8834 speculation is not done; dump info when indirect call removal
8835 can happen at LTO.
8836 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8837 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8838 (__gcov_indirect_call_profiler_v2): .. this one.
8839 * profile.h (init_node_map): Update.
8840 * coverage.c (coverage_compute_profile_id): New function.
8841 * coverage.h (coverage_compute_profile_id): Declare.
8842 * tree-profile.c (init_ic_make_global_vars): Make
8843 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8844 (gimple_init_edge_profiler): Update prototype of
8845 __gcov_indirect_call_profiler.
8846 (gimple_gen_ic_func_profiler): Simplify.
8847 (tree_profiling): Use init_node_map
8848
042ae7d2
JH		 88492013-08-09 Jan Hubicka <jh@suse.cz>
8850
c97fd4b8
UB		 8851 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8852 non-speculative refs.
042ae7d2
JH		 8853 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8854 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8855 (cgraph_set_call_stmt): Likewise.
8856 (cgraph_create_edge_1): Fix release checking compilatoin;
8857 clear lto_stmt_uid.
8858 (cgraph_free_edge): Free indirect info.
8859 (cgraph_turn_edge_to_speculative): New function.
8860 (cgraph_speculative_call_info): New function.
8861 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 8862 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 8863 (dump_cgraph_node): Dump speculation.
8864 (verify_edge_count_and_frequency): Accept speculative edges.
8865 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8866 (verify_cgraph_node): Handle speculation.
8867 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8868 (cgraph_set_call_stmt): Update prototype.
8869 (cgraph_make_edge_direct): Update prototype.
8870 (cgraph_speculative_call_info): Declare.
8871 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8872 to change; update call of ipa_find_references.
8873 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8874 lto_stmt_uid and speculative flags.
8875 (ipa_dump_references): Dump speculation.
8876 (ipa_clone_references): Clone speculative flag.
8877 (ipa_clone_referring): Likewise.
8878 (ipa_clone_ref): New function.
8879 (ipa_find_reference): Look into lto_stmt_uids
8880 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8881 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8882 (ipa_find_reference): Update declaration.
8883 (ipa_clone_ref): Declare.
8884 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8885 stream speculative flag.
8886 (lto_output_ref): Stream statements uids and speculation.
8887 (input_ref): Likewise.
8888 (input_edge): Stream speuclation.
8889 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8890 (cgraph_set_call_stmt_including_clones): Handle speculation.
8891 * ipa-inline.c (heap_edge_removal_hook): New function.
8892 (inline_small_functions): Register it.
8893 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8894 also initialize refs.
8895 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8896 edge to change.
8897 (try_make_edge_direct_simple_call): Likewise.
8898 (try_make_edge_direct_simple_call): Likewise.
8899 (update_indirect_edges_after_inlining): Likewise.
8900 (remove_described_reference): Look proper lto_stmt_uid.
8901 (propagate_controlled_uses): Likewise.
8902 (propagate_controlled_uses): Liekwise.
8903 * tree-inline.c (copy_bb): Copy speculative edges.
8904 (redirect_all_calls): New function.
c97fd4b8
UB		 8905 (copy_cfg_body): Do redirection after loop info is updated.
8906 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 8907
aa1e10cc
JH		 89082013-08-09 Jan Hubicka <jh@suse.cz>
8909
8910 * lto-streamer-out.c (output_function): Renumber PHIs.
8911 * lto-streamer-in.c (input_function): Likewise.
8912
66adb8eb
JG		 89132013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8914
8915 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8916 (get_lane_unsigned): Likewise.
8917 (dup_lane_scalar): Likewise.
8918 (get_lane): enable for VALL.
8919 * config/aarch64/aarch64-simd.md
8920 (aarch64_dup_lane_scalar<mode>): Remove.
8921 (aarch64_get_lane_signed<mode>): Likewise.
8922 (aarch64_get_lane_unsigned<mode>): Likewise.
8923 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8924 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8925 (aarch64_get_lane<mode>): Enable for all vector modes.
8926 (aarch64_get_lanedi): Remove misleading constraints.
8927 * config/aarch64/arm_neon.h
8928 (__aarch64_vget_lane_any): Define.
8929 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
8930 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
8931 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
8932 * config/aarch64/iterators.md (VDQQH): New.
8933 (VDQQHS): Likewise.
8934 (vwcore): Likewise.
8935
1f65ae7a
EB		 89362013-08-09 Eric Botcazou <ebotcazou@adacore.com>
8937
8938 * configure.ac: Add GAS check for LEON instructions on SPARC.
8939 * configure: Regenerate.
8940 * config.in: Likewise.
8941 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
8942 sparc*-*-* block.
8943 * config/sparc/sparc.opt (LEON, LEON3): New masks.
8944 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
8945 for LEON or LEON3.
8946 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
8947 (AS_LEON_FLAG): New macro.
8948 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
8949 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
8950 Deal with LEON and LEON3 for the memory model.
23a6cb78 8951 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 8952 (atomic_compare_and_swap<mode>_1): Likewise.
8953 (*atomic_compare_and_swap<mode>_1): Likewise.
8954
ff522f7f
ZC		 89552013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8956
8957 * config/arm/neon.md (vcond): Fix floating-point vector
8958 comparisons against 0.
8959
cb1cca12
VM		 89602013-08-08 Vladimir Makarov <vmakarov@redhat.com>
8961
8962 * lra-constraints.c (emit_spill_move): Remove assert.
8963 (process_alt_operands): Add more debugging
8964 output. Increase reject for spilling into memory. Decrease
8965 reject for reloading scratch.
8966 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
8967
a3719e31
SE		 89682013-08-08 Steve Ellcey <sellcey@mips.com>
8969
8970 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
8971 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
8972 micromips incompatible. Add nan2008.
8973 (MULTILIB_DIRNAMES): Add nan2008.
8974 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8975 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
8976 and micromips incompatible. Add nan2008.
8977 (MULTILIB_DIRNAMES): Add nan2008.
8978 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8979
f40423e2
RS		 89802013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
8981
8982 PR rtl-optimization/58079
8983 * combine.c (combine_simplify_rtx): Avoid using SUBST if
8984 simplify_comparison has widened a comparison with an integer.
8985
869b9125
KT		 89862013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8987
8988 * config/arm/neon.md (movmisalign<mode>): Disable when we
8989 don't allow unaligned accesses.
8990 (*movmisalign<mode>_neon_store): Likewise.
8991 (*movmisalign<mode>_neon_load): Likewise.
8992 (*movmisalign<mode>_neon_store): Likewise.
8993 (*movmisalign<mode>_neon_load): Likewise.
8994
71cafea9
JH		 89952013-08-08 Jan Hubicka <jh@suse.cz>
8996
8997 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
8998 (make_pass_rebuild_cgraph_edges): Also clear references.
8999 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9000 * ipa-inline-transform.c (inline_transform): Remove all references
9001 after inlining.
c97fd4b8
UB		 9002 * cgraphunit.c (expand_function): Remove all references after
9003 expansion.
71cafea9
JH		 9004 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9005 (ipa_find_reference): Rewrite to iterator.
9006 (remove_stmt_references): Likewise.
9007 (ipa_clear_stmts_in_references): New function.
9008 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 9009 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9010 clear references.
71cafea9
JH		 9011 * ipa-split.c (split_function): Remove references in split function.
9012
27d2e612
RE		 90132013-08-08 Richard Earnshaw <rearnsha@arm.com>
9014
9015 PR target/57431
1f65ae7a 9016 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 9017 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9018
5efc661c
RE		 90192013-08-08 Richard Earnshaw <rearnsha@arm.com>
9020
9021 PR target/56979
1f65ae7a 9022 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 9023 suggested mode for the assignment isn't compatible with the
9024 registers required.
9025
f276d31d
BE		 90262013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9027
9028 PR target/58065
9029 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9030
4b8bc035
XDL		 90312013-08-07 Xinliang David Li <davidxl@google.com>
9032
9033 * config/i386/i386.opt: New option -mtune-ctrl=.
9034 * config/i386/x86-tune.def: New file.
9035 * config/i386/i386.h: include x86-tune.def.
9036 * config/i386/i386.c (ix86_option_override_internal):
9037 Parsing -mtune-ctrl= option and set tune features.
9038
12211b99 90392013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 9040
9041 PR other/12081
03b0ee0a 9042 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 9043 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9044 to gen_2arg_fn_t.
9045
5cf6635b
EB		 90462013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9047
9048 * rtl.h (update_alignments): Declare.
9049 * final.c (grow_label_align): New function extracted from...
9050 (shorten_branches): ...here. Call it.
9051 (update_alignments): New function.
9052 * reorg.c (sibling_labels): New variable.
9053 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9054 the new label along with it onto the sibling_labels vector.
9055 (fill_simple_delay_slots): Adjust call to get_label_before.
9056 (fill_slots_from_thread): Likewise.
9057 (relax_delay_slots): Likewise.
9058 (make_return_insns): Likewise.
9059 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9060
3c8ca1ab
EB		 90612013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9062
9063 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9064 document its semantics.
9065 (diagnostic_report_diagnostic): Adjust accordingly.
9066
5ee5b32c
DM		 90672013-08-07 David Malcolm <dmalcolm@redhat.com>
9068
03b0ee0a 9069 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 9070 (sparc_option_override): ...and port to new C++ pass API.
9071 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9072
abd566fa
PB		 90732013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9074
9075 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9076
2077db1b
CT		 90772013-08-06 Caroline Tice <cmtice@google.com>
9078
9079 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9080 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9081 * tree-pass.h: Add pass_vtable_verify.
9082 * varasm.c (assemble_variable): Add code to properly set the comdat
9083 section and name for the .vtable_map_vars section.
9084 (assemble_vtyv_preinit_initializer): New function.
9085 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9086 LINK_ONCE flag.
9087 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9088 * vtable-verify.c: New file.
9089 * vtable-verify.h: New file.
9090 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9091 initialiation levels.
9092 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9093 * passes.def: Insert pass_vtable_verify.
9094 * aclocal.m4: Reorder includes.
9095 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9096 -fvtv-counts options.
9097 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9098 as appropriate, if -fvtable-verify=... is used.
9099 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9100 -fvtable-verify=... is used.
9101 * Makefile.in (OBJS): Add vtable-verify.o to list.
9102 (vtable-verify.o): Add new build rule.
9103 (GTFILES): Add vtable-verify.c to list.
9104 * common.opt (fvtable-verify=): New flag.
9105 (vtv_priority): Values for fvtable-verify= flag.
9106 (fvtv-counts): New flag.
9107 (fvtv-debug): New flag.
9108 * tree.h (save_vtable_map_decl): New extern function decl.
9109
03085d1c
DM		 91102013-08-07 David Malcolm <dmalcolm@redhat.com>
9111
9112 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9113 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9114 (pass_data_rl78_devirt): ...new pass_data instance and...
9115 (make_pass_rl78_devirt): ...new function.
9116 (rl78_asm_file_start): Port pass registration to new C++ API.
9117
05555c4a
DM		 91182013-08-07 David Malcolm <dmalcolm@redhat.com>
9119
9120 * coretypes.h (rtl_opt_pass): Add.
9121 (gcc::context): Add.
9122 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9123 (epiphany_init): Port to new C++ pass API.
9124 (epiphany_optimize_mode_switching): Likewise.
9125 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9126 (pass_manager::get_pass_mode_switching): New.
9127 (pass_manager::get_pass_peephole2): New.
9128 * mode-switching.c (pass_mode_switching): Add clone method.
9129 * recog.c (pass_peephole2): Add clone method.
9130 (pass_split_all_insns): Add clone method.
9131
c5a12e26
DM		 91322013-08-06 David Malcolm <dmalcolm@redhat.com>
9133
03b0ee0a
UB		 9134 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9135 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 9136
a2e2a668
JH		 91372013-08-06 Jan Hubicka <jh@suse.cz>
9138
9139 * cgraph.c (cgraph_get_body): New function based on lto.c
9140 implementation.
9141 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 9142 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9143 LTO paths.
a2e2a668
JH		 9144 * cgraphunit.c (expand_function): Get body prior expanding.
9145 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 9146 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9147 really need.
a2e2a668
JH		 9148 * passes.c (do_per_function_toporder): Get body.
9149 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9150 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9151
15682f24
MJ		 91522013-08-06 Martin Jambor <mjambor@suse.cz>
9153
9154 PR fortran/57987
9155 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9156 is not re-finalized. Rename second parameter to no_collect.
9157
78f6dd68
MJ		 91582013-08-06 Martin Jambor <mjambor@suse.cz>
9159
9160 PR middle-end/58041
9161 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9162 MEM_REF has proper alignment information.
9163
12211b99 91642013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 9165
9166 PR other/12081
9167 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9168 class insn_gen_fn.
9169 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9170 rtx (*) (rtx, ...) with insn_gen_fn.
9171 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9172 insn_gen_fn::stored_funcptr. Add initializer braces.
9173
8ac69a6c
DM		 91742013-08-05 David Malcolm <dmalcolm@redhat.com>
9175
9176 Rewrite how instances of passes are cloned to remove assumptions
9177 about their sizes (thus allowing pass subclasses to have
9178 additional data fields, albeit non-GC-managed ones at this point).
9179
9180 * passes.c (make_pass_instance): Now that passes have clone
9181 methods, rewrite this function to eliminate XNEW and memcpy
9182 calls that used hardcoded sizes. Since this function no longer
9183 creates pass instances, rename it to...
03b0ee0a
UB		 9184 (add_pass_instance): ...this. Document the old way that passes were
9185 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 9186 (next_pass_1): Add an initial_pass argument for use by
9187 add_pass_instance.
9188 (position_pass): When adding multiple instances of a pass, use
9189 the pass's clone method, rather than relying on the XNEW/memcpy
9190 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 9191 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9192 the initial instance of the current pass within the pass manager.
8ac69a6c 9193
27a4cd48
DM		 91942013-08-05 David Malcolm <dmalcolm@redhat.com>
9195
9196 This is the automated part of the conversion of passes from C
9197 structs to C++ classes.
9198
9199 Patch autogenerated by refactor_passes.py from
9200 https://github.com/davidmalcolm/gcc-refactoring-scripts
9201 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9202
9203 * asan.c (pass_asan): Convert from a global struct to a subclass of
9204 gimple_opt_pass along with...
9205 (pass_data_asan): ...new pass_data instance and...
9206 (make_pass_asan): ...new function.
9207 (pass_asan_O0): Convert from a global struct to a subclass of
9208 gimple_opt_pass along with...
9209 (pass_data_asan_O0): ...new pass_data instance and...
9210 (make_pass_asan_O0): ...new function.
9211 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9212 subclass of rtl_opt_pass along with...
9213 (pass_data_inc_dec): ...new pass_data instance and...
9214 (make_pass_inc_dec): ...new function.
9215 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9216 a subclass of rtl_opt_pass along with...
9217 (pass_data_reorder_blocks): ...new pass_data instance and...
9218 (make_pass_reorder_blocks): ...new function.
9219 (pass_duplicate_computed_gotos): Convert from a global struct to a
9220 subclass of rtl_opt_pass along with...
9221 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9222 (make_pass_duplicate_computed_gotos): ...new function.
9223 (pass_partition_blocks): Convert from a global struct to a subclass of
9224 rtl_opt_pass along with...
9225 (pass_data_partition_blocks): ...new pass_data instance and...
9226 (make_pass_partition_blocks): ...new function.
9227 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9228 struct to a subclass of rtl_opt_pass along with...
9229 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9230 and...
9231 (make_pass_branch_target_load_optimize1): ...new function.
9232 (pass_branch_target_load_optimize2): Convert from a global struct to a
9233 subclass of rtl_opt_pass along with...
9234 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9235 and...
9236 (make_pass_branch_target_load_optimize2): ...new function.
9237 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9238 of rtl_opt_pass along with...
9239 (pass_data_jump): ...new pass_data instance and...
9240 (make_pass_jump): ...new function.
9241 (pass_jump2): Convert from a global struct to a subclass of
9242 rtl_opt_pass along with...
9243 (pass_data_jump2): ...new pass_data instance and...
9244 (make_pass_jump2): ...new function.
9245 * cfgexpand.c (pass_expand): Convert from a global struct to a
9246 subclass of rtl_opt_pass along with...
9247 (pass_data_expand): ...new pass_data instance and...
9248 (make_pass_expand): ...new function.
9249 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9250 of rtl_opt_pass along with...
9251 (pass_data_free_cfg): ...new pass_data instance and...
9252 (make_pass_free_cfg): ...new function.
9253 (pass_into_cfg_layout_mode): Convert from a global struct to a
9254 subclass of rtl_opt_pass along with...
9255 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9256 (make_pass_into_cfg_layout_mode): ...new function.
9257 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9258 subclass of rtl_opt_pass along with...
9259 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9260 (make_pass_outof_cfg_layout_mode): ...new function.
9261 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9262 struct to a subclass of gimple_opt_pass along with...
9263 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9264 (make_pass_build_cgraph_edges): ...new function.
9265 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9266 subclass of gimple_opt_pass along with...
9267 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9268 (make_pass_rebuild_cgraph_edges): ...new function.
9269 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9270 subclass of gimple_opt_pass along with...
9271 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9272 and...
9273 (make_pass_remove_cgraph_callee_edges): ...new function.
9274 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9275 struct to a subclass of rtl_opt_pass along with...
9276 (pass_data_stack_adjustments): ...new pass_data instance and...
9277 (make_pass_stack_adjustments): ...new function.
9278 * combine.c (pass_combine): Convert from a global struct to a subclass
9279 of rtl_opt_pass along with...
9280 (pass_data_combine): ...new pass_data instance and...
9281 (make_pass_combine): ...new function.
9282 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9283 global struct to a subclass of rtl_opt_pass along with...
9284 (pass_data_compare_elim_after_reload): ...new pass_data instance
9285 and...
9286 (make_pass_compare_elim_after_reload): ...new function.
9287 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9288 of rtl_opt_pass along with...
9289 (pass_data_rtl_cprop): ...new pass_data instance and...
9290 (make_pass_rtl_cprop): ...new function.
9291 * cse.c (pass_cse): Convert from a global struct to a subclass of
9292 rtl_opt_pass along with...
9293 (pass_data_cse): ...new pass_data instance and...
9294 (make_pass_cse): ...new function.
9295 (pass_cse2): Convert from a global struct to a subclass of
9296 rtl_opt_pass along with...
9297 (pass_data_cse2): ...new pass_data instance and...
9298 (make_pass_cse2): ...new function.
9299 (pass_cse_after_global_opts): Convert from a global struct to a
9300 subclass of rtl_opt_pass along with...
9301 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9302 (make_pass_cse_after_global_opts): ...new function.
9303 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9304 of rtl_opt_pass along with...
9305 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9306 (make_pass_ud_rtl_dce): ...new function.
9307 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9308 rtl_opt_pass along with...
9309 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9310 (make_pass_fast_rtl_dce): ...new function.
9311 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9312 a subclass of rtl_opt_pass along with...
9313 (pass_data_df_initialize_opt): ...new pass_data instance and...
9314 (make_pass_df_initialize_opt): ...new function.
9315 (pass_df_initialize_no_opt): Convert from a global struct to a
9316 subclass of rtl_opt_pass along with...
9317 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9318 (make_pass_df_initialize_no_opt): ...new function.
9319 (pass_df_finish): Convert from a global struct to a subclass of
9320 rtl_opt_pass along with...
9321 (pass_data_df_finish): ...new pass_data instance and...
9322 (make_pass_df_finish): ...new function.
9323 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9324 rtl_opt_pass along with...
9325 (pass_data_rtl_dse1): ...new pass_data instance and...
9326 (make_pass_rtl_dse1): ...new function.
9327 (pass_rtl_dse2): Convert from a global struct to a subclass of
9328 rtl_opt_pass along with...
9329 (pass_data_rtl_dse2): ...new pass_data instance and...
9330 (make_pass_rtl_dse2): ...new function.
9331 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9332 subclass of rtl_opt_pass along with...
9333 (pass_data_dwarf2_frame): ...new pass_data instance and...
9334 (make_pass_dwarf2_frame): ...new function.
9335 * except.c (pass_set_nothrow_function_flags): Convert from a global
9336 struct to a subclass of rtl_opt_pass along with...
9337 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9338 and...
9339 (make_pass_set_nothrow_function_flags): ...new function.
9340 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9341 subclass of rtl_opt_pass along with...
9342 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9343 and...
9344 (make_pass_convert_to_eh_region_ranges): ...new function.
9345 * final.c (pass_compute_alignments): Convert from a global struct to a
9346 subclass of rtl_opt_pass along with...
9347 (pass_data_compute_alignments): ...new pass_data instance and...
9348 (make_pass_compute_alignments): ...new function.
9349 (pass_final): Convert from a global struct to a subclass of
9350 rtl_opt_pass along with...
9351 (pass_data_final): ...new pass_data instance and...
9352 (make_pass_final): ...new function.
9353 (pass_shorten_branches): Convert from a global struct to a subclass of
9354 rtl_opt_pass along with...
9355 (pass_data_shorten_branches): ...new pass_data instance and...
9356 (make_pass_shorten_branches): ...new function.
9357 (pass_clean_state): Convert from a global struct to a subclass of
9358 rtl_opt_pass along with...
9359 (pass_data_clean_state): ...new pass_data instance and...
9360 (make_pass_clean_state): ...new function.
9361 * function.c (pass_instantiate_virtual_regs): Convert from a global
9362 struct to a subclass of rtl_opt_pass along with...
9363 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9364 (make_pass_instantiate_virtual_regs): ...new function.
9365 (pass_leaf_regs): Convert from a global struct to a subclass of
9366 rtl_opt_pass along with...
9367 (pass_data_leaf_regs): ...new pass_data instance and...
9368 (make_pass_leaf_regs): ...new function.
9369 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9370 subclass of rtl_opt_pass along with...
9371 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9372 and...
9373 (make_pass_thread_prologue_and_epilogue): ...new function.
9374 (pass_match_asm_constraints): Convert from a global struct to a
9375 subclass of rtl_opt_pass along with...
9376 (pass_data_match_asm_constraints): ...new pass_data instance and...
9377 (make_pass_match_asm_constraints): ...new function.
9378 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9379 subclass of rtl_opt_pass along with...
9380 (pass_data_rtl_fwprop): ...new pass_data instance and...
9381 (make_pass_rtl_fwprop): ...new function.
9382 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9383 rtl_opt_pass along with...
9384 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9385 (make_pass_rtl_fwprop_addr): ...new function.
9386 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9387 rtl_opt_pass along with...
9388 (pass_data_rtl_pre): ...new pass_data instance and...
9389 (make_pass_rtl_pre): ...new function.
9390 (pass_rtl_hoist): Convert from a global struct to a subclass of
9391 rtl_opt_pass along with...
9392 (pass_data_rtl_hoist): ...new pass_data instance and...
9393 (make_pass_rtl_hoist): ...new function.
9394 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9395 subclass of gimple_opt_pass along with...
9396 (pass_data_lower_cf): ...new pass_data instance and...
9397 (make_pass_lower_cf): ...new function.
9398 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9399 from a global struct to a subclass of gimple_opt_pass along with...
9400 (pass_data_strength_reduction): ...new pass_data instance and...
9401 (make_pass_strength_reduction): ...new function.
9402 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9403 of rtl_opt_pass along with...
9404 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9405 (make_pass_rtl_ifcvt): ...new function.
9406 (pass_if_after_combine): Convert from a global struct to a subclass of
9407 rtl_opt_pass along with...
9408 (pass_data_if_after_combine): ...new pass_data instance and...
9409 (make_pass_if_after_combine): ...new function.
9410 (pass_if_after_reload): Convert from a global struct to a subclass of
9411 rtl_opt_pass along with...
9412 (pass_data_if_after_reload): ...new pass_data instance and...
9413 (make_pass_if_after_reload): ...new function.
9414 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9415 a subclass of rtl_opt_pass along with...
9416 (pass_data_initialize_regs): ...new pass_data instance and...
9417 (make_pass_initialize_regs): ...new function.
9418 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9419 of ipa_opt_pass_d along with...
9420 (pass_data_ipa_cp): ...new pass_data instance and...
9421 (make_pass_ipa_cp): ...new function.
9422 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9423 global struct to a subclass of gimple_opt_pass along with...
9424 (pass_data_inline_parameters): ...new pass_data instance and...
9425 (make_pass_inline_parameters): ...new function.
9426 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9427 subclass of gimple_opt_pass along with...
9428 (pass_data_early_inline): ...new pass_data instance and...
9429 (make_pass_early_inline): ...new function.
9430 (pass_ipa_inline): Convert from a global struct to a subclass of
9431 ipa_opt_pass_d along with...
9432 (pass_data_ipa_inline): ...new pass_data instance and...
9433 (make_pass_ipa_inline): ...new function.
9434 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9435 struct to a subclass of gimple_opt_pass along with...
9436 (pass_data_local_pure_const): ...new pass_data instance and...
9437 (make_pass_local_pure_const): ...new function.
9438 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9439 ipa_opt_pass_d along with...
9440 (pass_data_ipa_pure_const): ...new pass_data instance and...
9441 (make_pass_ipa_pure_const): ...new function.
9442 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9443 to a subclass of ipa_opt_pass_d along with...
9444 (pass_data_ipa_reference): ...new pass_data instance and...
9445 (make_pass_ipa_reference): ...new function.
9446 * ipa-split.c (pass_split_functions): Convert from a global struct to
9447 a subclass of gimple_opt_pass along with...
9448 (pass_data_split_functions): ...new pass_data instance and...
9449 (make_pass_split_functions): ...new function.
9450 (pass_feedback_split_functions): Convert from a global struct to a
9451 subclass of gimple_opt_pass along with...
9452 (pass_data_feedback_split_functions): ...new pass_data instance and...
9453 (make_pass_feedback_split_functions): ...new function.
9454 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9455 global struct to a subclass of simple_ipa_opt_pass along with...
9456 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9457 instance and...
9458 (make_pass_ipa_function_and_variable_visibility): ...new function.
9459 (pass_ipa_free_inline_summary): Convert from a global struct to a
9460 subclass of simple_ipa_opt_pass along with...
9461 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9462 (make_pass_ipa_free_inline_summary): ...new function.
9463 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9464 subclass of ipa_opt_pass_d along with...
9465 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9466 and...
9467 (make_pass_ipa_whole_program_visibility): ...new function.
9468 (pass_ipa_profile): Convert from a global struct to a subclass of
9469 ipa_opt_pass_d along with...
9470 (pass_data_ipa_profile): ...new pass_data instance and...
9471 (make_pass_ipa_profile): ...new function.
9472 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9473 ipa_opt_pass_d along with...
9474 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9475 (make_pass_ipa_cdtor_merge): ...new function.
9476 * ira.c (pass_ira): Convert from a global struct to a subclass of
9477 rtl_opt_pass along with...
9478 (pass_data_ira): ...new pass_data instance and...
9479 (make_pass_ira): ...new function.
9480 (pass_reload): Convert from a global struct to a subclass of
9481 rtl_opt_pass along with...
9482 (pass_data_reload): ...new pass_data instance and...
9483 (make_pass_reload): ...new function.
9484 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9485 subclass of rtl_opt_pass along with...
9486 (pass_data_cleanup_barriers): ...new pass_data instance and...
9487 (make_pass_cleanup_barriers): ...new function.
9488 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9489 of rtl_opt_pass along with...
9490 (pass_data_loop2): ...new pass_data instance and...
9491 (make_pass_loop2): ...new function.
9492 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9493 rtl_opt_pass along with...
9494 (pass_data_rtl_loop_init): ...new pass_data instance and...
9495 (make_pass_rtl_loop_init): ...new function.
9496 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9497 rtl_opt_pass along with...
9498 (pass_data_rtl_loop_done): ...new pass_data instance and...
9499 (make_pass_rtl_loop_done): ...new function.
9500 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9501 subclass of rtl_opt_pass along with...
9502 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9503 (make_pass_rtl_move_loop_invariants): ...new function.
9504 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9505 rtl_opt_pass along with...
9506 (pass_data_rtl_unswitch): ...new pass_data instance and...
9507 (make_pass_rtl_unswitch): ...new function.
9508 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9509 subclass of rtl_opt_pass along with...
9510 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9511 and...
9512 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9513 (pass_rtl_doloop): Convert from a global struct to a subclass of
9514 rtl_opt_pass along with...
9515 (pass_data_rtl_doloop): ...new pass_data instance and...
9516 (make_pass_rtl_doloop): ...new function.
9517 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9518 a subclass of rtl_opt_pass along with...
9519 (pass_data_lower_subreg): ...new pass_data instance and...
9520 (make_pass_lower_subreg): ...new function.
9521 (pass_lower_subreg2): Convert from a global struct to a subclass of
9522 rtl_opt_pass along with...
9523 (pass_data_lower_subreg2): ...new pass_data instance and...
9524 (make_pass_lower_subreg2): ...new function.
9525 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9526 struct to a subclass of ipa_opt_pass_d along with...
9527 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9528 (make_pass_ipa_lto_gimple_out): ...new function.
9529 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9530 of ipa_opt_pass_d along with...
9531 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9532 (make_pass_ipa_lto_finish_out): ...new function.
9533 * mode-switching.c (pass_mode_switching): Convert from a global struct
9534 to a subclass of rtl_opt_pass along with...
9535 (pass_data_mode_switching): ...new pass_data instance and...
9536 (make_pass_mode_switching): ...new function.
9537 * modulo-sched.c (pass_sms): Convert from a global struct to a
9538 subclass of rtl_opt_pass along with...
9539 (pass_data_sms): ...new pass_data instance and...
9540 (make_pass_sms): ...new function.
9541 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9542 subclass of gimple_opt_pass along with...
9543 (pass_data_expand_omp): ...new pass_data instance and...
9544 (make_pass_expand_omp): ...new function.
9545 (pass_lower_omp): Convert from a global struct to a subclass of
9546 gimple_opt_pass along with...
9547 (pass_data_lower_omp): ...new pass_data instance and...
9548 (make_pass_lower_omp): ...new function.
9549 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9550 of gimple_opt_pass along with...
9551 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9552 (make_pass_diagnose_omp_blocks): ...new function.
9553 * passes.c (pass_early_local_passes): Convert from a global struct to
9554 a subclass of simple_ipa_opt_pass along with...
9555 (pass_data_early_local_passes): ...new pass_data instance and...
9556 (make_pass_early_local_passes): ...new function.
9557 (pass_all_early_optimizations): Convert from a global struct to a
9558 subclass of gimple_opt_pass along with...
9559 (pass_data_all_early_optimizations): ...new pass_data instance and...
9560 (make_pass_all_early_optimizations): ...new function.
9561 (pass_all_optimizations): Convert from a global struct to a subclass
9562 of gimple_opt_pass along with...
9563 (pass_data_all_optimizations): ...new pass_data instance and...
9564 (make_pass_all_optimizations): ...new function.
9565 (pass_all_optimizations_g): Convert from a global struct to a subclass
9566 of gimple_opt_pass along with...
9567 (pass_data_all_optimizations_g): ...new pass_data instance and...
9568 (make_pass_all_optimizations_g): ...new function.
9569 (pass_rest_of_compilation): Convert from a global struct to a subclass
9570 of rtl_opt_pass along with...
9571 (pass_data_rest_of_compilation): ...new pass_data instance and...
9572 (make_pass_rest_of_compilation): ...new function.
9573 (pass_postreload): Convert from a global struct to a subclass of
9574 rtl_opt_pass along with...
9575 (pass_data_postreload): ...new pass_data instance and...
9576 (make_pass_postreload): ...new function.
9577 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9578 subclass of rtl_opt_pass along with...
9579 (pass_data_gcse2): ...new pass_data instance and...
9580 (make_pass_gcse2): ...new function.
9581 * postreload.c (pass_postreload_cse): Convert from a global struct to
9582 a subclass of rtl_opt_pass along with...
9583 (pass_data_postreload_cse): ...new pass_data instance and...
9584 (make_pass_postreload_cse): ...new function.
9585 * predict.c (pass_profile): Convert from a global struct to a subclass
9586 of gimple_opt_pass along with...
9587 (pass_data_profile): ...new pass_data instance and...
9588 (make_pass_profile): ...new function.
9589 (pass_strip_predict_hints): Convert from a global struct to a subclass
9590 of gimple_opt_pass along with...
9591 (pass_data_strip_predict_hints): ...new pass_data instance and...
9592 (make_pass_strip_predict_hints): ...new function.
9593 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9594 of rtl_opt_pass along with...
9595 (pass_data_peephole2): ...new pass_data instance and...
9596 (make_pass_peephole2): ...new function.
9597 (pass_split_all_insns): Convert from a global struct to a subclass of
9598 rtl_opt_pass along with...
9599 (pass_data_split_all_insns): ...new pass_data instance and...
9600 (make_pass_split_all_insns): ...new function.
9601 (pass_split_after_reload): Convert from a global struct to a subclass
9602 of rtl_opt_pass along with...
9603 (pass_data_split_after_reload): ...new pass_data instance and...
9604 (make_pass_split_after_reload): ...new function.
9605 (pass_split_before_regstack): Convert from a global struct to a
9606 subclass of rtl_opt_pass along with...
9607 (pass_data_split_before_regstack): ...new pass_data instance and...
9608 (make_pass_split_before_regstack): ...new function.
9609 (pass_split_before_sched2): Convert from a global struct to a subclass
9610 of rtl_opt_pass along with...
9611 (pass_data_split_before_sched2): ...new pass_data instance and...
9612 (make_pass_split_before_sched2): ...new function.
9613 (pass_split_for_shorten_branches): Convert from a global struct to a
9614 subclass of rtl_opt_pass along with...
9615 (pass_data_split_for_shorten_branches): ...new pass_data instance
9616 and...
9617 (make_pass_split_for_shorten_branches): ...new function.
9618 * ree.c (pass_ree): Convert from a global struct to a subclass of
9619 rtl_opt_pass along with...
9620 (pass_data_ree): ...new pass_data instance and...
9621 (make_pass_ree): ...new function.
9622 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9623 subclass of rtl_opt_pass along with...
9624 (pass_data_stack_regs): ...new pass_data instance and...
9625 (make_pass_stack_regs): ...new function.
9626 (pass_stack_regs_run): Convert from a global struct to a subclass of
9627 rtl_opt_pass along with...
9628 (pass_data_stack_regs_run): ...new pass_data instance and...
9629 (make_pass_stack_regs_run): ...new function.
9630 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9631 subclass of rtl_opt_pass along with...
9632 (pass_data_cprop_hardreg): ...new pass_data instance and...
9633 (make_pass_cprop_hardreg): ...new function.
9634 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9635 subclass of rtl_opt_pass along with...
9636 (pass_data_reginfo_init): ...new pass_data instance and...
9637 (make_pass_reginfo_init): ...new function.
9638 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9639 of rtl_opt_pass along with...
9640 (pass_data_regmove): ...new pass_data instance and...
9641 (make_pass_regmove): ...new function.
9642 * regrename.c (pass_regrename): Convert from a global struct to a
9643 subclass of rtl_opt_pass along with...
9644 (pass_data_regrename): ...new pass_data instance and...
9645 (make_pass_regrename): ...new function.
9646 * reorg.c (pass_delay_slots): Convert from a global struct to a
9647 subclass of rtl_opt_pass along with...
9648 (pass_data_delay_slots): ...new pass_data instance and...
9649 (make_pass_delay_slots): ...new function.
9650 (pass_machine_reorg): Convert from a global struct to a subclass of
9651 rtl_opt_pass along with...
9652 (pass_data_machine_reorg): ...new pass_data instance and...
9653 (make_pass_machine_reorg): ...new function.
9654 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9655 of rtl_opt_pass along with...
9656 (pass_data_sched): ...new pass_data instance and...
9657 (make_pass_sched): ...new function.
9658 (pass_sched2): Convert from a global struct to a subclass of
9659 rtl_opt_pass along with...
9660 (pass_data_sched2): ...new pass_data instance and...
9661 (make_pass_sched2): ...new function.
9662 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9663 to a subclass of rtl_opt_pass along with...
9664 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9665 (make_pass_stack_ptr_mod): ...new function.
9666 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9667 to a subclass of rtl_opt_pass along with...
9668 (pass_data_rtl_store_motion): ...new pass_data instance and...
9669 (make_pass_rtl_store_motion): ...new function.
9670 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9671 of gimple_opt_pass along with...
9672 (pass_data_tracer): ...new pass_data instance and...
9673 (make_pass_tracer): ...new function.
9674 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9675 to a subclass of gimple_opt_pass along with...
9676 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9677 (make_pass_diagnose_tm_blocks): ...new function.
9678 (pass_lower_tm): Convert from a global struct to a subclass of
9679 gimple_opt_pass along with...
9680 (pass_data_lower_tm): ...new pass_data instance and...
9681 (make_pass_lower_tm): ...new function.
9682 (pass_tm_init): Convert from a global struct to a subclass of
9683 gimple_opt_pass along with...
9684 (pass_data_tm_init): ...new pass_data instance and...
9685 (make_pass_tm_init): ...new function.
9686 (pass_tm_mark): Convert from a global struct to a subclass of
9687 gimple_opt_pass along with...
9688 (pass_data_tm_mark): ...new pass_data instance and...
9689 (make_pass_tm_mark): ...new function.
9690 (pass_tm_edges): Convert from a global struct to a subclass of
9691 gimple_opt_pass along with...
9692 (pass_data_tm_edges): ...new pass_data instance and...
9693 (make_pass_tm_edges): ...new function.
9694 (pass_tm_memopt): Convert from a global struct to a subclass of
9695 gimple_opt_pass along with...
9696 (pass_data_tm_memopt): ...new pass_data instance and...
9697 (make_pass_tm_memopt): ...new function.
9698 (pass_ipa_tm): Convert from a global struct to a subclass of
9699 simple_ipa_opt_pass along with...
9700 (pass_data_ipa_tm): ...new pass_data instance and...
9701 (make_pass_ipa_tm): ...new function.
9702 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9703 subclass of gimple_opt_pass along with...
9704 (pass_data_call_cdce): ...new pass_data instance and...
9705 (make_pass_call_cdce): ...new function.
9706 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9707 subclass of gimple_opt_pass along with...
9708 (pass_data_build_cfg): ...new pass_data instance and...
9709 (make_pass_build_cfg): ...new function.
9710 (pass_split_crit_edges): Convert from a global struct to a subclass of
9711 gimple_opt_pass along with...
9712 (pass_data_split_crit_edges): ...new pass_data instance and...
9713 (make_pass_split_crit_edges): ...new function.
9714 (pass_warn_function_return): Convert from a global struct to a
9715 subclass of gimple_opt_pass along with...
9716 (pass_data_warn_function_return): ...new pass_data instance and...
9717 (make_pass_warn_function_return): ...new function.
9718 (pass_warn_function_noreturn): Convert from a global struct to a
9719 subclass of gimple_opt_pass along with...
9720 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9721 (make_pass_warn_function_noreturn): ...new function.
9722 (pass_warn_unused_result): Convert from a global struct to a subclass
9723 of gimple_opt_pass along with...
9724 (pass_data_warn_unused_result): ...new pass_data instance and...
9725 (make_pass_warn_unused_result): ...new function.
9726 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9727 a subclass of gimple_opt_pass along with...
9728 (pass_data_merge_phi): ...new pass_data instance and...
9729 (make_pass_merge_phi): ...new function.
9730 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9731 a subclass of gimple_opt_pass along with...
9732 (pass_data_lower_complex): ...new pass_data instance and...
9733 (make_pass_lower_complex): ...new function.
9734 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9735 gimple_opt_pass along with...
9736 (pass_data_lower_complex_O0): ...new pass_data instance and...
9737 (make_pass_lower_complex_O0): ...new function.
9738 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9739 subclass of gimple_opt_pass along with...
9740 (pass_data_lower_eh): ...new pass_data instance and...
9741 (make_pass_lower_eh): ...new function.
9742 (pass_refactor_eh): Convert from a global struct to a subclass of
9743 gimple_opt_pass along with...
9744 (pass_data_refactor_eh): ...new pass_data instance and...
9745 (make_pass_refactor_eh): ...new function.
9746 (pass_lower_resx): Convert from a global struct to a subclass of
9747 gimple_opt_pass along with...
9748 (pass_data_lower_resx): ...new pass_data instance and...
9749 (make_pass_lower_resx): ...new function.
9750 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9751 of gimple_opt_pass along with...
9752 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9753 (make_pass_lower_eh_dispatch): ...new function.
9754 (pass_cleanup_eh): Convert from a global struct to a subclass of
9755 gimple_opt_pass along with...
9756 (pass_data_cleanup_eh): ...new pass_data instance and...
9757 (make_pass_cleanup_eh): ...new function.
9758 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9759 to a subclass of simple_ipa_opt_pass along with...
9760 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9761 (make_pass_ipa_lower_emutls): ...new function.
9762 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9763 a subclass of gimple_opt_pass along with...
9764 (pass_data_if_conversion): ...new pass_data instance and...
9765 (make_pass_if_conversion): ...new function.
9766 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9767 subclass of gimple_opt_pass along with...
9768 (pass_data_build_ssa): ...new pass_data instance and...
9769 (make_pass_build_ssa): ...new function.
9770 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9771 global struct to a subclass of gimple_opt_pass along with...
9772 (pass_data_loop_distribution): ...new pass_data instance and...
9773 (make_pass_loop_distribution): ...new function.
9774 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9775 subclass of gimple_opt_pass along with...
9776 (pass_data_mudflap_1): ...new pass_data instance and...
9777 (make_pass_mudflap_1): ...new function.
9778 (pass_mudflap_2): Convert from a global struct to a subclass of
9779 gimple_opt_pass along with...
9780 (pass_data_mudflap_2): ...new pass_data instance and...
9781 (make_pass_mudflap_2): ...new function.
9782 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9783 subclass of gimple_opt_pass along with...
9784 (pass_data_mudflap_1): ...new pass_data instance and...
9785 (make_pass_mudflap_1): ...new function.
9786 (pass_mudflap_2): Convert from a global struct to a subclass of
9787 gimple_opt_pass along with...
9788 (pass_data_mudflap_2): ...new pass_data instance and...
9789 (make_pass_mudflap_2): ...new function.
9790 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9791 gimple_opt_pass along with...
9792 (pass_data_nrv): ...new pass_data instance and...
9793 (make_pass_nrv): ...new function.
9794 (pass_return_slot): Convert from a global struct to a subclass of
9795 gimple_opt_pass along with...
9796 (pass_data_return_slot): ...new pass_data instance and...
9797 (make_pass_return_slot): ...new function.
9798 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9799 to a subclass of gimple_opt_pass along with...
9800 (pass_data_object_sizes): ...new pass_data instance and...
9801 (make_pass_object_sizes): ...new function.
9802 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9803 global struct to a subclass of gimple_opt_pass along with...
9804 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9805 and...
9806 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9807 (pass_fixup_cfg): Convert from a global struct to a subclass of
9808 gimple_opt_pass along with...
9809 (pass_data_fixup_cfg): ...new pass_data instance and...
9810 (make_pass_fixup_cfg): ...new function.
9811 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9812 (make_pass_mudflap_1): ...new function.
9813 (pass_mudflap_2): Replace declaration with that of...
9814 (make_pass_mudflap_2): ...new function.
9815 (pass_asan): Replace declaration with that of...
9816 (make_pass_asan): ...new function.
9817 (pass_asan_O0): Replace declaration with that of...
9818 (make_pass_asan_O0): ...new function.
9819 (pass_tsan): Replace declaration with that of...
9820 (make_pass_tsan): ...new function.
9821 (pass_tsan_O0): Replace declaration with that of...
9822 (make_pass_tsan_O0): ...new function.
9823 (pass_lower_cf): Replace declaration with that of...
9824 (make_pass_lower_cf): ...new function.
9825 (pass_refactor_eh): Replace declaration with that of...
9826 (make_pass_refactor_eh): ...new function.
9827 (pass_lower_eh): Replace declaration with that of...
9828 (make_pass_lower_eh): ...new function.
9829 (pass_lower_eh_dispatch): Replace declaration with that of...
9830 (make_pass_lower_eh_dispatch): ...new function.
9831 (pass_lower_resx): Replace declaration with that of...
9832 (make_pass_lower_resx): ...new function.
9833 (pass_build_cfg): Replace declaration with that of...
9834 (make_pass_build_cfg): ...new function.
9835 (pass_early_tree_profile): Replace declaration with that of...
9836 (make_pass_early_tree_profile): ...new function.
9837 (pass_cleanup_eh): Replace declaration with that of...
9838 (make_pass_cleanup_eh): ...new function.
9839 (pass_sra): Replace declaration with that of...
9840 (make_pass_sra): ...new function.
9841 (pass_sra_early): Replace declaration with that of...
9842 (make_pass_sra_early): ...new function.
9843 (pass_early_ipa_sra): Replace declaration with that of...
9844 (make_pass_early_ipa_sra): ...new function.
9845 (pass_tail_recursion): Replace declaration with that of...
9846 (make_pass_tail_recursion): ...new function.
9847 (pass_tail_calls): Replace declaration with that of...
9848 (make_pass_tail_calls): ...new function.
9849 (pass_tree_loop): Replace declaration with that of...
9850 (make_pass_tree_loop): ...new function.
9851 (pass_tree_loop_init): Replace declaration with that of...
9852 (make_pass_tree_loop_init): ...new function.
9853 (pass_lim): Replace declaration with that of...
9854 (make_pass_lim): ...new function.
9855 (pass_tree_unswitch): Replace declaration with that of...
9856 (make_pass_tree_unswitch): ...new function.
9857 (pass_predcom): Replace declaration with that of...
9858 (make_pass_predcom): ...new function.
9859 (pass_iv_canon): Replace declaration with that of...
9860 (make_pass_iv_canon): ...new function.
9861 (pass_scev_cprop): Replace declaration with that of...
9862 (make_pass_scev_cprop): ...new function.
9863 (pass_empty_loop): Replace declaration with that of...
9864 (make_pass_empty_loop): ...new function.
9865 (pass_record_bounds): Replace declaration with that of...
9866 (make_pass_record_bounds): ...new function.
9867 (pass_graphite): Replace declaration with that of...
9868 (make_pass_graphite): ...new function.
9869 (pass_graphite_transforms): Replace declaration with that of...
9870 (make_pass_graphite_transforms): ...new function.
9871 (pass_if_conversion): Replace declaration with that of...
9872 (make_pass_if_conversion): ...new function.
9873 (pass_loop_distribution): Replace declaration with that of...
9874 (make_pass_loop_distribution): ...new function.
9875 (pass_vectorize): Replace declaration with that of...
9876 (make_pass_vectorize): ...new function.
9877 (pass_slp_vectorize): Replace declaration with that of...
9878 (make_pass_slp_vectorize): ...new function.
9879 (pass_complete_unroll): Replace declaration with that of...
9880 (make_pass_complete_unroll): ...new function.
9881 (pass_complete_unrolli): Replace declaration with that of...
9882 (make_pass_complete_unrolli): ...new function.
9883 (pass_parallelize_loops): Replace declaration with that of...
9884 (make_pass_parallelize_loops): ...new function.
9885 (pass_loop_prefetch): Replace declaration with that of...
9886 (make_pass_loop_prefetch): ...new function.
9887 (pass_iv_optimize): Replace declaration with that of...
9888 (make_pass_iv_optimize): ...new function.
9889 (pass_tree_loop_done): Replace declaration with that of...
9890 (make_pass_tree_loop_done): ...new function.
9891 (pass_ch): Replace declaration with that of...
9892 (make_pass_ch): ...new function.
9893 (pass_ccp): Replace declaration with that of...
9894 (make_pass_ccp): ...new function.
9895 (pass_phi_only_cprop): Replace declaration with that of...
9896 (make_pass_phi_only_cprop): ...new function.
9897 (pass_build_ssa): Replace declaration with that of...
9898 (make_pass_build_ssa): ...new function.
9899 (pass_build_alias): Replace declaration with that of...
9900 (make_pass_build_alias): ...new function.
9901 (pass_build_ealias): Replace declaration with that of...
9902 (make_pass_build_ealias): ...new function.
9903 (pass_dominator): Replace declaration with that of...
9904 (make_pass_dominator): ...new function.
9905 (pass_dce): Replace declaration with that of...
9906 (make_pass_dce): ...new function.
9907 (pass_dce_loop): Replace declaration with that of...
9908 (make_pass_dce_loop): ...new function.
9909 (pass_cd_dce): Replace declaration with that of...
9910 (make_pass_cd_dce): ...new function.
9911 (pass_call_cdce): Replace declaration with that of...
9912 (make_pass_call_cdce): ...new function.
9913 (pass_merge_phi): Replace declaration with that of...
9914 (make_pass_merge_phi): ...new function.
9915 (pass_split_crit_edges): Replace declaration with that of...
9916 (make_pass_split_crit_edges): ...new function.
9917 (pass_pre): Replace declaration with that of...
9918 (make_pass_pre): ...new function.
9919 (pass_profile): Replace declaration with that of...
9920 (make_pass_profile): ...new function.
9921 (pass_strip_predict_hints): Replace declaration with that of...
9922 (make_pass_strip_predict_hints): ...new function.
9923 (pass_lower_complex_O0): Replace declaration with that of...
9924 (make_pass_lower_complex_O0): ...new function.
9925 (pass_lower_complex): Replace declaration with that of...
9926 (make_pass_lower_complex): ...new function.
9927 (pass_lower_vector): Replace declaration with that of...
9928 (make_pass_lower_vector): ...new function.
9929 (pass_lower_vector_ssa): Replace declaration with that of...
9930 (make_pass_lower_vector_ssa): ...new function.
9931 (pass_lower_omp): Replace declaration with that of...
9932 (make_pass_lower_omp): ...new function.
9933 (pass_diagnose_omp_blocks): Replace declaration with that of...
9934 (make_pass_diagnose_omp_blocks): ...new function.
9935 (pass_expand_omp): Replace declaration with that of...
9936 (make_pass_expand_omp): ...new function.
9937 (pass_expand_omp_ssa): Replace declaration with that of...
9938 (make_pass_expand_omp_ssa): ...new function.
9939 (pass_object_sizes): Replace declaration with that of...
9940 (make_pass_object_sizes): ...new function.
9941 (pass_strlen): Replace declaration with that of...
9942 (make_pass_strlen): ...new function.
9943 (pass_fold_builtins): Replace declaration with that of...
9944 (make_pass_fold_builtins): ...new function.
9945 (pass_stdarg): Replace declaration with that of...
9946 (make_pass_stdarg): ...new function.
9947 (pass_early_warn_uninitialized): Replace declaration with that of...
9948 (make_pass_early_warn_uninitialized): ...new function.
9949 (pass_late_warn_uninitialized): Replace declaration with that of...
9950 (make_pass_late_warn_uninitialized): ...new function.
9951 (pass_cse_reciprocals): Replace declaration with that of...
9952 (make_pass_cse_reciprocals): ...new function.
9953 (pass_cse_sincos): Replace declaration with that of...
9954 (make_pass_cse_sincos): ...new function.
9955 (pass_optimize_bswap): Replace declaration with that of...
9956 (make_pass_optimize_bswap): ...new function.
9957 (pass_optimize_widening_mul): Replace declaration with that of...
9958 (make_pass_optimize_widening_mul): ...new function.
9959 (pass_warn_function_return): Replace declaration with that of...
9960 (make_pass_warn_function_return): ...new function.
9961 (pass_warn_function_noreturn): Replace declaration with that of...
9962 (make_pass_warn_function_noreturn): ...new function.
9963 (pass_cselim): Replace declaration with that of...
9964 (make_pass_cselim): ...new function.
9965 (pass_phiopt): Replace declaration with that of...
9966 (make_pass_phiopt): ...new function.
9967 (pass_forwprop): Replace declaration with that of...
9968 (make_pass_forwprop): ...new function.
9969 (pass_phiprop): Replace declaration with that of...
9970 (make_pass_phiprop): ...new function.
9971 (pass_tree_ifcombine): Replace declaration with that of...
9972 (make_pass_tree_ifcombine): ...new function.
9973 (pass_dse): Replace declaration with that of...
9974 (make_pass_dse): ...new function.
9975 (pass_nrv): Replace declaration with that of...
9976 (make_pass_nrv): ...new function.
9977 (pass_rename_ssa_copies): Replace declaration with that of...
9978 (make_pass_rename_ssa_copies): ...new function.
9979 (pass_sink_code): Replace declaration with that of...
9980 (make_pass_sink_code): ...new function.
9981 (pass_fre): Replace declaration with that of...
9982 (make_pass_fre): ...new function.
9983 (pass_check_data_deps): Replace declaration with that of...
9984 (make_pass_check_data_deps): ...new function.
9985 (pass_copy_prop): Replace declaration with that of...
9986 (make_pass_copy_prop): ...new function.
9987 (pass_vrp): Replace declaration with that of...
9988 (make_pass_vrp): ...new function.
9989 (pass_uncprop): Replace declaration with that of...
9990 (make_pass_uncprop): ...new function.
9991 (pass_return_slot): Replace declaration with that of...
9992 (make_pass_return_slot): ...new function.
9993 (pass_reassoc): Replace declaration with that of...
9994 (make_pass_reassoc): ...new function.
9995 (pass_rebuild_cgraph_edges): Replace declaration with that of...
9996 (make_pass_rebuild_cgraph_edges): ...new function.
9997 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
9998 (make_pass_remove_cgraph_callee_edges): ...new function.
9999 (pass_build_cgraph_edges): Replace declaration with that of...
10000 (make_pass_build_cgraph_edges): ...new function.
10001 (pass_local_pure_const): Replace declaration with that of...
10002 (make_pass_local_pure_const): ...new function.
10003 (pass_tracer): Replace declaration with that of...
10004 (make_pass_tracer): ...new function.
10005 (pass_warn_unused_result): Replace declaration with that of...
10006 (make_pass_warn_unused_result): ...new function.
10007 (pass_diagnose_tm_blocks): Replace declaration with that of...
10008 (make_pass_diagnose_tm_blocks): ...new function.
10009 (pass_lower_tm): Replace declaration with that of...
10010 (make_pass_lower_tm): ...new function.
10011 (pass_tm_init): Replace declaration with that of...
10012 (make_pass_tm_init): ...new function.
10013 (pass_tm_mark): Replace declaration with that of...
10014 (make_pass_tm_mark): ...new function.
10015 (pass_tm_memopt): Replace declaration with that of...
10016 (make_pass_tm_memopt): ...new function.
10017 (pass_tm_edges): Replace declaration with that of...
10018 (make_pass_tm_edges): ...new function.
10019 (pass_split_functions): Replace declaration with that of...
10020 (make_pass_split_functions): ...new function.
10021 (pass_feedback_split_functions): Replace declaration with that of...
10022 (make_pass_feedback_split_functions): ...new function.
10023 (pass_strength_reduction): Replace declaration with that of...
10024 (make_pass_strength_reduction): ...new function.
10025 (pass_ipa_lower_emutls): Replace declaration with that of...
10026 (make_pass_ipa_lower_emutls): ...new function.
10027 (pass_ipa_function_and_variable_visibility): Replace declaration with
10028 that of...
10029 (make_pass_ipa_function_and_variable_visibility): ...new function.
10030 (pass_ipa_tree_profile): Replace declaration with that of...
10031 (make_pass_ipa_tree_profile): ...new function.
10032 (pass_early_local_passes): Replace declaration with that of...
10033 (make_pass_early_local_passes): ...new function.
10034 (pass_ipa_whole_program_visibility): Replace declaration with that
10035 of...
10036 (make_pass_ipa_whole_program_visibility): ...new function.
10037 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10038 (make_pass_ipa_lto_gimple_out): ...new function.
10039 (pass_ipa_increase_alignment): Replace declaration with that of...
10040 (make_pass_ipa_increase_alignment): ...new function.
10041 (pass_ipa_inline): Replace declaration with that of...
10042 (make_pass_ipa_inline): ...new function.
10043 (pass_ipa_free_lang_data): Replace declaration with that of...
10044 (make_pass_ipa_free_lang_data): ...new function.
10045 (pass_ipa_free_inline_summary): Replace declaration with that of...
10046 (make_pass_ipa_free_inline_summary): ...new function.
10047 (pass_ipa_cp): Replace declaration with that of...
10048 (make_pass_ipa_cp): ...new function.
10049 (pass_ipa_reference): Replace declaration with that of...
10050 (make_pass_ipa_reference): ...new function.
10051 (pass_ipa_pure_const): Replace declaration with that of...
10052 (make_pass_ipa_pure_const): ...new function.
10053 (pass_ipa_pta): Replace declaration with that of...
10054 (make_pass_ipa_pta): ...new function.
10055 (pass_ipa_lto_finish_out): Replace declaration with that of...
10056 (make_pass_ipa_lto_finish_out): ...new function.
10057 (pass_ipa_tm): Replace declaration with that of...
10058 (make_pass_ipa_tm): ...new function.
10059 (pass_ipa_profile): Replace declaration with that of...
10060 (make_pass_ipa_profile): ...new function.
10061 (pass_ipa_cdtor_merge): Replace declaration with that of...
10062 (make_pass_ipa_cdtor_merge): ...new function.
10063 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10064 of...
10065 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10066 (pass_init_datastructures): Replace declaration with that of...
10067 (make_pass_init_datastructures): ...new function.
10068 (pass_fixup_cfg): Replace declaration with that of...
10069 (make_pass_fixup_cfg): ...new function.
10070 (pass_expand): Replace declaration with that of...
10071 (make_pass_expand): ...new function.
10072 (pass_instantiate_virtual_regs): Replace declaration with that of...
10073 (make_pass_instantiate_virtual_regs): ...new function.
10074 (pass_rtl_fwprop): Replace declaration with that of...
10075 (make_pass_rtl_fwprop): ...new function.
10076 (pass_rtl_fwprop_addr): Replace declaration with that of...
10077 (make_pass_rtl_fwprop_addr): ...new function.
10078 (pass_jump): Replace declaration with that of...
10079 (make_pass_jump): ...new function.
10080 (pass_jump2): Replace declaration with that of...
10081 (make_pass_jump2): ...new function.
10082 (pass_lower_subreg): Replace declaration with that of...
10083 (make_pass_lower_subreg): ...new function.
10084 (pass_cse): Replace declaration with that of...
10085 (make_pass_cse): ...new function.
10086 (pass_fast_rtl_dce): Replace declaration with that of...
10087 (make_pass_fast_rtl_dce): ...new function.
10088 (pass_ud_rtl_dce): Replace declaration with that of...
10089 (make_pass_ud_rtl_dce): ...new function.
10090 (pass_rtl_dce): Replace declaration with that of...
10091 (make_pass_rtl_dce): ...new function.
10092 (pass_rtl_dse1): Replace declaration with that of...
10093 (make_pass_rtl_dse1): ...new function.
10094 (pass_rtl_dse2): Replace declaration with that of...
10095 (make_pass_rtl_dse2): ...new function.
10096 (pass_rtl_dse3): Replace declaration with that of...
10097 (make_pass_rtl_dse3): ...new function.
10098 (pass_rtl_cprop): Replace declaration with that of...
10099 (make_pass_rtl_cprop): ...new function.
10100 (pass_rtl_pre): Replace declaration with that of...
10101 (make_pass_rtl_pre): ...new function.
10102 (pass_rtl_hoist): Replace declaration with that of...
10103 (make_pass_rtl_hoist): ...new function.
10104 (pass_rtl_store_motion): Replace declaration with that of...
10105 (make_pass_rtl_store_motion): ...new function.
10106 (pass_cse_after_global_opts): Replace declaration with that of...
10107 (make_pass_cse_after_global_opts): ...new function.
10108 (pass_rtl_ifcvt): Replace declaration with that of...
10109 (make_pass_rtl_ifcvt): ...new function.
10110 (pass_into_cfg_layout_mode): Replace declaration with that of...
10111 (make_pass_into_cfg_layout_mode): ...new function.
10112 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10113 (make_pass_outof_cfg_layout_mode): ...new function.
10114 (pass_loop2): Replace declaration with that of...
10115 (make_pass_loop2): ...new function.
10116 (pass_rtl_loop_init): Replace declaration with that of...
10117 (make_pass_rtl_loop_init): ...new function.
10118 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10119 (make_pass_rtl_move_loop_invariants): ...new function.
10120 (pass_rtl_unswitch): Replace declaration with that of...
10121 (make_pass_rtl_unswitch): ...new function.
10122 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10123 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10124 (pass_rtl_doloop): Replace declaration with that of...
10125 (make_pass_rtl_doloop): ...new function.
10126 (pass_rtl_loop_done): Replace declaration with that of...
10127 (make_pass_rtl_loop_done): ...new function.
10128 (pass_web): Replace declaration with that of...
10129 (make_pass_web): ...new function.
10130 (pass_cse2): Replace declaration with that of...
10131 (make_pass_cse2): ...new function.
10132 (pass_df_initialize_opt): Replace declaration with that of...
10133 (make_pass_df_initialize_opt): ...new function.
10134 (pass_df_initialize_no_opt): Replace declaration with that of...
10135 (make_pass_df_initialize_no_opt): ...new function.
10136 (pass_reginfo_init): Replace declaration with that of...
10137 (make_pass_reginfo_init): ...new function.
10138 (pass_inc_dec): Replace declaration with that of...
10139 (make_pass_inc_dec): ...new function.
10140 (pass_stack_ptr_mod): Replace declaration with that of...
10141 (make_pass_stack_ptr_mod): ...new function.
10142 (pass_initialize_regs): Replace declaration with that of...
10143 (make_pass_initialize_regs): ...new function.
10144 (pass_combine): Replace declaration with that of...
10145 (make_pass_combine): ...new function.
10146 (pass_if_after_combine): Replace declaration with that of...
10147 (make_pass_if_after_combine): ...new function.
10148 (pass_ree): Replace declaration with that of...
10149 (make_pass_ree): ...new function.
10150 (pass_partition_blocks): Replace declaration with that of...
10151 (make_pass_partition_blocks): ...new function.
10152 (pass_match_asm_constraints): Replace declaration with that of...
10153 (make_pass_match_asm_constraints): ...new function.
10154 (pass_regmove): Replace declaration with that of...
10155 (make_pass_regmove): ...new function.
10156 (pass_split_all_insns): Replace declaration with that of...
10157 (make_pass_split_all_insns): ...new function.
10158 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10159 (make_pass_fast_rtl_byte_dce): ...new function.
10160 (pass_lower_subreg2): Replace declaration with that of...
10161 (make_pass_lower_subreg2): ...new function.
10162 (pass_mode_switching): Replace declaration with that of...
10163 (make_pass_mode_switching): ...new function.
10164 (pass_sms): Replace declaration with that of...
10165 (make_pass_sms): ...new function.
10166 (pass_sched): Replace declaration with that of...
10167 (make_pass_sched): ...new function.
10168 (pass_ira): Replace declaration with that of...
10169 (make_pass_ira): ...new function.
10170 (pass_reload): Replace declaration with that of...
10171 (make_pass_reload): ...new function.
10172 (pass_clean_state): Replace declaration with that of...
10173 (make_pass_clean_state): ...new function.
10174 (pass_branch_prob): Replace declaration with that of...
10175 (make_pass_branch_prob): ...new function.
10176 (pass_value_profile_transformations): Replace declaration with that
10177 of...
10178 (make_pass_value_profile_transformations): ...new function.
10179 (pass_postreload_cse): Replace declaration with that of...
10180 (make_pass_postreload_cse): ...new function.
10181 (pass_gcse2): Replace declaration with that of...
10182 (make_pass_gcse2): ...new function.
10183 (pass_split_after_reload): Replace declaration with that of...
10184 (make_pass_split_after_reload): ...new function.
10185 (pass_branch_target_load_optimize1): Replace declaration with that
10186 of...
10187 (make_pass_branch_target_load_optimize1): ...new function.
10188 (pass_thread_prologue_and_epilogue): Replace declaration with that
10189 of...
10190 (make_pass_thread_prologue_and_epilogue): ...new function.
10191 (pass_stack_adjustments): Replace declaration with that of...
10192 (make_pass_stack_adjustments): ...new function.
10193 (pass_peephole2): Replace declaration with that of...
10194 (make_pass_peephole2): ...new function.
10195 (pass_if_after_reload): Replace declaration with that of...
10196 (make_pass_if_after_reload): ...new function.
10197 (pass_regrename): Replace declaration with that of...
10198 (make_pass_regrename): ...new function.
10199 (pass_cprop_hardreg): Replace declaration with that of...
10200 (make_pass_cprop_hardreg): ...new function.
10201 (pass_reorder_blocks): Replace declaration with that of...
10202 (make_pass_reorder_blocks): ...new function.
10203 (pass_branch_target_load_optimize2): Replace declaration with that
10204 of...
10205 (make_pass_branch_target_load_optimize2): ...new function.
10206 (pass_leaf_regs): Replace declaration with that of...
10207 (make_pass_leaf_regs): ...new function.
10208 (pass_split_before_sched2): Replace declaration with that of...
10209 (make_pass_split_before_sched2): ...new function.
10210 (pass_compare_elim_after_reload): Replace declaration with that of...
10211 (make_pass_compare_elim_after_reload): ...new function.
10212 (pass_sched2): Replace declaration with that of...
10213 (make_pass_sched2): ...new function.
10214 (pass_stack_regs): Replace declaration with that of...
10215 (make_pass_stack_regs): ...new function.
10216 (pass_stack_regs_run): Replace declaration with that of...
10217 (make_pass_stack_regs_run): ...new function.
10218 (pass_df_finish): Replace declaration with that of...
10219 (make_pass_df_finish): ...new function.
10220 (pass_compute_alignments): Replace declaration with that of...
10221 (make_pass_compute_alignments): ...new function.
10222 (pass_duplicate_computed_gotos): Replace declaration with that of...
10223 (make_pass_duplicate_computed_gotos): ...new function.
10224 (pass_variable_tracking): Replace declaration with that of...
10225 (make_pass_variable_tracking): ...new function.
10226 (pass_free_cfg): Replace declaration with that of...
10227 (make_pass_free_cfg): ...new function.
10228 (pass_machine_reorg): Replace declaration with that of...
10229 (make_pass_machine_reorg): ...new function.
10230 (pass_cleanup_barriers): Replace declaration with that of...
10231 (make_pass_cleanup_barriers): ...new function.
10232 (pass_delay_slots): Replace declaration with that of...
10233 (make_pass_delay_slots): ...new function.
10234 (pass_split_for_shorten_branches): Replace declaration with that of...
10235 (make_pass_split_for_shorten_branches): ...new function.
10236 (pass_split_before_regstack): Replace declaration with that of...
10237 (make_pass_split_before_regstack): ...new function.
10238 (pass_convert_to_eh_region_ranges): Replace declaration with that
10239 of...
10240 (make_pass_convert_to_eh_region_ranges): ...new function.
10241 (pass_shorten_branches): Replace declaration with that of...
10242 (make_pass_shorten_branches): ...new function.
10243 (pass_set_nothrow_function_flags): Replace declaration with that of...
10244 (make_pass_set_nothrow_function_flags): ...new function.
10245 (pass_dwarf2_frame): Replace declaration with that of...
10246 (make_pass_dwarf2_frame): ...new function.
10247 (pass_final): Replace declaration with that of...
10248 (make_pass_final): ...new function.
10249 (pass_rtl_seqabstr): Replace declaration with that of...
10250 (make_pass_rtl_seqabstr): ...new function.
10251 (pass_release_ssa_names): Replace declaration with that of...
10252 (make_pass_release_ssa_names): ...new function.
10253 (pass_early_inline): Replace declaration with that of...
10254 (make_pass_early_inline): ...new function.
10255 (pass_inline_parameters): Replace declaration with that of...
10256 (make_pass_inline_parameters): ...new function.
10257 (pass_update_address_taken): Replace declaration with that of...
10258 (make_pass_update_address_taken): ...new function.
10259 (pass_convert_switch): Replace declaration with that of...
10260 (make_pass_convert_switch): ...new function.
10261 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10262 to a subclass of simple_ipa_opt_pass along with...
10263 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10264 (make_pass_ipa_tree_profile): ...new function.
10265 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10266 subclass of gimple_opt_pass along with...
10267 (pass_data_sra_early): ...new pass_data instance and...
10268 (make_pass_sra_early): ...new function.
10269 (pass_sra): Convert from a global struct to a subclass of
10270 gimple_opt_pass along with...
10271 (pass_data_sra): ...new pass_data instance and...
10272 (make_pass_sra): ...new function.
10273 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10274 gimple_opt_pass along with...
10275 (pass_data_early_ipa_sra): ...new pass_data instance and...
10276 (make_pass_early_ipa_sra): ...new function.
10277 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10278 subclass of gimple_opt_pass along with...
10279 (pass_data_ccp): ...new pass_data instance and...
10280 (make_pass_ccp): ...new function.
10281 (pass_fold_builtins): Convert from a global struct to a subclass of
10282 gimple_opt_pass along with...
10283 (pass_data_fold_builtins): ...new pass_data instance and...
10284 (make_pass_fold_builtins): ...new function.
10285 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10286 subclass of gimple_opt_pass along with...
10287 (pass_data_copy_prop): ...new pass_data instance and...
10288 (make_pass_copy_prop): ...new function.
10289 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10290 global struct to a subclass of gimple_opt_pass along with...
10291 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10292 (make_pass_rename_ssa_copies): ...new function.
10293 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10294 subclass of gimple_opt_pass along with...
10295 (pass_data_dce): ...new pass_data instance and...
10296 (make_pass_dce): ...new function.
10297 (pass_dce_loop): Convert from a global struct to a subclass of
10298 gimple_opt_pass along with...
10299 (pass_data_dce_loop): ...new pass_data instance and...
10300 (make_pass_dce_loop): ...new function.
10301 (pass_cd_dce): Convert from a global struct to a subclass of
10302 gimple_opt_pass along with...
10303 (pass_data_cd_dce): ...new pass_data instance and...
10304 (make_pass_cd_dce): ...new function.
10305 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10306 subclass of gimple_opt_pass along with...
10307 (pass_data_dominator): ...new pass_data instance and...
10308 (make_pass_dominator): ...new function.
10309 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10310 gimple_opt_pass along with...
10311 (pass_data_phi_only_cprop): ...new pass_data instance and...
10312 (make_pass_phi_only_cprop): ...new function.
10313 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10314 subclass of gimple_opt_pass along with...
10315 (pass_data_dse): ...new pass_data instance and...
10316 (make_pass_dse): ...new function.
10317 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10318 a subclass of gimple_opt_pass along with...
10319 (pass_data_forwprop): ...new pass_data instance and...
10320 (make_pass_forwprop): ...new function.
10321 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10322 struct to a subclass of gimple_opt_pass along with...
10323 (pass_data_tree_ifcombine): ...new pass_data instance and...
10324 (make_pass_tree_ifcombine): ...new function.
10325 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10326 subclass of gimple_opt_pass along with...
10327 (pass_data_ch): ...new pass_data instance and...
10328 (make_pass_ch): ...new function.
10329 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10330 subclass of gimple_opt_pass along with...
10331 (pass_data_tree_loop): ...new pass_data instance and...
10332 (make_pass_tree_loop): ...new function.
10333 (pass_tree_loop_init): Convert from a global struct to a subclass of
10334 gimple_opt_pass along with...
10335 (pass_data_tree_loop_init): ...new pass_data instance and...
10336 (make_pass_tree_loop_init): ...new function.
10337 (pass_lim): Convert from a global struct to a subclass of
10338 gimple_opt_pass along with...
10339 (pass_data_lim): ...new pass_data instance and...
10340 (make_pass_lim): ...new function.
10341 (pass_tree_unswitch): Convert from a global struct to a subclass of
10342 gimple_opt_pass along with...
10343 (pass_data_tree_unswitch): ...new pass_data instance and...
10344 (make_pass_tree_unswitch): ...new function.
10345 (pass_predcom): Convert from a global struct to a subclass of
10346 gimple_opt_pass along with...
10347 (pass_data_predcom): ...new pass_data instance and...
10348 (make_pass_predcom): ...new function.
10349 (pass_vectorize): Convert from a global struct to a subclass of
10350 gimple_opt_pass along with...
10351 (pass_data_vectorize): ...new pass_data instance and...
10352 (make_pass_vectorize): ...new function.
10353 (pass_graphite): Convert from a global struct to a subclass of
10354 gimple_opt_pass along with...
10355 (pass_data_graphite): ...new pass_data instance and...
10356 (make_pass_graphite): ...new function.
10357 (pass_graphite_transforms): Convert from a global struct to a subclass
10358 of gimple_opt_pass along with...
10359 (pass_data_graphite_transforms): ...new pass_data instance and...
10360 (make_pass_graphite_transforms): ...new function.
10361 (pass_check_data_deps): Convert from a global struct to a subclass of
10362 gimple_opt_pass along with...
10363 (pass_data_check_data_deps): ...new pass_data instance and...
10364 (make_pass_check_data_deps): ...new function.
10365 (pass_iv_canon): Convert from a global struct to a subclass of
10366 gimple_opt_pass along with...
10367 (pass_data_iv_canon): ...new pass_data instance and...
10368 (make_pass_iv_canon): ...new function.
10369 (pass_scev_cprop): Convert from a global struct to a subclass of
10370 gimple_opt_pass along with...
10371 (pass_data_scev_cprop): ...new pass_data instance and...
10372 (make_pass_scev_cprop): ...new function.
10373 (pass_record_bounds): Convert from a global struct to a subclass of
10374 gimple_opt_pass along with...
10375 (pass_data_record_bounds): ...new pass_data instance and...
10376 (make_pass_record_bounds): ...new function.
10377 (pass_complete_unroll): Convert from a global struct to a subclass of
10378 gimple_opt_pass along with...
10379 (pass_data_complete_unroll): ...new pass_data instance and...
10380 (make_pass_complete_unroll): ...new function.
10381 (pass_complete_unrolli): Convert from a global struct to a subclass of
10382 gimple_opt_pass along with...
10383 (pass_data_complete_unrolli): ...new pass_data instance and...
10384 (make_pass_complete_unrolli): ...new function.
10385 (pass_parallelize_loops): Convert from a global struct to a subclass
10386 of gimple_opt_pass along with...
10387 (pass_data_parallelize_loops): ...new pass_data instance and...
10388 (make_pass_parallelize_loops): ...new function.
10389 (pass_loop_prefetch): Convert from a global struct to a subclass of
10390 gimple_opt_pass along with...
10391 (pass_data_loop_prefetch): ...new pass_data instance and...
10392 (make_pass_loop_prefetch): ...new function.
10393 (pass_iv_optimize): Convert from a global struct to a subclass of
10394 gimple_opt_pass along with...
10395 (pass_data_iv_optimize): ...new pass_data instance and...
10396 (make_pass_iv_optimize): ...new function.
10397 (pass_tree_loop_done): Convert from a global struct to a subclass of
10398 gimple_opt_pass along with...
10399 (pass_data_tree_loop_done): ...new pass_data instance and...
10400 (make_pass_tree_loop_done): ...new function.
10401 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10402 struct to a subclass of gimple_opt_pass along with...
10403 (pass_data_cse_reciprocals): ...new pass_data instance and...
10404 (make_pass_cse_reciprocals): ...new function.
10405 (pass_cse_sincos): Convert from a global struct to a subclass of
10406 gimple_opt_pass along with...
10407 (pass_data_cse_sincos): ...new pass_data instance and...
10408 (make_pass_cse_sincos): ...new function.
10409 (pass_optimize_bswap): Convert from a global struct to a subclass of
10410 gimple_opt_pass along with...
10411 (pass_data_optimize_bswap): ...new pass_data instance and...
10412 (make_pass_optimize_bswap): ...new function.
10413 (pass_optimize_widening_mul): Convert from a global struct to a
10414 subclass of gimple_opt_pass along with...
10415 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10416 (make_pass_optimize_widening_mul): ...new function.
10417 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10418 subclass of gimple_opt_pass along with...
10419 (pass_data_phiopt): ...new pass_data instance and...
10420 (make_pass_phiopt): ...new function.
10421 (pass_cselim): Convert from a global struct to a subclass of
10422 gimple_opt_pass along with...
10423 (pass_data_cselim): ...new pass_data instance and...
10424 (make_pass_cselim): ...new function.
10425 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10426 subclass of gimple_opt_pass along with...
10427 (pass_data_phiprop): ...new pass_data instance and...
10428 (make_pass_phiprop): ...new function.
10429 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10430 subclass of gimple_opt_pass along with...
10431 (pass_data_pre): ...new pass_data instance and...
10432 (make_pass_pre): ...new function.
10433 (pass_fre): Convert from a global struct to a subclass of
10434 gimple_opt_pass along with...
10435 (pass_data_fre): ...new pass_data instance and...
10436 (make_pass_fre): ...new function.
10437 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10438 subclass of gimple_opt_pass along with...
10439 (pass_data_reassoc): ...new pass_data instance and...
10440 (make_pass_reassoc): ...new function.
10441 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10442 subclass of gimple_opt_pass along with...
10443 (pass_data_sink_code): ...new pass_data instance and...
10444 (make_pass_sink_code): ...new function.
10445 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10446 subclass of gimple_opt_pass along with...
10447 (pass_data_strlen): ...new pass_data instance and...
10448 (make_pass_strlen): ...new function.
10449 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10450 struct to a subclass of gimple_opt_pass along with...
10451 (pass_data_build_alias): ...new pass_data instance and...
10452 (make_pass_build_alias): ...new function.
10453 (pass_build_ealias): Convert from a global struct to a subclass of
10454 gimple_opt_pass along with...
10455 (pass_data_build_ealias): ...new pass_data instance and...
10456 (make_pass_build_ealias): ...new function.
10457 (pass_ipa_pta): Convert from a global struct to a subclass of
10458 simple_ipa_opt_pass along with...
10459 (pass_data_ipa_pta): ...new pass_data instance and...
10460 (make_pass_ipa_pta): ...new function.
10461 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10462 subclass of gimple_opt_pass along with...
10463 (pass_data_uncprop): ...new pass_data instance and...
10464 (make_pass_uncprop): ...new function.
10465 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10466 global struct to a subclass of gimple_opt_pass along with...
10467 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10468 (make_pass_late_warn_uninitialized): ...new function.
10469 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10470 to a subclass of gimple_opt_pass along with...
10471 (pass_data_init_datastructures): ...new pass_data instance and...
10472 (make_pass_init_datastructures): ...new function.
10473 (pass_early_warn_uninitialized): Convert from a global struct to a
10474 subclass of gimple_opt_pass along with...
10475 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10476 (make_pass_early_warn_uninitialized): ...new function.
10477 (pass_update_address_taken): Convert from a global struct to a
10478 subclass of gimple_opt_pass along with...
10479 (pass_data_update_address_taken): ...new pass_data instance and...
10480 (make_pass_update_address_taken): ...new function.
10481 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10482 struct to a subclass of gimple_opt_pass along with...
10483 (pass_data_release_ssa_names): ...new pass_data instance and...
10484 (make_pass_release_ssa_names): ...new function.
10485 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10486 subclass of gimple_opt_pass along with...
10487 (pass_data_stdarg): ...new pass_data instance and...
10488 (make_pass_stdarg): ...new function.
10489 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10490 global struct to a subclass of gimple_opt_pass along with...
10491 (pass_data_convert_switch): ...new pass_data instance and...
10492 (make_pass_convert_switch): ...new function.
10493 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10494 to a subclass of gimple_opt_pass along with...
10495 (pass_data_tail_recursion): ...new pass_data instance and...
10496 (make_pass_tail_recursion): ...new function.
10497 (pass_tail_calls): Convert from a global struct to a subclass of
10498 gimple_opt_pass along with...
10499 (pass_data_tail_calls): ...new pass_data instance and...
10500 (make_pass_tail_calls): ...new function.
10501 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10502 struct to a subclass of gimple_opt_pass along with...
10503 (pass_data_lower_vector): ...new pass_data instance and...
10504 (make_pass_lower_vector): ...new function.
10505 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10506 gimple_opt_pass along with...
10507 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10508 (make_pass_lower_vector_ssa): ...new function.
10509 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10510 to a subclass of gimple_opt_pass along with...
10511 (pass_data_slp_vectorize): ...new pass_data instance and...
10512 (make_pass_slp_vectorize): ...new function.
10513 (pass_ipa_increase_alignment): Convert from a global struct to a
10514 subclass of simple_ipa_opt_pass along with...
10515 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10516 (make_pass_ipa_increase_alignment): ...new function.
10517 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10518 gimple_opt_pass along with...
10519 (pass_data_vrp): ...new pass_data instance and...
10520 (make_pass_vrp): ...new function.
10521 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10522 subclass of simple_ipa_opt_pass along with...
10523 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10524 (make_pass_ipa_free_lang_data): ...new function.
10525 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10526 gimple_opt_pass along with...
10527 (pass_data_tsan): ...new pass_data instance and...
10528 (make_pass_tsan): ...new function.
10529 (pass_tsan_O0): Convert from a global struct to a subclass of
10530 gimple_opt_pass along with...
10531 (pass_data_tsan_O0): ...new pass_data instance and...
10532 (make_pass_tsan_O0): ...new function.
10533 * var-tracking.c (pass_variable_tracking): Convert from a global
10534 struct to a subclass of rtl_opt_pass along with...
10535 (pass_data_variable_tracking): ...new pass_data instance and...
10536 (make_pass_variable_tracking): ...new function.
10537 * web.c (pass_web): Convert from a global struct to a subclass of
10538 rtl_opt_pass along with...
10539 (pass_data_web): ...new pass_data instance and...
10540 (make_pass_web): ...new function.
10541 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10542 declaration with that of...
10543 (make_pass_mode_switch_use): ...new function.
10544 (pass_resolve_sw_modes): Replace declaration with that of...
10545 (make_pass_resolve_sw_modes): ...new function.
10546 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10547 from a global struct to a subclass of rtl_opt_pass along with...
10548 (pass_data_mode_switch_use): ...new pass_data instance and...
10549 (make_pass_mode_switch_use): ...new function.
10550 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10551 from a global struct to a subclass of rtl_opt_pass along with...
10552 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10553 (make_pass_resolve_sw_modes): ...new function.
10554 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10555 struct to a subclass of rtl_opt_pass along with...
10556 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10557 (make_pass_insert_vzeroupper): ...new function.
10558 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10559 global struct to a subclass of rtl_opt_pass along with...
10560 (pass_data_work_around_errata): ...new pass_data instance and...
10561 (make_pass_work_around_errata): ...new function.
10562 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10563 struct to a subclass of rtl_opt_pass along with...
10564 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10565 (make_pass_mips_machine_reorg2): ...new function.
10566
6a389ed5
DM		 105672013-08-05 David Malcolm <dmalcolm@redhat.com>
10568
10569 * passes.c (pass_manager::operator new): New.
10570
f7695dbf
DM		 105712013-08-05 David Malcolm <dmalcolm@redhat.com>
10572
10573 Handwritten part of conversion of passes to C++ classes.
10574
10575 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10576 (toplev.o): Add dep on PASS_MANAGER_H.
10577 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10578 of early local pases to reflect this moving from a global to a
10579 member of gcc::pass_manager.
10580 (cgraph_add_new_function): Likewise.
10581 * lto-cgraph.c (lto_output_node): Update for conversion of
10582 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10583 * passes.c (opt_pass::clone): New.
10584 (opt_pass::gate): New.
10585 (opt_pass::execute): New.
10586 (opt_pass::opt_pass): New.
10587 (pass_manager::execute_early_local_passes): New.
10588 (pass_manager::execute_pass_mode_switching): new.
10589 (finish_optimization_passes): Convert to...
10590 (pass_manager::finish_optimization_passes): ...this.
10591 (finish_optimization_passes): Update for conversion of passes to
10592 C++ classes.
10593 (register_dump_files_1): Use has_gate since we cannot portably
10594 check a vtable entry against NULL.
10595 (dump_one_pass): Likewise.
10596 (ipa_write_summaries_2): Likewise.
10597 (ipa_write_optimization_summaries_1): Likewise.
10598 (ipa_read_summaries_1): Likewise.
10599 (ipa_read_optimization_summaries_1): Likewise.
10600 (execute_ipa_stmt_fixups): Likewise.
10601 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10602 pass-creation functions rather than wiring up globals, and
10603 storing the results in fields of pass_manager generated using
10604 pass-instances.def.
10605 (pass_manager::dump_profile_report): Update for conversion of
10606 passes to C++ classes.
10607 (pass_manager::execute_ipa_summary_passes): Likewise.
10608 (execute_one_ipa_transform_pass): Likewise.
10609 (execute_one_pass): Use has_gate and has_execute since we cannot
10610 portably check a vtable entry against NULL.
10611 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10612 (pass_manager): Use pass-instances.def to add fields for the
10613 various pass instances.
10614 * toplev.c (finalize): Update for move of
10615 finish_optimization_passes to a method of gcc::pass_manager.
10616 * toplev.h (finish_optimization_passes): Move to method of class
10617 pass_manager.
10618 * tree-pass.h (struct pass_data): New.
03b0ee0a 10619 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 10620 (opt_pass::gate): Convert to virtual function.
10621 (opt_pass::~opt_pass): New.
10622 (opt_pass::clone): New.
10623 (opt_pass::execute): Convert to virtual function.
10624 (opt_pass::opt_pass): New.
10625 (opt_pass::ctxt_): new.
10626 (gimple_opt_pass): Convert to subclass of opt_pass.
10627 (gimple_opt_pass::gimple_opt_pass): New.
10628 (rtl_opt_pass): Convert to subclass of opt_pass.
10629 (rtl_opt_pass::rtl_opt_pass): New.
10630 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10631 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10632 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10633 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10634 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10635 invocation of pass_mode_switching to reflect this moving from a
10636 global to a member of gcc::pass_manager.
10637 (ix86_option_override): Rework how pass_insert_vzeroupper is
10638 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 10639 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 10640
d0b2f831
RE		 106412013-08-05 Richard Earnshaw <rearnsha@arm.com>
10642
10643 PR rtl-optimization/57708
10644 * recog.c (peep2_find_free_register): Validate all regs in a
10645 multi-reg mode.
10646
51a5c0c2
JH		 106472013-08-05 Jan Hubicka <jh@suse.cz>
10648
10649 PR lto/57602
03b0ee0a
UB		 10650 * cgraph.c (verify_cgraph_node): Accept local flags from other
10651 partitions.
51a5c0c2
JH		 10652 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10653 (function_and_variable_visibility): Likewise.
10654 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10655
65f0a120
GDR		 106562013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10657
10658 * graph.c (init_graph_slim_pretty_print): Remove.
10659 (print_graph_cfg): Do not call it. Use local pretty printer.
10660 (start_graph_dump): Likewise.
10661
b3f80694
GDR		 106622013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10663
10664 * gimple-pretty-print.c (buffer): Remove.
10665 (initialized): Likewise.
10666 (maybe_init_pretty_print): Likewise.
10667 (print_gimple_stmt): Do not call it. Use non-static local
10668 pretty_printer variable.
10669 (print_gimple_expr): Likewise.
10670 (print_gimple_seq): Likewise.
10671 (gimple_dump_bb): Likewise.
10672
11a877b3
GDR		 106732013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10674
10675 * asan.c (asan_pp): Remove.
10676 (asan_pp_initialized): Likewise.
10677 (asan_pp_initialize): Likewise.
10678 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10679 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10680 (asan_add_global): Likewise.
10681
b066401f
GDR		 106822013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10683
10684 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10685 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10686 * pretty-print.c (pp_formatted_text_data): Likewise.
10687 (pp_indent): Rename from pp_base_indent.
10688 (pp_format): Rename from pp_base_format.
10689 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10690 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10691 (pp_flush): Rename from pp_base_flush.
10692 (pp_set_line_maximum_length): Rename from
10693 pp_base_set_line_maximum_length.
10694 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10695 (pp_set_prefix): Rename from pp_base_set_prefix.
10696 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10697 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10698 (pp_append_text): Rename from pp_base_append_text.
10699 (pp_formatted_text): Rename from pp_base_formatted_text.
10700 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10701 (pp_remaining_character_count_for_line): Rename from
10702 pp_base_remaining_character_count_for_line.
10703 (pp_newline): Rename from pp_base_newline.
10704 (pp_character): Rename from pp_base_character.
10705 (pp_string): Rename from pp_base_string.
10706 (pp_maybe_space): Rename from pp_base_maybe_space.
10707 * asan.c (asan_pp_string): Adjust.
10708 (asan_emit_stack_protection): Likewise.
10709 (asan_add_global): Likewise.
10710 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10711 * tree-mudflap.c (mf_varname_tree): Likewise.
10712 * tree-pretty-print.c (pp_tree_identifier): Rename from
10713 pp_base_tree_identifier.
10714 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10715 Declare as function.
10716
137a1a27
GDR		 107172013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10718
10719 * pretty-print.h (pp_bar_bar): New.
10720 (pp_ampersand_ampersand): Likewise.
10721 (pp_less_equal): Likewise.
10722 (pp_greater_equal): Likewise.
10723 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10724 printer functions instead of pp_string or operators and punctuators.
10725 (dump_gimple_call): Likewise.
10726 (dump_gimple_omp_for): Likewise.
10727 (dump_gimple_transaction): Likewise.
10728 (dump_gimple_phi): Likewise.
10729 (pp_gimple_stmt_1): Likewise.
10730 * sched-vis.c (print_insn): Likewise.
10731 * tree-mudflap.c (mf_varname_tree): Likewise.
10732 * tree-pretty-print.c (dump_block_node): Likewise.
10733 (dump_generic_node): Likewise.
10734
815effe1
JH		 107352013-08-02 Jan Hubicka <jh@suse.cz>
10736
10737 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10738 boundaries.
10739 * lto-streamer-out.c (tree_is_indexable): Results decls and
10740 parm decls are not indexable.
10741 (DFS_write_tree_body): Do not follow args and results.
10742 (hash_tree): Likewise.
10743 (output_functions): Rearrange so struct function is needed
10744 only when real body is output; be able to also ouptut abstract
10745 functions; output DECL_ARGUMENTS and DECL_RESULT.
10746 (lto_output): When not in WPA, ale store abstract functions.
10747 (write_symbol): Do not care about RESULT_DECL.
10748 (output_symbol_p): Handle correctly sbtract decls.
10749 * lto-streamer-in.c (input_function): Rearrange so struct
10750 function can be NULL at entry; allow streaming of
10751 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10752 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10753 sanity check during LTO.
10754 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10755 RESULT_DECl and DECL_ARGUMENTS.
10756 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10757 Likewise.
10758
07838b13
GDR		 107592013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10760
10761 * pretty-print.h (pp_underscore): New.
10762 (pp_comma): Tidy.
10763 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10764 printer functions instead of pp_character.
10765 (dump_binary_rhs): Likewise.
10766 (dump_ternary_rhs): Likewise.
10767 (dump_gimple_call_args): Likewise.
10768 (pp_points_to_solution): Likewise.
10769 (dump_gimple_call): Likewise.
10770 (dump_gimple_switch): Likewise.
10771 (dump_gimple_cond): Likewise.
10772 (dump_gimple_bind): Likewise.
10773 (dump_gimple_try): Likewise.
10774 (dump_gimple_omp_for): Likewise.
10775 (dump_gimple_omp_continue): Likewise.
10776 (dump_gimple_omp_single): Likewise.
10777 (dump_gimple_omp_sections): Likewise.
10778 (dump_gimple_omp_block): Likewise.
10779 (dump_gimple_omp_critical): Likewise.
10780 (dump_gimple_transaction): Likewise.
10781 (dump_gimple_asm): Likewise.
10782 (dump_gimple_phi): Likewise.
10783 (dump_gimple_omp_parallel): Likewise.
10784 (dump_gimple_omp_task): Likewise.
10785 (dump_gimple_omp_atomic_load): Likewise.
10786 (dump_gimple_omp_atomic_store): Likewise.
10787 (dump_gimple_mem_ops): Likewise.
10788 (pp_gimple_stmt_1): Likewise.
10789 (pp_cfg_jump): Likewise.
10790 (dump_implicit_edges): Likewise.
10791 (gimple_dump_bb_for_graph): Likewise.
10792 * graph.c (draw_cfg_node): Likewise.
10793 * langhooks.c (lhd_print_error_function): Likewise.
10794 * sched-vis.c (print_exp): Likewise.
10795 (print_value): Likewise.
10796 (print_pattern): Likewise.
10797 (print_insn): Likewise.
10798 (rtl_dump_bb_for_graph): Likewise.
10799 * tree-pretty-print.c (dump_function_declaration): Likewise.
10800 (dump_array_domain): Likewise.
10801 (dump_omp_clause): Likewise.
10802 (dump_location): Likewise.
10803 (dump_generic_node): Likewise.
10804 (print_struct_decl): Likewise.
10805 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10806
bb0d2039
BS		 108072013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10808
10809 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10810 candidate table when replacing a candidate statement.
10811 (replace_rhs_if_not_dup): Likewise.
10812 (replace_one_candidate): Likewise.
10813
bc0ec027 108142013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 10815 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 10816
10817 * cgraphunit.c (add_new_function): Fix logic when adding from
10818 late IPA pass.
10819 (assemble_thunk): Rename to ...
03b0ee0a
UB		 10820 (expand_thunk); .. this one; export; get it working with
10821 general functions; make produced gimple valid.
bc0ec027
JH		 10822 * cgraph.h (expand_thunk): Declare.
10823
0e8853ee
JH		 108242013-08-02 Jan Hubicka <jh@suse.cz>
10825
03b0ee0a
UB		 10826 * ipa-cp.c (gather_context_independent_values): Use
10827 ipa_get_param_move_cost.
10828 (get_replacement_map): Remove PARAM; move parameter folding
10829 into tree-inline.c
0e8853ee
JH		 10830 (create_specialized_node): Update.
10831 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10832 assert that we have gimple body; update move_cost.
10833 (count_formal_params): Assert that we have gimple body.
10834 (ipa_dump_param): New function.
10835 (ipa_alloc_node_params): Break out from ...
10836 (ipa_initialize_node_params): ... here.
10837 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10838 (ipa_write_node_info): Stream move costs.
10839 (ipa_read_node_info): Read move costs.
10840 (ipa_update_after_lto_read): Do not recompute node params.
10841 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10842 (ipa_get_param): Check we are not in WPA.
10843 (ipa_get_param_move_cost): New.
10844 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10845 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10846 parm numbers to be present.
10847
ecee672b
VM		 108482013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10849
10850 PR rtl-optimization/58048
10851 * lra-constraints.c (process_alt_operands): Don't check asm
10852 operand on register.
10853
13ec0527
EB		 108542013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10855
10856 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10857 the implied StoreLoad barrier for atomic operations if before.
10858
af15184a
JH		 108592013-08-02 Jan Hubicka <jh@suse.cz>
10860 Martin Liska <marxin.liska@gmail.com>
10861
03b0ee0a
UB		 10862 * cgraph.c (cgraph_function_body_availability): Do not check
10863 cgraph flags.
af15184a
JH		 10864 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10865 symtab_node_availability): Declare.
10866 * ipa.c (can_replace_by_local_alias): New.
10867 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 10868 * symtab.c (symtab_for_node_and_aliases,
10869 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 10870
01e54ef8
VM		 108712013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10872
10873 PR rtl-optimization/57963
03b0ee0a 10874 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 10875 (lra_constraints): Use them.
10876
e2fd7ca7
SN		 108772013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10878
03b0ee0a
UB		 10879 * config/arm/types.md (define_attr "type"): Add "load_acq"
10880 and "store_rel".
e2fd7ca7
SN		 10881 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10882 changes.
10883 (cortex_a53_store1): Likewise.
10884
0dcc26c3
JH		 108852013-08-01 Jan Hubicka <jh@suse.cz>
10886
03b0ee0a
UB		 10887 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10888 partitions are not needed.
0dcc26c3 10889
0bd72901
UB		 108902013-08-01 Uros Bizjak <ubizjak@gmail.com>
10891
10892 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10893 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10894 MAYBE_NON_Q_CLASS_P where appropriate.
10895
12123452
JH		 108962013-08-01 Jan Hubicka <jh@suse.cz>
10897
10898 * cgraph.h (release_function_body): Declare.
10899 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10900 of unused delcarations.
10901
ce852f9c
KT		 109022013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10903
03b0ee0a
UB		 10904 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10905 RTL form when subtracting a constant.
ce852f9c 10906
c743b246
KT		 109072013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10908
10909 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10910 Generate canonical plus rtx with negated immediate instead of minus
10911 where appropriate.
10912 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10913
c0c123ef
JH		 109142013-08-01 Jan Hubicka <jh@suse.cz>
10915
10916 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10917 (cgraph_release_function_body): Likewise.
10918 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10919 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10920 to used_as_abstract_origin.
10921 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10922 symbols used as abstract origins.
10923 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 10924 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10925 used_as_abstract_origin.
c0c123ef 10926 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 10927 used_as_abstract_origin; be ready for DECL_RESULT and
10928 DECL_ARGUMENTS to be NULL.
c0c123ef 10929
0bd72901
UB		 10930 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
10931 for abstract functions.
10932 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
10933 real symbols.
c0c123ef 10934
1f029433
JH		 109352013-08-01 Jan Hubicka <jh@suse.cz>
10936
10937 * profile.c (compute_value_histograms): Fix thinko.
10938
c451f4d6
SN		 109392013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10940
10941 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
10942 aarch-common-protos.h to extra_headers.
10943 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10944 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
10945 * config/aarch64/t-aarch64 (aarch-common.o): Define.
10946
34b512f5
SN		 109472013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10948
10949 * config/aarch64/aarch64.md (define_attr "type"): Delete.
10950 Include "../arm/types.md". Define "type" attribute for all patterns.
10951 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
10952 attribute changes.
10953
d86e633a
MM		 109542013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
10955
10956 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
10957 to support power8 load fusion.
10958 (fusion_gpr_mem_load): Likewise.
10959
10960 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
10961
10962 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
10963 declarations for power8 load fusion.
10964 (emit_fusion_gpr_load): Likewise.
10965
10966 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
10967 tuning for power8, turn on fusion mode by default. Turn on sign
10968 extending fusion mode if normal fusion mode is on, and we are at
10969 -O2 or -O3.
10970 (fusion_gpr_load_p): New function, return true if we can fuse an
10971 addis instruction with a dependent load to a GPR.
10972 (emit_fusion_gpr_load): Emit the instructions for power8 load
10973 fusion to GPRs.
10974
0bd72901 10975 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 10976 (VSX load fusion peepholes): New peepholes to fuse together an
10977 addi instruction with a VSX load instruction.
10978
10979 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
10980 peepholes to fuse an addis instruction with a load to a GPR base
10981 register. If we are supporting sign extending fusions, convert
10982 sign extending loads to zero extending loads and add an explicit
10983 sign extension.
10984
c3f35647
SN		 109852013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10986
10987 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
10988 aarch-common-protos.h to extra_headers.
10989 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10990 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
10991 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 10992 (arm_no_early_alu_shift_dep): Likewise.
10993 (arm_no_early_alu_shift_value_dep): Likewise.
10994 (arm_no_early_mul_dep): Likewise.
10995 (arm_no_early_store_addr_dep): Likewise.
10996 (arm_mac_accumulator_is_mul_result): Likewise.
10997 (arm_mac_accumulator_is_result): Likewise.
c3f35647 10998 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 10999 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11000 here to ...
c3f35647 11001 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 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.
c3f35647
SN		 11008 * config/arm/aarch-common-protos.h: ... here. New file.
11009 * config/arm/t-arm (aarch-common.o): Define.
11010
110112013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11012
11013 * config/arm/arm.md: Include new file "types.md".
11014 (define_attr "type"): Move from here to ...
11015 (define_attr "mul32"): Likewise.
11016 (define_attr "mul64"): Likewise.
11017 * config/arm/types.md: ... here. New file.
11018
228c1313
SH		 110192013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11020
11021 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11022 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11023
d6e9046d 110242013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 11025
11026 * gen-pass-instances.awk: Fix offset of substr().
11027
a167b052
DM		 110282013-07-31 David Malcolm <dmalcolm@redhat.com>
11029
11030 * Makefile.in (pass-instances.def): New.
11031 (passes.o): Replace dependency on passes.def with one on
11032 pass-instances.def
11033
11034 * gen-pass-instances.awk: New.
11035
11036 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11037 rather than passes.def, updating local definition of NEXT_PASS
11038 macro to add an extra NUM parameter (currently unused).
11039
315f8c0e
DM		 110402013-07-30 David Malcolm <dmalcolm@redhat.com>
11041
11042 * Makefile.in (PASS_MANAGER_H): New.
11043 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11044 (passes.o): Likewise.
11045 (statistics.o): Likewise.
11046 (cgraphunit.o): Likewise.
11047 (context.o): Depend on PASS_MANAGER_H.
11048
11049 * pass_manager.h: New.
11050
11051 * cgraphunit.c (cgraph_add_new_function): Update for moves
11052 of globals to fields of pass_manager.
11053 (analyze_function): Likewise.
11054 (expand_function): Likewise.
11055 (ipa_passes): Likewise.
11056 (compile): Likewise.
11057
11058 * context.c (context::context): New.
11059 * context.h (context::context): New.
11060 (context::get_passes): New.
11061 (context::passes_): New.
11062
11063 * lto-cgraph.c (input_node): Update for moves of globals to
11064 fields of pass_manager.
11065
11066 * passes.c (all_passes): Remove, in favor of a field of the
11067 same name within the new class pass_manager.
11068 (all_small_ipa_passes): Likewise.
11069 (all_lowering_passes): Likewise.
11070 (all_regular_ipa_passes): Likewise.
11071 (all_late_ipa_passes): Likewise.
11072 (all_lto_gen_passes): Likewise.
11073 (passes_by_id): Likewise.
11074 (passes_by_id_size): Likewise.
11075 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11076 the new class pass_manager.
11077 (set_pass_for_id): Convert to...
11078 (pass_manager::set_pass_for_id): ...method.
11079 (get_pass_for_id): Convert to...
11080 (pass_manager::get_pass_for_id): ...method.
11081 (register_one_dump_file): Move body of implementation into...
11082 (pass_manager::register_one_dump_file): ...here.
11083 (register_dump_files_1): Convert to...
11084 (pass_manager::register_dump_files_1): ...method.
11085 (register_dump_files): Convert to...
11086 (pass_manager::register_dump_files): ...method.
11087 (create_pass_tab): Update for moves of globals to fields of
11088 pass_manager.
11089 (dump_passes): Move body of implementation into...
11090 (pass_manager::dump_passes): ...here.
11091 (register_pass): Move body of implementation into...
11092 (pass_manager::register_pass): ...here.
11093 (init_optimization_passes): Convert into...
11094 (pass_manager::pass_manager): ...constructor for new
11095 pass_manager class, and initialize the pass_lists array.
11096 (check_profile_consistency): Update for moves of globals to
11097 fields of pass_manager.
11098 (dump_profile_report): Move body of implementation into...
11099 (pass_manager::dump_profile_report): ...here.
11100 (ipa_write_summaries_1): Update for moves of pass lists from
11101 being globals to fields of pass_manager.
11102 (ipa_write_optimization_summaries): Likewise.
11103 (ipa_read_summaries): Likewise.
11104 (ipa_read_optimization_summaries): Likewise.
11105 (execute_all_ipa_stmt_fixups): Likewise.
11106
11107 * statistics.c (statistics_fini): Update for moves of globals to
11108 fields of pass_manager.
11109
11110 * toplev.c (general_init): Replace call to
11111 init_optimization_passes with construction of the pass_manager
11112 instance.
11113
11114 * tree-pass.h (all_passes): Remove, in favor of a field of the
11115 same name within the new class pass_manager.
11116 (all_small_ipa_passes): Likewise.
11117 (all_lowering_passes): Likewise.
11118 (all_regular_ipa_passes): Likewise.
11119 (all_lto_gen_passes): Likewise.
11120 (all_late_ipa_passes): Likewise.
11121 (passes_by_id): Likewise.
11122 (passes_by_id_size): Likewise.
11123 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11124 the new class pass_manager.
11125 (get_pass_for_id): Remove.
11126
02cab1c1
RE		 111272013-07-30 Richard Earnshaw <rearnsha@arm.com>
11128
11129 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11130 configs.
11131
b2d3f886
RE		 111322013-07-30 Richard Earnshaw <rearnsha@arm.com>
11133
11134 * arm.md (mulhi3): New expand pattern.
11135
605e86fa
JH		 111362013-07-30 Jan Hubicka <jh@suse.cz>
11137 Martin Liska <marxin.liska@gmail.com>
11138
11139 * profile.c (compute_value_histograms): Do not ICE when
11140 there is mismatch only on some counters.
11141
14a87636
ZC		 111422013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11143
11144 PR rtl-optimization/57637
11145 * function.c (move_insn_for_shrink_wrap): Also check the
11146 GEN set of the LIVE problem for the liveness analysis
11147 if it exists, otherwise give up.
11148
0100cd3f
BS		 111492013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11150
11151 PR tree-optimization/57993
11152 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11153 replaced statement in the candidate table.
11154 (phi_add_costs): Return infinite cost when the hidden basis does
11155 not dominate all phis on which the candidate is dependent.
11156 (replace_one_candidate): Record replaced statement in the
11157 candidate table.
11158
f0d811f7
JR		 111592013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11160
11161 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11162 (ashlv2si3): New expander.
11163 (*ashlv2si3_i): New define_insn_and_split.
11164 * predicates.md (float_operation): Allow patterns with three
11165 basic sub-patterns.
11166
bba33211
JR		 11167 PR rtl-optimization/58021
11168 * mode-switching.c (create_pre_exit): Always split off preceding
11169 insns if we are not at the basic block head.
11170
ff3f3951
MR		 111712013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11172
11173 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11174 (UCLIBC_DYNAMIC_LINKER): New macro.
11175 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11176 `-mnan=2008'.
11177 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11178 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11179 `-mnan=2008'.
11180 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11181 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11182 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11183 for SF and DF modes. Use ieee_quad_format for TF mode.
11184 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11185 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11186 (mips_option_override): Handle `-mnan=legacy'.
11187 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11188 `-mabs=2008' and `-mnan=2008'.
11189 (OPTION_DEFAULT_SPECS): Add "nan" default.
11190 (ASM_SPEC): Handle `-mnan='.
11191 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11192 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11193 comment accordingly.
11194 (neg<mode>2): Likewise.
11195 * config/mips/mips.opt (mabs, mnan): New options.
11196 * doc/install.texi (Configuration): Document `--with-nan=' option.
11197 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11198 `-mnan=' options.
11199 (MIPS Options): Document them.
11200 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11201 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11202 * configure: Regenerate.
11203 * config.in: Regenerate.
11204
576e2f37
UB		 112052013-07-29 Uros Bizjak <ubizjak@gmail.com>
11206
11207 * config/i386/i386.md (float post-reload splitters): Do not check
11208 for subregs of SSE registers.
11209
112102013-07-29 Uros Bizjak <ubizjak@gmail.com>
11211 H.J. Lu <hongjiu.lu@intel.com>
11212
11213 PR target/57954
11214 PR target/57988
11215 * config/i386/i386.md (post-reload splitter
11216 to avoid partial SSE reg dependency stalls): New pattern.
11217
a71f0749
DV		 112182013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11219
11220 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11221
4c97f1cc
JR		 112222013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11223
11224 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11225 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11226
b0c733d8
CC		 112272013-07-26 Cary Coutant <ccoutant@google.com>
11228
11229 * dwarf2out.c (die_checksum_ordered): Don't include template
11230 instantiations in signature.
11231 (is_template_parameter): New function.
11232 (is_template_instantiation): New function.
11233 (generate_skeleton_bottom_up): Don't include template instantiations
11234 in type unit DIE.
11235 (generate_skeleton): Likewise.
11236 (break_out_comdat_types): Move recursive call to break out nested
11237 types earlier.
11238 (prune_unused_types_mark_generic_parms_dies): Call
11239 is_template_parameter.
11240
040d8a1c
IB		 112412013-07-26 Ian Bolton <ian.bolton@arm.com>
11242
11243 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11244 uses vector registers.
89fdc743 11245 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11246
85bd4ac6 112472013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11248 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 11249
11250 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11251 where appropriate.
85bd4ac6
KT		 11252 * config/arm/arm.md (movcond_addsi): New splitter.
11253
b6051207
SA		 112542013-07-25 Sterling Augustine <saugustine@google.com>
11255
11256 * dwarf2out.c (size_of_pubnames): Move code to...
11257 (include_pubname_in_output): ...here. New.
11258 (want_pubnames): Rearrange.
11259 (output_pubnames): Call include_pubname_in_output. Move assertion.
11260
41a7c215
CM		 112612013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11262
11263 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11264
112652013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11266
11267 PR target/38836
11268 * doc/extend.texi: Remove obsolete builtins. Fix
11269 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11270
49bde175
JH		 112712013-07-25 Jan Hubicka <jh@suse.cz>
11272
11273 * cgraph.c (release_function_body): Break out from ...
11274 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11275 and DECL_ARGUMENTS.
11276 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11277 old_tree in the map.
11278 (create_specialized_node): Update.
11279 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11280 into index.
41a7c215
CM		 11281 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11282 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 11283 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11284 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11285 DECL_RESULT.
11286
4870352d
KT		 112872013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11288
11289 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11290 addsi3_carryin_alt2_<optab>): Correct output template.
11291
1d6aee1c
KT		 112922013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11293
11294 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11295 Adjust for arm_restrict_it.
11296 Remove trailing whitespace.
11297
c7cafd75
MK		 112982013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11299
017d38f5
MK		 11300 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11301 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11302
c7cafd75
MK		 11303 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11304
00b7527b
VM		 113052013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11306
11307 PR rtl-optimization/57960
11308 * lra-constraints.c (process_alt_operands): Use the right mode
11309 when checking strict_low.
11310
4f63dfc6
JH		 113112013-07-25 Jan Hubicka <jh@suse.cz>
11312
11313 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11314 * cgraph.c (cgraph_remove_node): Do not release function body
11315 when in cgraph streaming.
11316 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 11317 in other partitions are not considered reachable; fix handling of
11318 clones.
4f63dfc6 11319
e6dcfa18
RR		 113202013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11321
11322 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11323
0b93d3b6
RR		 113242013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11325
11326 PR target/19599
11327 PR target/57731
bb80c2eb 11328 PR target/57837
e6dcfa18 11329 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11330 Ss with US. Adjust output for v5 and v4t.
41a7c215 11331 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11332
0b93d3b6
RR		 11333 * config/arm/constraints.md ("Ss"): Rename to US.
11334
690688b3
TG		 113352013-07-25 Terry Guo <terry.guo@arm.com>
11336
11337 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11338 shift_add/shift_sub0/shift_sub1 RTXs.
11339
a4ad093b 113402013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11341 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 11342
11343 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11344 (altivec_vpks<VI_char>ss): Likewise.
11345 (altivec_vpks<VI_char>us): Likewise.
11346 (altivec_vpku<VI_char>us): Likewise.
11347 (altivec_vpku<VI_char>um): Likewise.
11348
8c5005ce
DM		 113492013-07-24 David Malcolm <dmalcolm@redhat.com>
11350
11351 Introduce context class.
11352
11353 * Makefile.in (CONTEXT_H): New.
11354 (OBJS): Add context.o.
11355 (toplev.o): Add CONTEXT_H to dependencies.
11356 (context.o): New.
11357
41a7c215 11358 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 11359
11360 * context.c: New.
11361
11362 * context.h: New.
11363
ce4a9422
JR		 113642013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11365
11366 PR rtl-optimization/57968
11367 * mode-switching.c (create_pre_exit): Allow instructions that
11368 don't set a return register to need a non-exit mode.
11369
493f4c9e 113702013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11371 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11372
b5860fd3
WS		 11373 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11374 operands to vperm for little endian.
11375 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11376 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11377
73792b92 113782013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11379 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 11380
11381 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11382 two operands for little-endian.
11383
14c21302
SE		 113842013-07-23 Steve Ellcey <sellcey@mips.com>
11385
11386 * config/mips/mips.c (mips_case_values_threshold): New.
11387 (TARGET_CASE_VALUES_THRESHOLD): Define.
11388
a6056198 113892013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11390 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 11391
11392 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11393 selection of field for vector splat in little endian mode.
11394
dd7a40e1
MM		 113952013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11396
11397 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11398 expanders to rs6000.md.
11399 (ior<mode>3): Likewise.
11400 (and<mode>3): Likewise.
11401 (one_cmpl<mode>2): Likewise.
11402 (nor<mode>3): Likewise.
11403 (andc<mode>3): Likewise.
11404 (eqv<mode>3): Likewise.
11405 (nand<mode>3): Likewise.
11406 (orc<mode>3): Likewise.
11407
11408 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11409 declaration.
11410
11411 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11412 to split multi-word logical operations.
11413 (rs6000_split_logical_di): Likewise.
11414 (rs6000_split_logical): Likewise.
11415
11416 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11417 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11418 and allow TImode operations in 32-bit.
11419 (vsx_and<mode>3_64bit): Likewise.
11420 (vsx_ior<mode>3_32bit): Likewise.
11421 (vsx_ior<mode>3_64bit): Likewise.
11422 (vsx_xor<mode>3_32bit): Likewise.
11423 (vsx_xor<mode>3_64bit): Likewise.
11424 (vsx_one_cmpl<mode>2_32bit): Likewise.
11425 (vsx_one_cmpl<mode>2_64bit): Likewise.
11426 (vsx_nor<mode>3_32bit): Likewise.
11427 (vsx_nor<mode>3_64bit): Likewise.
11428 (vsx_andc<mode>3_32bit): Likewise.
11429 (vsx_andc<mode>3_64bit): Likewise.
11430 (vsx_eqv<mode>3_32bit): Likewise.
11431 (vsx_eqv<mode>3_64bit): Likewise.
11432 (vsx_nand<mode>3_32bit): Likewise.
11433 (vsx_nand<mode>3_64bit): Likewise.
11434 (vsx_orc<mode>3_32bit): Likewise.
11435 (vsx_orc<mode>3_64bit): Likewise.
11436
11437 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11438 logical types in GPRs.
11439
11440 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11441 logical insns to rs6000.md, and allow TImode operations in
11442 32-bit.
11443 (altivec_ior<mode>3): Likewise.
11444 (altivec_xor<mode>3): Likewise.
11445 (altivec_one_cmpl<mode>2): Likewise.
11446 (altivec_nor<mode>3): Likewise.
11447 (altivec_andc<mode>3): Likewise.
11448
11449 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11450 attributes for moving the 128-bit logical operations into
11451 rs6000.md.
11452 (BOOL_REGS_OUTPUT): Likewise.
11453 (BOOL_REGS_OP1): Likewise.
11454 (BOOL_REGS_OP2): Likewise.
11455 (BOOL_REGS_UNARY): Likewise.
11456 (BOOL_REGS_AND_CR0): Likewise.
11457 (one_cmpl<mode>2): Add support for DI logical operations on
11458 32-bit, splitting the operations to 32-bit.
11459 (anddi3): Likewise.
11460 (iordi3): Likewise.
11461 (xordi3): Likewise.
11462 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11463 changes to combine the 32/64-bit code, allow logical operations on
11464 TI mode in 32-bit, and to use similar match_operator patterns like
11465 scalar mode uses. Combine the Altivec and VSX code for logical
11466 operations, and move it here.
11467 (ior<mode>3, 128-bit types): Likewise.
11468 (xor<mode>3, 128-bit types): Likewise.
11469 (one_cmpl<mode>3, 128-bit types): Likewise.
11470 (nor<mode>3, 128-bit types): Likewise.
11471 (andc<mode>3, 128-bit types): Likewise.
11472 (eqv<mode>3, 128-bit types): Likewise.
11473 (nand<mode>3, 128-bit types): Likewise.
11474 (orc<mode>3, 128-bit types): Likewise.
11475 (and<mode>3_internal): Likewise.
11476 (bool<mode>3_internal): Likewise.
11477 (boolc<mode>3_internal1): Likewise.
11478 (boolc<mode>3_internal2): Likewise.
11479 (boolcc<mode>3_internal1): Likewise.
11480 (boolcc<mode>3_internal2): Likewise.
11481 (eqv<mode>3_internal1): Likewise.
11482 (eqv<mode>3_internal2): Likewise.
11483 (one_cmpl1<mode>3_internal): Likewise.
11484
12211b99 114852013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 11486
11487 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11488 Rename flag_stack_usage to flag_stack_usage_info.
11489
12211b99 114902013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 11491
a5f257fd
ME		 11492 * config/microblaze/sync.md: New file.
11493 * config/microblaze/microblaze.md: Include sync.md
11494 * config/microblaze/microblaze.c: Add print_operand 'y'.
11495 * config/microblaze/constraints.md: Add memory_contraint
11496 'Q' which is a single register.
450b0ebb 11497
49fba14f
EB		 114982013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11499
11500 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11501
c7b3b99f
PCC		 115022013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11503
11504 * reload.c (find_reloads): Exit loop once we find this operand
11505 cannot be reloaded somehow for this alternative.
11506
11507 * reload.c (find_reloads): Exit loop once we find a hard register.
11508
11509 * rtlanal.c (computed_jump_p): Exit loop once we find label
11510 reference is used.
11511
11512 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11513
11514 * cfgloopmanip.c (remove_path): Exit loop after setting
11515 irred_invalidated.
11516
11517 * gensupport.c (subst_dup): Avoid loop if code is not
11518 MATCH_DUP nor MATCH_OP_DUP.
11519
c8fbf1fa
NBJ		 115202013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11521
11522 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11523
9259db42
YZ		 115242013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11525
11526 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11527 true for SP_REGNUM if mode == ptr_mode.
11528 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11529 with value R0_REGNUM + 31.
11530
00edcfbe
YZ		 115312013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11532
11533 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11534 pad pointer-typed argument downward.
11535
43be9a95
YZ		 115362013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11537
11538 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11539 and __ILP32__ when the ILP32 model is in use.
11540
28514dda
YZ		 115412013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11542
11543 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11544 (aarch64_load_symref_appropriately): In the case of
11545 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11546 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11547 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11548 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11549 if 'mode' doesn't equal to 'ptr_mode'.
11550 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11551 'vcall_offset'; change to call aarch64_emit_move differently depending
11552 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11553 to calculate the upper bound of 'vcall_offset'.
11554 (aarch64_cannot_force_const_mem): Change to also return true if
11555 mode != ptr_mode.
11556 (aarch64_legitimize_reload_address): In the case of large
11557 displacements, add new local variable 'xmode' and an assertion
11558 based on it; change to use 'xmode' to generate the new rtx and
11559 reload.
11560 (aarch64_asm_trampoline_template): Change to generate the template
11561 differently depending on TARGET_ILP32 or not; change to use
11562 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11563 (aarch64_trampoline_size): Removed.
11564 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11565 and replace immediate literals with it. Change to use 'ptr_mode'
11566 instead of 'DImode' and call convert_memory_address if the mode
11567 of 'fnaddr' doesn't equal to 'ptr_mode'.
11568 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11569 to output symbol.
11570 (aarch64_elf_asm_destructor): Likewise.
11571 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11572 on TARGET_ILP32 instead of aarch64_trampoline_size.
11573 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11574 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11575 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11576 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11577 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11578 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11579 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11580 depending on the value of 'mode'.
11581 (add_losym_<mode>): New.
11582 (ldr_got_small_<mode>): New, based on ldr_got_small.
11583 (ldr_got_small): Remove.
11584 (ldr_got_small_sidi): New.
11585 * config/aarch64/iterators.md (P): New.
11586 (PTR): Change to 'ptr_mode' in the condition.
11587
17a819cb
YZ		 115882013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11589
11590 * config.gcc (aarch64*-*-*): Support --with-abi.
11591 (aarch64*-*-elf): Support --with-multilib-list.
11592 (aarch64*-*-linux*): Likewise.
11593 (supported_defaults): Add abi to aarch64*-*-*.
11594 * configure.ac: Mention AArch64 for --with-multilib-list.
11595 * configure: Re-generated.
11596 * config/aarch64/biarchilp32.h: New file.
11597 * config/aarch64/biarchlp64.h: New file.
11598 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11599 (ABI_SPEC): Ditto.
11600 (MULTILIB_DEFAULTS): Ditto.
11601 (DRIVER_SELF_SPECS): Ditto.
11602 (ASM_SPEC): Update to also substitute -mabi.
11603 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11604 file whose name depends on -mabi= and -mbig-endian.
11605 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11606 TARGET_ILP32.
11607 (POINTER_SIZE): New define.
11608 (POINTERS_EXTEND_UNSIGNED): Ditto.
11609 (enum aarch64_abi_type): New enumeration tag.
11610 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11611 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11612 (TARGET_ILP32): New define.
11613 * config/aarch64/aarch64.opt (mabi): New.
11614 (aarch64_abi): New.
11615 (ilp32, lp64): New values for -mabi.
11616 * config/aarch64/t-aarch64 (comma): New define.
11617 (MULTILIB_OPTIONS): Ditto.
11618 (MULTILIB_DIRNAMES): Ditto.
11619 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11620 * doc/invoke.texi: Document -mabi for AArch64.
11621
9f9cbdce
GJL		 116222013-07-23 Georg-Johann Lay <avr@gjlay.de>
11623
11624 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11625
7061977b 116262013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11627 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 11628
11629 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11630 endianness when selecting field to splat.
11631
75a41878
EC		 116322013-07-22 Eric Christopher <echristo@gmail.com>
11633
11634 * dwarf2out.c (die_odr_checksum): New function to use
11635 CHECKSUM_ macros and ULEB128 for DIE tag.
11636 (generate_type_signature): Use.
41a7c215 11637
38ae58ca
EB		 116382013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11639
11640 * config.gcc (sparc*-*-*): Accept leon3 processor.
11641 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11642 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11643 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11644 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11645 (mfix-ut699): Adjust comment.
11646 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11647 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11648 (CPP_CPU_SPEC): Likewise.
11649 (ASM_CPU_SPEC): Likewise.
11650 * config/sparc/sparc.c (leon3_cost): New constant.
11651 (sparc_option_override): Add leon3 support.
11652 (mem_ref): New function.
11653 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11654 (sparc_do_work_around_errata): Look into the instruction in the delay
11655 slot and adjust accordingly. Add fix for the data cache nullify issues
11656 of the UT699. Change insertion position for the NOP.
11657 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11658 (leon3_load): New reservation.
11659 (leon_store): Bump latency to 2.
11660 (grfpu): New automaton.
11661 (grfpu_alu): New unit.
11662 (grfpu_ds): Likewise.
11663 (leon_fp_alu): Adjust.
11664 (leon_fp_mult): Delete.
11665 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11666 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11667 * config/sparc/sparc.md (cpu): Add leon3.
11668 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11669 (swapsi): Likewise.
11670 (atomic_test_and_set): Likewise.
11671 (ldstub): Likewise.
11672
04dfc6df
JU		 116732013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11674
11675 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11676 default for R5900 targets.
11677 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11678 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11679 * config/mips/mips.c (mips_option_override): Report an error for
11680 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11681 for -march=r5900 -mhard-float.
11682
c7b3b99f 116832013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 11684
11685 * df-problems.c (can_move_insns_across): Exit loop once we
11686 find a non-fixed, non-global register.
11687
11688 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11689 setting can_throw.
11690
11691 * omega.c (omega_eliminate_red): Break after setting red_found.
11692 (omega_problem_has_red_equations): Similarly after setting found.
11693 (omega_query_variable): Similarly after setting coupled.
11694
9b6e6981
MP		 116952013-07-22 Marek Polacek <polacek@redhat.com>
11696
11697 * gimplify.c: Don't include gimple.h twice.
11698
5e5f7673
KT		 116992013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11700
11701 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11702 instead of TARGET_THUMB1.
11703 (Pz): New constraint.
11704 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11705 encodings.
11706 (compare_negsi_si): Likewise.
11707 (compare_addsi2_op0): Likewise.
11708 (compare_addsi2_op1): Likewise.
11709 (addsi3_carryin_<optab>): Likewise.
11710 (addsi3_carryin_alt2_<optab>): Likewise.
11711 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11712 for arm_restrict_it.
11713 (subsi3_carryin): Likewise.
11714 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11715 (minmax_arithsi): Disable for arm_restrict_it.
11716 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11717 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11718 (satsi_<SAT:code>_shift): Likewise.
11719 (arm_shiftsi3): Add alternative for 16-bit encoding.
11720 (arm32_movhf): Disable for arm_restrict_it.
11721 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11722 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11723
859abddd
SN		 117242013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11725
11726 * config/arm/arm.md (attribute "insn"): Delete.
11727 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11728 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11729 (not_shiftsi): Update for attribute change.
11730 (not_shiftsi_compare0): Likewise.
11731 (not_shiftsi_compare0_scratch): Likewise.
11732 (arm_one_cmplsi2): Likewise.
11733 (thumb1_one_cmplsi2): Likewise.
11734 (notsi_compare0): Likewise.
11735 (notsi_compare0_scratch): Likewise.
11736 (thumb1_movdi_insn): Likewise.
11737 (arm_movsi_insn): Likewise.
11738 (movhi_insn_arch4): Likewise.
11739 (movhi_bytes): Likewise.
11740 (arm_movqi_insn): Likewise.
11741 (thumb1_movqi_insn): Likewise.
11742 (arm32_movhf): Likewise.
11743 (thumb1_movhf): Likewise.
11744 (arm_movsf_soft_insn): Likewise.
11745 (thumb1_movsf_insn): Likewise.
11746 (thumb_movdf_insn): Likewise.
11747 (movsicc_insn): Likewise.
11748 (movsfcc_soft_insn): Likewise.
11749 (and_scc): Likewise.
11750 (cond_move): Likewise.
11751 (if_move_not): Likewise.
11752 (if_not_move): Likewise.
11753 (if_shift_move): Likewise.
11754 (if_move_shift): Likewise.
11755 (if_shift_shift): Likewise.
11756 (if_not_arith): Likewise.
11757 (if_arith_not): Likewise.
11758 (cond_move_not): Likewise.
11759 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11760 (neon_mov<mode>): Likewise.
11761 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11762 (thumb2_movsi_vfp): Likewise.
11763 (movsf_vfp): Likewise.
11764 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 11765 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11766 change.
859abddd
SN		 11767 (cortexa7_older_only): Likewise.
11768 (cortexa7_younger): Likewise.
11769 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11770 (1020alu_shift_op): Likewise.
11771 (1020alu_shift_reg_op): Likewise.
11772 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11773 (alu_shift_op): Likewise.
11774 (alu_shift_reg_op): Likewise.
11775 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11776 (11_alu_shift_op): Likewise.
11777 (11_alu_shift_reg_op): Likewise.
11778 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11779 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 11780 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11781 change.
859abddd
SN		 11782 (cortex_a15_alu_shift): Likewise.
11783 (cortex_a15_alu_shift_reg): Likewise.
11784 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11785 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 11786 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11787 change.
859abddd 11788 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 11789 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11790 change.
859abddd
SN		 11791 (cortex_a7_alu_reg): Likewise.
11792 (cortex_a7_alu_shift): Likewise.
11793 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11794 (cortex_a8_alu_shift): Likewise.
11795 (cortex_a8_alu_shift_reg): Likewise.
11796 (cortex_a8_mov): Likewise.
11797 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11798 (cortex_a9_dp_shift): Likewise.
11799 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11800 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11801 (cortex_r4_mov): Likewise.
11802 (cortex_r4_alu_shift): Likewise.
11803 (cortex_r4_alu_shift_reg): Likewise.
11804 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11805 (526_alu_shift_op): Likewise.
11806 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11807 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11808 (626te_alu_shift_op): Likewise.
11809 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11810 (726te_alu_op): Likewise.
11811 (726te_alu_shift_op): Likewise.
11812 (726te_alu_shift_reg_op): Likewise.
11813 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11814 (mp626_alu_shift_op): Likewise.
11815 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11816 (pj4_alu_e1_conds): Likewise.
11817 (pj4_alu): Likewise.
11818 (pj4_alu_conds): Likewise.
11819 (pj4_shift): Likewise.
11820 (pj4_shift_conds): Likewise.
11821 (pj4_alu_shift): Likewise.
11822 (pj4_alu_shift_conds): Likewise.
11823
95b97fac
KT		 118242013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11825
11826 * config/arm/predicates.md (shiftable_operator_strict_it):
11827 New predicate.
11828 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11829 Disable cond_exec version for arm_restrict_it.
11830 (thumb2_smaxsi3): Convert to generate cond_exec.
11831 (thumb2_sminsi3): Likewise.
11832 (thumb32_umaxsi3): Likewise.
11833 (thumb2_uminsi3): Likewise.
11834 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11835 (thumb2_neg_abssi2): Likewise.
11836 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11837 (thumb2_movsicc_insn): Adjust alternatives.
11838 (thumb2_mov_negscc): Disable for arm_restrict_it.
11839 (thumb2_mov_negscc_strict_it): New pattern.
11840 (thumb2_mov_notscc_strict_it): New pattern.
11841 (thumb2_mov_notscc): Disable for arm_restrict_it.
11842 (thumb2_ior_scc): Likewise.
11843 (thumb2_ior_scc_strict_it): New pattern.
11844 (thumb2_cond_move): Adjust for arm_restrict_it.
11845 (thumb2_cond_arith): Disable for arm_restrict_it.
11846 (thumb2_cond_arith_strict_it): New pattern.
11847 (thumb2_cond_sub): Adjust for arm_restrict_it.
11848 (thumb2_movcond): Likewise.
11849 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11850 (thumb2_zero_extendhisi2_v6): Likewise.
11851 (thumb2_zero_extendqisi2_v6): Likewise.
11852 (orsi_notsi_si): Likewise.
11853 (orsi_not_shiftsi_si): Likewise.
11854
d5a6ef82
GJL		 118552013-07-22 Georg-Johann Lay <avr@gjlay.de>
11856
11857 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11858 instruction sequence is 1 byte shorter.
11859
d9e71154
UB		 118602013-07-22 Uros Bizjak <ubizjak@gmail.com>
11861
11862 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11863 it is not needed after split.
11864
8c6381ec
IS		 118652013-07-20 Iain Sandoe <iain@codesourcery.com>
11866
11867 PR target/51784
11868 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11869 second label for nonlocal goto receivers. Don't output pic base labels
11870 unless we're producing PIC; mark that action unreachable().
11871 (ix86_save_reg): If the function contains a nonlocal label, save the
11872 PIC base reg.
11873 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 11874 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 11875 (update_pic_label_number_if_needed): New.
11876 (machopic_output_function_base_name): Adjust for nonlocal receiver
11877 case.
11878 (machopic_should_output_picbase_label): New.
11879 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11880 (nonlocal_goto_receiver): New insn and split.
11881
d05d0709
JG		 118822013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11883
11884 * config/aarch64/aarch64-builtins.c
11885 (aarch64_fold_builtin): Fold abs in all modes.
11886 * config/aarch64/aarch64-simd-builtins.def
11887 (abs): Enable for all modes.
11888 * config/aarch64/arm_neon.h
11889 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11890 (vabs_f64): Add missing intrinsic.
11891
ba590e63
IB		 118922013-07-19 Ian Bolton <ian.bolton@arm.com>
11893
11894 * config/aarch64/arm_neon.h (vabs_s64): New function
11895
02371798
GJL		 118962013-07-19 Georg-Johann Lay <avr@gjlay.de>
11897
11898 PR target/57516
11899 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11900 * config/avr/avr.md (adjust_len): Add `round'.
11901 * config/avr/avr-protos.h (avr_out_round): New prototype.
11902 (avr_out_plus): Add `out_label' argument.
11903 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11904 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11905 Handle the case where `insn' is just a pattern.
11906 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11907 (avr_out_round): New function.
11908 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11909
12211b99 119102013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 11911
11912 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11913 Add check for flag_stack_usage to handle -fstack-usage support
11914
12211b99 119152013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 11916
11917 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11918 interaction for new Power8 flags and VSX.
11919
df9dda2d
ST		 119202013-07-18 Sriraman Tallam <tmsriram@google.com>
11921
11922 PR middle-end/57698
11923 * tree-inline.c (expand_call_inline): Emit errors during
11924 early_inlining only if optimization is not turned on.
11925
0f9cdef4
DM		 119262013-07-18 David Malcolm <dmalcolm@redhat.com>
11927
11928 * passes.def: New.
11929
11930 * passes.c (init_optimization_passes): Move the construction of
11931 the pass hierarchy into a new passes.def file.
11932
11933 * Makefile.in (passes.o): Add dependency on passes.def.
11934
2efa4087
DM		 119352013-07-18 David Malcolm <dmalcolm@redhat.com>
11936
11937 * passes.c (init_optimization_passes): Introduce macros for
11938 constructing the tree of passes (INSERT_PASSES_AFTER,
11939 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
11940 TERMINATE_PASS_LIST).
11941
bc404e1b
VM		 119422013-07-18 Vladimir Makarov <vmakarov@redhat.com>
11943 Wei Mi <wmi@google.com>
11944
11945 PR rtl-optimization/57878
11946 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
11947 top.
11948 (reload_pseudo_compare_func): Check nregs first for reload
11949 pseudos.
11950
d675843f
DM		 119512013-07-18 David Malcolm <dmalcolm@redhat.com>
11952
11953 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
11954
c7b3b99f 119552013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 11956
11957 * read-rtl.c (validate_const_int): Once an invalid character is
11958 seen, quit the loop.
11959
11960 * gengtype.c (write_roots): Similarly once we find the "deletable"
11961 or "if_marked" option.
11962
52fceb44
SN		 119632013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11964
11965 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
11966 "xtab" and "sat". Move value "clz" from here to ...
11967 (attriubte "type"): ... here.
11968 (satsi_<SAT:code>): Delete "insn" attribute.
11969 (satsi_<SAT:code>_shift): Likewise.
11970 (arm_zero_extendqisi2addsi): Likewise.
11971 (arm_extendqisi2addsi): Likewise.
11972 (clzsi2): Update for attribute changes.
11973 (rbitsi2): Likewise.
371e77e3
UB		 11974 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
11975 attribute.
52fceb44
SN		 11976 (arm_usatsihi): Likewise.
11977 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11978
006bd006
SN		 119792013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11980
11981 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 11982 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
11983 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 11984 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
11985 in alphabetical order.
11986 (attribute "core_cycles"): Update for attribute changes.
11987 (arm_addsi3): Likewise.
11988 (addsi3_compare0): Likewise.
11989 (addsi3_compare0_scratch): Likewise.
11990 (addsi3_compare_op1): Likewise.
11991 (addsi3_compare_op2): Likewise.
11992 (compare_addsi2_op0): Likewise.
11993 (compare_addsi2_op1): Likewise.
11994 (addsi3_carryin_shift_<optab>): Likewise.
11995 (subsi3_carryin_shift): Likewise.
11996 (rsbsi3_carryin_shift): Likewise.
11997 (arm_subsi3_insn): Likewise.
11998 (subsi3_compare0): Likewise.
11999 (subsi3_compare): Likewise.
12000 (arm_andsi3_insn): Likewise.
12001 (thumb1_andsi3_insn): Likewise.
12002 (andsi3_compare0): Likewise.
12003 (andsi3_compare0_scratch): Likewise.
12004 (zeroextractsi_compare0_scratch
12005 (andsi_not_shiftsi_si): Likewise.
12006 (iorsi3_insn): Likewise.
12007 (iorsi3_compare0): Likewise.
12008 (iorsi3_compare0_scratch): Likewise.
12009 (arm_xorsi3): Likewise.
12010 (thumb1_xorsi3_insn): Likewise.
12011 (xorsi3_compare0): Likewise.
12012 (xorsi3_compare0_scratch): Likewise.
12013 (satsi_<SAT:code>_shift): Likewise.
12014 (rrx): Likewise.
12015 (arm_shiftsi3): Likewise.
12016 (shiftsi3_compare0): Likewise.
12017 (not_shiftsi): Likewise.
12018 (not_shiftsi_compare0): Likewise.
12019 (not_shiftsi_compare0_scratch): Likewise.
12020 (arm_one_cmplsi2): Likewise.
12021 (thumb_one_complsi2): Likewise.
12022 (notsi_compare0): Likewise.
12023 (notsi_compare0_scratch): Likewise.
12024 (thumb1_zero_extendhisi2): Likewise.
12025 (arm_zero_extendhisi2): Likewise.
12026 (arm_zero_extendhisi2_v6): Likewise.
12027 (arm_zero_extendhisi2addsi): Likewise.
12028 (thumb1_zero_extendqisi2): Likewise.
12029 (thumb1_zero_extendqisi2_v6): Likewise.
12030 (arm_zero_extendqisi2): Likewise.
12031 (arm_zero_extendqisi2_v6): Likewise.
12032 (arm_zero_extendqisi2addsi): Likewise.
12033 (thumb1_extendhisi2): Likewise.
12034 (arm_extendhisi2): Likewise.
12035 (arm_extendhisi2_v6): Likewise.
12036 (arm_extendqisi): Likewise.
12037 (arm_extendqisi_v6): Likewise.
12038 (arm_extendqisi2addsi): Likewise.
12039 (thumb1_extendqisi2): Likewise.
12040 (thumb1_movdi_insn): Likewise.
12041 (arm_movsi_insn): Likewise.
12042 (movsi_compare0): Likewise.
12043 (movhi_insn_arch4): Likewise.
12044 (movhi_bytes): Likewise.
12045 (arm_movqi_insn): Likewise.
12046 (thumb1_movqi_insn): Likewise.
12047 (arm32_movhf): Likewise.
12048 (thumb1_movhf): Likewise.
12049 (arm_movsf_soft_insn): Likewise.
12050 (thumb1_movsf_insn): Likewise.
12051 (movdf_soft_insn): Likewise.
12052 (thumb_movdf_insn): Likewise.
12053 (arm_cmpsi_insn): Likewise.
12054 (cmpsi_shiftsi): Likewise.
12055 (cmpsi_shiftsi_swp): Likewise.
12056 (arm_cmpsi_negshiftsi_si): Likewise.
12057 (movsicc_insn): Likewise.
12058 (movsfcc_soft_insn): Likewise.
12059 (arith_shiftsi): Likewise.
12060 (arith_shiftsi_compare0
12061 (arith_shiftsi_compare0_scratch
12062 (sub_shiftsi): Likewise.
12063 (sub_shiftsi_compare0
12064 (sub_shiftsi_compare0_scratch
12065 (and_scc): Likewise.
12066 (cond_move): Likewise.
12067 (if_plus_move): Likewise.
12068 (if_move_plus): Likewise.
12069 (if_move_not): Likewise.
12070 (if_not_move): Likewise.
12071 (if_shift_move): Likewise.
12072 (if_move_shift): Likewise.
12073 (if_shift_shift): Likewise.
12074 (if_not_arith): Likewise.
12075 (if_arith_not): Likewise.
12076 (cond_move_not): Likewise.
12077 (thumb1_ashlsi3): Set type attribute.
12078 (thumb1_ashrsi3): Likewise.
12079 (thumb1_lshrsi3): Likewise.
12080 (thumb1_rotrsi3): Likewise.
12081 (shiftsi3_compare0_scratch): Likewise.
12082 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12083 (neon_mov<mode>): Likewise.
371e77e3
UB		 12084 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12085 attribute changes.
006bd006
SN		 12086 (thumb2_movsi_insn): Likewise.
12087 (thumb2_cmpsi_neg_shiftsi): Likewise.
12088 (thumb2_extendqisi_v6): Likewise.
12089 (thumb2_zero_extendhisi2_v6): Likewise.
12090 (thumb2_zero_extendqisi2_v6): Likewise.
12091 (thumb2_shiftsi3_short): Likewise.
12092 (thumb2_addsi3_compare0_scratch): Likewise.
12093 (orsi_not_shiftsi_si): Likewise.
12094 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12095 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12096 changes.
12097 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12098 (1020alu_shift_op): Likewise.
12099 (1020alu_shift_reg_op): Likewise.
12100 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12101 (alu_shift_op): Likewise.
12102 (alu_shift_reg_op): Likewise.
12103 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12104 (11_alu_shift_op): Likewise.
12105 (11_alu_shift_reg_op): Likewise.
12106 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12107 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12108 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12109 changes.
006bd006
SN		 12110 (cortex_a15_alu_shift): Likewise.
12111 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 12112 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12113 changes.
006bd006
SN		 12114 (cortex_a5_alu_shift): Likewise.
12115 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12116 changes.
12117 (cortex_a53_alu_shift): Likewise.
12118 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12119 changes.
12120 (cortex_a7_alu_reg): Likewise.
12121 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 12122 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12123 changes.
006bd006
SN		 12124 (cortex_a8_alu_shift): Likewise.
12125 (cortex_a8_alu_shift_reg): Likewise.
12126 (cortex_a8_mov): Likewise.
12127 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12128 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 12129 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12130 changes.
12131 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12132 changes.
006bd006
SN		 12133 (cortex_r4_mov): Likewise.
12134 (cortex_r4_alu_shift): Likewise.
12135 (cortex_r4_alu_shift_reg): Likewise.
12136 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12137 (526_alu_shift_op): Likewise.
12138 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12139 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12140 (626te_alu_shift_op): Likewise.
12141 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12142 (726te_alu_op): Likewise.
12143 (726te_alu_shift_op): Likewise.
12144 (726te_alu_shift_reg_op): Likewise.
12145 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12146 (mp626_alu_shift_op): Likewise.
12147 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12148 (pj4_alu_e1_conds): Likewise.
12149 (pj4_alu): Likewise.
12150 (pj4_alu_conds): Likewise.
12151 (pj4_shift): Likewise.
12152 (pj4_shift_conds): Likewise.
12153 (pj4_alu_shift): Likewise.
12154 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 12155 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12156 changes.
006bd006
SN		 12157 (cortexa7_older_only): Likewise.
12158 (cortexa7_younger): Likewise.
12159
651df1b2
DM		 121602013-07-18 David Malcolm <dmalcolm@redhat.com>
12161
12162 * ipa-pure-const.c (generate_summary): Rename to...
12163 (pure_const_generate_summary): ... this.
12164
ea717bd7
IS		 121652013-07-17 Iain Sandoe <iain@codesourcery.com>
12166
12167 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12168
a1259a13
YR		 121692013-07-17 Yvan Roux <yvan.roux@linaro.org>
12170
12171 PR target/57909
12172 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12173 usage in HI mode.
12174
167f68ed
AK		 121752013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12176
12177 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12178 enabled without -march=zEC12.
12179 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12180 flags to be set.
12181
f900a982
MR		 121822013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12183
12184 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12185 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12186 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12187 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12188 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12189 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12190 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12191 ISA_HAS_FP4.
12192 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12193 and ISA_HAS_NMADD3_NMSUB3.
12194 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12195 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12196 (nmsub4<mode>, nmsub3<mode>): Likewise.
12197 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12198
d4ed27eb
MR		 121992013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12200
12201 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12202 TARGET_MIPS5400 checking.
12203
7a5add18
PB		 122042013-07-16 Jakub Jelinek <jakub@redhat.com>
12205 Peter Bergner <bergner@vnet.ibm.com>
12206
12207 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12208 registers in the comment.
12209 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12210 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12211 rather than FIRST_PSEUDO_REGISTERS.
12212
122132013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 12214
12215 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12216 enable extra ISA flags with TARGET_HTM.
12217
d2ab0929
MR		 122182013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12219
12220 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12221 Fix comment typos.
12222
839c74bc
CH		 122232013-07-15 Cong Hou <congh@google.com>
12224
12225 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12226 in compare function for sorting.
12227
7a5add18 122282013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 12229
12230 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12231 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12232 * config/rs6000/rs6000.opt: Add -mhtm option.
12233 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12234 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12235 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12236 __HTM__ if the HTM instructions are available.
12237 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12238 htm_spr_reg_operand): New define_predicates.
12239 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12240 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12241 Include htm.md.
12242 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12243 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12244 HTM builtin functions.
12245 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12246 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12247 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12248 (rs6000_builtin_mask_calculate): Likewise.
12249 (rs6000_option_override_internal): Likewise.
12250 (bdesc_htm): Add new HTM builtin support.
12251 (htm_spr_num): New function.
12252 (htm_spr_regno): Likewise.
12253 (rs6000_htm_spr_icode): Likewise.
12254 (htm_expand_builtin): Likewise.
12255 (htm_init_builtins): Likewise.
12256 (rs6000_expand_builtin): Add support for HTM builtin functions.
12257 (rs6000_init_builtins): Likewise.
371e77e3
UB		 12258 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12259 option.
0258b6e4
PB		 12260 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12261 (TARGET_HTM, MASK_HTM): Define macros.
12262 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12263 (FIXED_REGISTERS): Likewise.
12264 (CALL_USED_REGISTERS): Likewise.
12265 (CALL_REALLY_USED_REGISTERS): Likewise.
12266 (REG_ALLOC_ORDER): Likewise.
12267 (enum reg_class): Likewise.
12268 (REG_CLASS_NAMES): Likewise.
12269 (REG_CLASS_CONTENTS): Likewise.
12270 (REGISTER_NAMES): Likewise.
12271 (ADDITIONAL_REGISTER_NAMES): Likewise.
12272 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12273 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12274 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12275 * config/rs6000/htm.md: New file.
12276 * config/rs6000/htmintrin.h: New file.
12277 * config/rs6000/htmxlintrin.h: New file.
12278
87dd8ab0
MS		 122792013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12280
12281 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12282 Define SYMBOL_TINY_GOT, update comment.
12283 * config/aarch64/aarch64.c
12284 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12285 (aarch64_expand_mov_immediate): Likewise.
12286 (aarch64_print_operand): Likewise.
12287 (aarch64_classify_symbol): Likewise.
12288 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12289 (ldr_got_tiny): Define.
12290
52d676b6
TG		 122912013-07-13 Tobias Grosser <tobias@grosser.es>
12292
12293 PR tree-optimization/54094
12294 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12295 scheduling dimension for the parallelism check from the polyhedral
12296 information in the AST.
12297 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12298 in 2D + 1 form.
12299
286e8fc1
JM		 123002013-07-13 Jason Merrill <jason@redhat.com>
12301
12302 * print-tree.c (debug_vec_tree): Use debug_raw.
12303 (debug_raw (vec<tree, va_gc> &)): New.
12304 (debug_raw (vec<tree, va_gc> *)): New.
12305 * tree.h: Declare them.
12306
e43257e8
BC		 123072013-07-13 Bin Cheng <bin.cheng@arm.com>
12308
12309 * ifcvt.c (ifcvt_after_combine): New static variable.
12310 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12311 for size.
12312 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12313 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12314 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12315
da582d46
MR		 123162013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12317
12318 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12319
18b0ea8f
MM		 123202013-07-12 Michael Matz <matz@suse.de>
12321
12322 PR middle-end/55771
12323 * convert.c (convert_to_real): Reject non-float inner types.
12324
e4f0f84d
TB		 123252013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12326
12327 * config/aarch64/aarch64-protos.h
12328 (aarch64_simd_immediate_valid_for_move): Remove.
12329 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12330 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12331 cases.
12332 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12333
76500022
SE		 123342013-07-11 Steve Ellcey <sellcey@mips.com>
12335
12336 * config/mips/mips.c (mips_conditional_register_usage): Do not
12337 use t[0-7] registers in MIPS16 mode when optimizing for size.
12338
78039734
ST		 123392013-07-11 Sriraman Tallam <tmsriram@google.com>
12340
12341 * config/i386/i386.c (dispatch_function_versions): Fix array
12342 indexing of function_version_info to match actual_versions.
12343
7b24b675
TJ		 123442013-07-11 Teresa Johnson <tejohnson@google.com>
12345
12346 * vec.h (struct va_gc): Move release out-of-line.
12347 (va_gc::release): Call ggc_free on released vec.
12348
85d4cbb8
UW		 123492013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12350
12351 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12352 Require GOT register as additional operand in UNSPEC.
12353 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12354 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12355 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12356 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12357 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12358 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12359 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12360
52b3f9e2
GJL		 123612013-07-11 Georg-Johann Lay <avr@gjlay.de>
12362
12363 PR target/57631
12364 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12365 name seen by assembler/linker rather if available.
12366
3f3e266e
AS		 123672013-07-11 Andreas Schwab <schwab@suse.de>
12368
12369 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12370
21231ca6
VM		 123712013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12372
371e77e3 12373 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12374
d331d374
JM		 123752013-07-10 Joseph Myers <joseph@codesourcery.com>
12376
673c2f63
JM		 12377 * doc/tm.texi.in: Move hook documentation to ....
12378 * target.def: ... here.
12379
d331d374
JM		 12380 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12381 text on @hook line.
12382 * doc/tm.texi: Regenerate.
12383
dfeadaa0
PC		 123842013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12385
12386 PR c++/57869
12387 * doc/invoke.texi: Document -Wconditionally-supported.
12388
bb6f2bac
GJL		 123892013-07-10 Georg-Johann Lay <avr@gjlay.de>
12390
12391 PR target/57844
12392 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12393 of my_fp.
12394
5219b232
GJL		 123952013-07-10 Georg-Johann Lay <avr@gjlay.de>
12396
12397 PR target/57506
12398 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12399 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12400 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12401 Remove duplicate devices.
12402 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12403 * config/avr/t-multilib: Regenerate.
12404 * config/avr/avr-tables.opt: Regenerate.
12405 * doc/avr-mmcu.texi: Regenerate.
12406
283045c0
GJL		 124072013-07-10 Georg-Johann Lay <avr@gjlay.de>
12408
12409 PR target/56987
12410 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12411
ceb2d59f 124122013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 12413
ceb2d59f
GS		 12414 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12415 the cost of MULT when optimizing for size.
12416
055e0a99
JBG		 124172013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12418
12419 * config/cr16/cr16-protos.h: Don't include target.h.
12420
ef43e868
JM		 124212013-07-09 Joseph Myers <joseph@codesourcery.com>
12422
12423 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12424 adjust register size for TDmode and TFmode for VSX registers.
12425
4a283fdf
KT		 124262013-07-08 Kai Tietz <ktietz@redhat.com>
12427
12428 PR target/56892
12429 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12430 hook_bool_const_tree_true.
12431
2cf4c39e
AK		 124322013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12433
12434 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12435 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12436 * config/s390/s390.md: Define FPR*_REGNUM constants.
12437 Fix FPR2_REGNUM constant (18 -> 17).
12438 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12439 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12440 ("trunc<BFP:mode><DFP_ALL:mode>2")
12441 ("trunc<DFP_ALL:mode><BFP:mode>2")
12442 ("*extend<BFP:mode><DFP_ALL:mode>2")
12443 ("*extend<DFP_ALL:mode><BFP:mode>2")
12444 ("extend<BFP:mode><DFP_ALL:mode>2")
12445 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12446 FPR4_REGNUM.
12447
e6ac0270 124482013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 12449
6a2fa4b2 12450 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 12451
b89b22fc
AK		 124522013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12453
12454 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12455 and cfun_fpr_bit_p to cfun_fpr_save_p.
12456 (s390_frame_area, s390_register_info, s390_frame_info)
12457 (s390_emit_prologue, s390_emit_epilogue)
12458 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12459 register numbers.
12460 * config/s390/s390.h: Define *_REGNUM macros for floating point
12461 register numbers.
12462
07f398aa
EB		 124632013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12464
12465 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12466
75fefa91
PCC		 124672013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12468
12469 PR rtl-optimization/57786
12470 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12471 and break out of the loop when it is set to false.
12472
046f1eee
JJ		 124732013-07-08 Jakub Jelinek <jakub@redhat.com>
12474
8140c065
JJ		 12475 PR target/57819
12476 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12477 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12478 (const_int 63)) 0)).
12479 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12480 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12481 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12482
046f1eee
JJ		 12483 PR rtl-optimization/57829
12484 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12485 mask bits outside of mode are just sign-extension from mode to HWI.
12486
930b700b
MZ		 124872013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12488
12489 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12490 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12491 adjust_address instead of change_address to keep info about alignment.
12492 (emit_strmov): Remove.
12493 (emit_memmov): New function.
12494 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12495 (expand_movmem_epilogue): Likewise and return updated rtx for
12496 destination.
12497 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12498 destination and source.
12499 (decide_alignment): Refactor, handle vector_loop.
12500 (ix86_expand_movmem): Likewise.
12501 (ix86_expand_setmem): Likewise.
12502 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12503
7b9d1bd8
UB		 125042013-07-07 Uros Bizjak <ubizjak@gmail.com>
12505
12506 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12507 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12508
eabb5f48
UB		 125092013-07-06 Uros Bizjak <ubizjak@gmail.com>
12510
12511 * config/i386/sse.md (sse_movlhps): Change alternative 3
12512 of operand 2 to "m".
12513
125142013-07-06 Uros Bizjak <ubizjak@gmail.com>
12515
12516 PR target/57807
12517 * config/i386/sse.md (iptr): New mode attribute.
12518 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12519 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12520 (<sse>_vmmul<mode>3): Ditto.
12521 (<sse>_vmdiv<mode>3): Ditto.
12522 (sse_vmrcpv4sf2): Ditto.
12523 (<sse>_vmsqrt<mode>2): Ditto.
12524 (sse_vmrsqrtv4sf2): Ditto.
12525 (<sse>_vm<code><mode>3): Ditto.
12526 (avx_vmcmp<mode>3): Ditto.
12527 (<sse>_vmmaskcmp<mode>3): Ditto.
12528 (<sse>_comi): Ditto.
12529 (<sse>_ucomi): Ditto.
12530 (*xop_vmfrcz_<mode>): Ditto.
12531 (*fmai_fmadd_<mode>): Ditto.
12532 (*fmai_fmsub_<mode>): Ditto.
12533 (*fmai_fnmadd_<mode>): Ditto.
12534 (*fmai_fnmsub_<mode>): Ditto.
12535 (*fma4i_vmfmadd_<mode>): Ditto.
12536 (*fma4i_vmfmsub_<mode>): Ditto.
12537 (*fma4i_vmfnmadd_<mode>): Ditto.
12538 (*fma4i_vmfnmsub_<mode>): Ditto.
12539 (*xop_vmfrcz_<mode>): Ditto.
12540 (sse_cvtps2pi): Ditto.
12541 (sse_cvttps2pi): Ditto.
12542 (sse_cvtss2si): Ditto.
12543 (sse_cvtss2si_2): Ditto.
12544 (sse_cvtss2siq_2): Ditto.
12545 (sse_cvttss2si): Ditto.
12546 (sse_cvttss2siq): Ditto.
12547 (sse_cvtsd2si): Ditto.
12548 (sse_cvtsd2si_2): Ditto.
12549 (sse_cvtsd2siq_2): Ditto.
12550 (sse_cvttsd2si): Ditto.
12551 (sse_cvttsd2siq): Ditto.
12552 (sse_cvtsd2ss): Ditto.
12553 (sse_cvtss2sd): Ditto.
12554 (avx2_pbroadcast<mode>): Ditto.
12555 (avx2_pbroadcast<mode>_1): Ditto.
12556 (*avx_vperm_broadcast_v4sf): Ditto.
12557
12558 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12559 (sse_movlhps): Ditto.
12560 (sse_storehps): Ditto.
12561 (sse_loadhps): Ditto.
12562 (sse_storelps): Ditto.
12563 (sse_loadlps): Ditto.
12564 (*vec_concatv4sf): Ditto.
12565 (*vec_interleave_highv2df): Ditto.
12566 (*vec_interleave_lowv2df): Ditto.
12567 (*vec_extractv2df_1_sse): Ditto.
12568 (*vec_extractv2df_0_sse): Ditto.
12569 (sse2_storelpd): Ditto.
12570 (sse2_loadlpd): Ditto.
12571 (sse2_movsd): Ditto.
12572 (*vec_concatv4si): Ditto.
12573 (vec_concatv2di): Ditto.
12574
12575 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12576 for Intel asm dialect.
12577 (mmx_punpcklwd): Ditto.
12578 (mmx_punpckldq): Ditto.
12579
12580 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12581 for intel assembler dialect.
12582
1f6eac90
JJ		 125832013-07-06 Jakub Jelinek <jakub@redhat.com>
12584
12585 PR target/29776
12586 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12587 for BUILT_IN_C{LZ,LRSB}*.
12588 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12589 * tree-vrp.c (extract_range_basic): Handle
12590 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12591 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12592 fall thru to code calling set_value*.
12593 * builtins.c (expand_builtin): Remove *IMAX cases.
12594 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12595 if width is bigger than 2*HWI.
12596
2b778c9d
VM		 125972013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12598
12599 PR rtl-optimization/55342
12600 * lra-int.h (lra_subreg_reload_pseudos): New.
12601 * lra.c: Add undoing optional reloads to the block diagram.
12602 (lra_subreg_reload_pseudos): New.
12603 (lra_optional_reload_pseudos): Change comments.
12604 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12605 lra_optional_reload_pseudos after undo transformations.
12606 * lra-assigns.c (pseudo_prefix_title): New.
12607 (lra_setup_reg_renumber): Use it.
12608 (spill_for): Ditto. Check subreg reload pseudos too.
12609 (assign_by_spills): Consider subreg reload pseudos too.
12610 * lra-constraints.c (simplify_operand_subreg): Use
12611 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12612 (curr_insn_transform): Recognize and do optional reloads.
12613 (undo_optional_reloads): New.
12614 (lra_undo_inheritance): Call undo_optional_reloads.
12615
207156e4
TQ		 126162013-07-05 Thomas Quinot <quinot@adacore.com>
12617
12618 * tree-complex.c (expand_complex_operations_1): Fix typo.
12619
8990e73a
TB		 126202013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12621
12622 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12623 (tune_params): New member 'const vec_costs'.
12624 * config/aarch64/aarch64.c (generic_vector_cost): New.
12625 (generic_tunings): New member 'generic_vector_cost'.
12626 (aarch64_builtin_vectorization_cost): New.
12627 (aarch64_add_stmt_cost): New.
12628 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12629 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12630
f7bc421d
JJ		 126312013-07-03 Jakub Jelinek <jakub@redhat.com>
12632
12633 PR target/57777
12634 * config/i386/predicates.md (vsib_address_operand): Disallow
12635 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12636
e90d1568
HPN		 126372013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12638
12639 PR middle-end/55030
12640 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12641 expand_builtin_setjmp_receiver.
12642 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12643 with NULL for the label parameter.
12644 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12645 the frame-pointer. Adjust comments.
12646 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12647 only if LABEL is non-NULL.
12648
cbd7413b
YZ		 126492013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12650
12651 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12652 (ARM_ABI_AAPCS64): Ditto.
12653 (arm_abi): Ditto.
12654 (ARM_DEFAULT_ABI): Ditto.
12655
dec11868
JG		 126562013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12657
12658 * config/aarch64/aarch64-builtins.c
12659 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12660 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12661 (st1): Likewise.
12662 * config/aarch64/aarch64-simd.md
12663 (aarch64_ld1<VALL:mode>): New.
12664 (aarch64_st1<VALL:mode>): Likewise.
12665 * config/aarch64/arm_neon.h
12666 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12667
786fb9b6
ST		 126682013-07-02 Sriraman Tallam <tmsriram@google.com>
12669
0ba6aff3
UB		 12670 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12671 target ISA is AVX.
786fb9b6
ST		 12672 (ix86_option_override_internal):Turn on all -mavx target flags by
12673 default as they are dependent on AVX anyway.
12674
4b043553
CC		 126752013-07-02 Cary Coutant <ccoutant@google.com>
12676
12677 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12678 deterministic hash.
12679 (loc_checksum_ordered): Likewise.
12680 (hash_loc_operands): Remove inline keyword.
12681
c4e87a13
JJ		 126822013-07-02 Jakub Jelinek <jakub@redhat.com>
12683
12684 PR tree-optimization/57741
12685 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12686 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12687 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12688 Allow REAL_CST step_exprs if flag_associative_math.
12689 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12690
26366d28
IB		 126912013-07-02 Ian Bolton <ian.bolton@arm.com>
12692
0ba6aff3 12693 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 12694
2879bb2b
IB		 126952013-07-02 Ian Bolton <ian.bolton@arm.com>
12696
12697 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12698
afc5e8a6
KT		 126992013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12700
12701 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12702 encoding.
12703 (iorsi3_insn): Likewise.
12704 (arm_xorsi3): Likewise.
12705
31f8442b
SN		 127062013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12707
12708 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12709 to ...
12710 (attribute "type"): ... here, and prefix with "wmmx_".
12711 (attribute "core_cycles"): Update for attribute changes.
12712 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12713 (tbcstv4hi): Likewise.
12714 (tbcstv2si): Likewise.
12715 (iwmmxt_iordi3): Likewise.
12716 (iwmmxt_xordi3): Likewise.
12717 (iwmmxt_anddi3): Likewise.
12718 (iwmmxt_nanddi3): Likewise.
12719 (iwmmxt_arm_movdi): Likewise.
12720 (iwmmxt_movsi_insn): Likewise.
12721 (mov<mode>_internal): Likewise.
12722 (and<mode>3_iwmmxt): Likewise.
12723 (ior<mode>3_iwmmxt): Likewise.
12724 (xor<mode>3_iwmmxt): Likewise.
12725 (add<mode>3_iwmmxt): Likewise.
12726 (ssaddv8qi3): Likewise.
12727 (ssaddv4hi3): Likewise.
12728 (ssaddv2si3): Likewise.
12729 (usaddv8qi3): Likewise.
12730 (usaddv4hi3): Likewise.
12731 (usaddv2si3): Likewise.
12732 (sub<mode>3_iwmmxt): Likewise.
12733 (sssubv8qi3): Likewise.
12734 (sssubv4hi3): Likewise.
12735 (sssubv2si3): Likewise.
12736 (ussubv8qi3): Likewise.
12737 (ussubv4hi3): Likewise.
12738 (ussubv2si3): Likewise.
12739 (mulv4hi3_iwmmxt): Likewise.
12740 (smulv4hi3_highpart): Likewise.
12741 (umulv4hi3_highpart): Likewise.
12742 (iwmmxt_wmacs): Likewise.
12743 (iwmmxt_wmacsz): Likewise.
12744 (iwmmxt_wmacu): Likewise.
12745 (iwmmxt_wmacuz): Likewise.
12746 (iwmmxt_clrdi): Likewise.
12747 (iwmmxt_clrv8qi): Likewise.
12748 (iwmmxt_clr4hi): Likewise.
12749 (iwmmxt_clr2si): Likewise.
12750 (iwmmxt_uavgrndv8qi3): Likewise.
12751 (iwmmxt_uavgrndv4hi3): Likewise.
12752 (iwmmxt_uavgv8qi3): Likewise.
12753 (iwmmxt_uavgv4hi3): Likewise.
12754 (iwmmxt_tinsrb): Likewise.
12755 (iwmmxt_tinsrh): Likewise.
12756 (iwmmxt_tinsrw): Likewise.
12757 (iwmmxt_textrmub): Likewise.
12758 (iwmmxt_textrmsb): Likewise.
12759 (iwmmxt_textrmuh): Likewise.
12760 (iwmmxt_textrmsh): Likewise.
12761 (iwmmxt_textrmw): Likewise.
12762 (iwmxxt_wshufh): Likewise.
12763 (eqv8qi3): Likewise.
12764 (eqv4hi3): Likewise.
12765 (eqv2si3): Likewise.
12766 (gtuv8qi3): Likewise.
12767 (gtuv4hi3): Likewise.
12768 (gtuv2si3): Likewise.
12769 (gtv8qi3): Likewise.
12770 (gtv4hi3): Likewise.
12771 (gtv2si3): Likewise.
12772 (smax<mode>3_iwmmxt): Likewise.
12773 (umax<mode>3_iwmmxt): Likewise.
12774 (smin<mode>3_iwmmxt): Likewise.
12775 (umin<mode>3_iwmmxt): Likewise.
12776 (iwmmxt_wpackhss): Likewise.
12777 (iwmmxt_wpackwss): Likewise.
12778 (iwmmxt_wpackdss): Likewise.
12779 (iwmmxt_wpackhus): Likewise.
12780 (iwmmxt_wpackwus): Likewise.
12781 (iwmmxt_wpackdus): Likewise.
12782 (iwmmxt_wunpckihb): Likewise.
12783 (iwmmxt_wunpckihh): Likewise.
12784 (iwmmxt_wunpckihw): Likewise.
12785 (iwmmxt_wunpckilb): Likewise.
12786 (iwmmxt_wunpckilh): Likewise.
12787 (iwmmxt_wunpckilw): Likewise.
12788 (iwmmxt_wunpckehub): Likewise.
12789 (iwmmxt_wunpckehuh): Likewise.
12790 (iwmmxt_wunpckehuw): Likewise.
12791 (iwmmxt_wunpckehsb): Likewise.
12792 (iwmmxt_wunpckehsh): Likewise.
12793 (iwmmxt_wunpckehsw): Likewise.
12794 (iwmmxt_wunpckelub): Likewise.
12795 (iwmmxt_wunpckeluh): Likewise.
12796 (iwmmxt_wunpckeluw): Likewise.
12797 (iwmmxt_wunpckelsb): Likewise.
12798 (iwmmxt_wunpckelsh): Likewise.
12799 (iwmmxt_wunpckelsw): Likewise.
12800 (ror<mode>3): Likewise.
12801 (ashr<mode>3_iwmmxt): Likewise.
12802 (lshr<mode>3_iwmmxt): Likewise.
12803 (ashl<mode>3_iwmmxt): Likewise.
12804 (ror<mode>3_di): Likewise.
12805 (ashr<mode>3_di): Likewise.
12806 (lshr<mode>3_di): Likewise.
12807 (ashl<mode>3_di): Likewise.
12808 (iwmmxt_wmadds): Likewise.
12809 (iwmmxt_wmaddu): Likewise.
12810 (iwmmxt_tmia): Likewise.
12811 (iwmmxt_tmiaph): Likewise.
12812 (iwmmxt_tmiabb): Likewise.
12813 (iwmmxt_tmiatb): Likewise.
12814 (iwmmxt_tmiabt): Likewise.
12815 (iwmmxt_tmiatt): Likewise.
12816 (iwmmxt_tmovmskb): Likewise.
12817 (iwmmxt_tmovmskh): Likewise.
12818 (iwmmxt_tmovmskw): Likewise.
12819 (iwmmxt_waccb): Likewise.
12820 (iwmmxt_wacch): Likewise.
12821 (iwmmxt_waccw): Likewise.
12822 (iwmmxt_waligni): Likewise.
12823 (iwmmxt_walignr): Likewise.
12824 (iwmmxt_walignr0): Likewise.
12825 (iwmmxt_walignr1): Likewise.
12826 (iwmmxt_walignr2): Likewise.
12827 (iwmmxt_walignr3): Likewise.
12828 (iwmmxt_wsadb): Likewise.
12829 (iwmmxt_wsadh): Likewise.
12830 (iwmmxt_wsadbz): Likewise.
12831 (iwmmxt_wsadhz): Likewise.
12832 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12833 (iwmmxt_wabsdiffb): Likewise.
12834 (iwmmxt_wabsdiffh): Likewise.
12835 (iwmmxt_wabsdiffw): Likewise.
12836 (iwmmxt_waddsubhx): Likewise
12837 (iwmmxt_wsubaddhx): Likewise.
12838 (addc<mode>3): Likewise.
12839 (iwmmxt_avg4): Likewise.
12840 (iwmmxt_avg4r): Likewise.
12841 (iwmmxt_wmaddsx): Likewise.
12842 (iwmmxt_wmaddux): Likewise.
12843 (iwmmxt_wmaddsn): Likewise.
12844 (iwmmxt_wmaddun): Likewise.
12845 (iwmmxt_wmulwsm): Likewise.
12846 (iwmmxt_wmulwum): Likewise.
12847 (iwmmxt_wmulsmr): Likewise.
12848 (iwmmxt_wmulumr): Likewise.
12849 (iwmmxt_wmulwsmr): Likewise.
12850 (iwmmxt_wmulwumr): Likewise.
12851 (iwmmxt_wmulwl): Likewise.
12852 (iwmmxt_wqmulm): Likewise.
12853 (iwmmxt_wqmulwm): Likewise.
12854 (iwmmxt_wqmulmr): Likewise.
12855 (iwmmxt_wqmulwmr): Likewise.
12856 (iwmmxt_waddbhusm): Likewise.
12857 (iwmmxt_waddbhusl): Likewise.
12858 (iwmmxt_wqmiabb): Likewise.
12859 (iwmmxt_wqmiabt): Likewise.
12860 (iwmmxt_wqmiatb): Likewise.
12861 (iwmmxt_wqmiatt): Likewise.
12862 (iwmmxt_wqmiabbn): Likewise.
12863 (iwmmxt_wqmiabtn): Likewise.
12864 (iwmmxt_wqmiatbn): Likewise.
12865 (iwmmxt_wqmiattn): Likewise.
12866 (iwmmxt_wmiabb): Likewise.
12867 (iwmmxt_wmiabt): Likewise.
12868 (iwmmxt_wmiatb): Likewise.
12869 (iwmmxt_wmiatt): Likewise.
12870 (iwmmxt_wmiabbn): Likewise.
12871 (iwmmxt_wmiabtn): Likewise.
12872 (iwmmxt_wmiatbn): Likewise.
12873 (iwmmxt_wmiattn): Likewise.
12874 (iwmmxt_wmiawbb): Likewise.
12875 (iwmmxt_wmiawbt): Likewise.
12876 (iwmmxt_wmiawtb): Likewise.
12877 (iwmmxt_wmiawtt): Likewise.
12878 (iwmmxt_wmiawbbn): Likewise.
12879 (iwmmxt_wmiawbtn): Likewise.
12880 (iwmmxt_wmiawtbn): Likewise.
12881 (iwmmxt_wmiawttn): Likewise.
12882 (iwmmxt_wmerge): Likewise.
12883 (iwmmxt_tandc<mode>3): Likewise.
12884 (iwmmxt_torc<mode>3): Likewise.
12885 (iwmmxt_torvsc<mode>3): Likewise.
12886 (iwmmxt_textrc<mode>3): Likewise.
12887 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12888 (wmmxt_pack): Likewise.
12889 (wmmxt_mult_c1): Likewise.
12890 (wmmxt_mult_c2): Likewise.
12891 (wmmxt_alu_c1): Likewise.
12892 (wmmxt_alu_c2): Likewise.
12893 (wmmxt_alu_c3): Likewise.
12894 (wmmxt_transfer_c1): Likewise.
12895 (wmmxt_transfer_c2): Likewise.
12896 (wmmxt_transfer_c3): Likewise.
12897 (marvell_f_iwmmxt_wstr): Likewise.
12898 (marvell_f_iwmmxt_wldr): Likewise.
12899
bdb7bf8a
YZ		 129002013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12901
0ba6aff3 12902 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 12903
e6631d38
VM		 129042013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12905
12906 Revert:
12907 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12908 * lra-constraints.c (need_for_split_p): Check call used hard regs
12909 living through calls.
12910
12911 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12912 call used regs for call insn.
12913
53c1275b
JJ		 129142013-06-28 Jakub Jelinek <jakub@redhat.com>
12915
12916 PR target/57736
0ba6aff3
UB		 12917 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12918 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 12919
1141ed3f
BI		 129202013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12921
12922 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 12923
e32e4c4a
VM		 129242013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12925
12926 * lra-constraints.c (need_for_split_p): Check call used hard regs
12927 living through calls.
12928
f161bfd3
MM		 129292013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
12930
12931 PR target/57744
12932 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
12933 to tie with any other modes. Eliminate Altivec vector mode tests,
12934 since these are a subset of ALTIVEC or VSX vector modes. Simplify
12935 code, to return 0 if testing MODE2 for a condition, if we've
12936 already tested MODE1 for the same condition.
12937
7eda14e1
MS		 129382013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12939
12940 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
12941 layout.
12942
c822f852
MS		 129432013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12944
12945 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12946 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
12947
da4f13a4
MS		 129482013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12949
0ba6aff3
UB		 12950 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
12951 Define.
da4f13a4
MS		 12952 (aarch64_symbolic_constant_p): Remove.
12953 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
12954 static. Fix line length and white space.
12955 (aarch64_symbolic_constant_p): Remove.
12956 * config/aarch64/predicates.md (aarch64_valid_symref):
12957 Use aarch64_classify_symbol_expression.
12958
956a95a5
KT		 129592013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12960
12961 * config/arm/constraints.md (Ts): New constraint.
12962 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
12963 16-bit encodings.
12964 (compare_scc): Use "Ts" constraint for operand 0.
12965 (ior_scc_scc): Likewise.
12966 (and_scc_scc): Likewise.
12967 (and_scc_scc_nodom): Likewise.
12968 (ior_scc_scc_cmp): Likewise for operand 7.
12969 (and_scc_scc_cmp): Likewise.
12970 * config/arm/thumb2.md (thumb2_movsi_insn):
12971 Add alternatives for 16-bit encodings.
12972 (thumb2_movhi_insn): Likewise.
12973 (thumb2_movsicc_insn): Likewise.
12974 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
12975 (thumb2_negscc): Use "Ts" constraint.
12976 Move mvn instruction outside cond_exec block.
12977 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
12978 for 16-bit encodings.
12979
31dfce10
KT		 129802013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12981
12982 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
12983 encoding.
12984 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
12985 (mulsi3subsi): Likewise.
12986 (mulsidi3adddi): Likewise.
12987 (mulsidi3_v6): Likewise.
12988 (umulsidi3_v6): Likewise.
12989 (umulsidi3adddi_v6): Likewise.
12990 (smulsi3_highpart_v6): Likewise.
12991 (umulsi3_highpart_v6): Likewise.
12992 (mulhisi3tb): Likewise.
12993 (mulhisi3bt): Likewise.
12994 (mulhisi3tt): Likewise.
12995 (maddhisi4): Likewise.
12996 (maddhisi4tb): Likewise.
12997 (maddhisi4tt): Likewise.
12998 (maddhidi4): Likewise.
12999 (maddhidi4tb): Likewise.
13000 (maddhidi4tt): Likewise.
13001 (zeroextractsi_compare0_scratch): Likewise.
13002 (insv_zero): Likewise.
13003 (insv_t2): Likewise.
13004 (anddi_notzesidi_di): Likewise.
13005 (anddi_notsesidi_di): Likewise.
13006 (andsi_notsi_si): Likewise.
13007 (iordi_zesidi_di): Likewise.
13008 (xordi_zesidi_di): Likewise.
13009 (andsi_iorsi3_notsi): Likewise.
13010 (smax_0): Likewise.
13011 (smax_m1): Likewise.
13012 (smin_0): Likewise.
13013 (not_shiftsi): Likewise.
13014 (unaligned_loadsi): Likewise.
13015 (unaligned_loadhis): Likewise.
13016 (unaligned_loadhiu): Likewise.
13017 (unaligned_storesi): Likewise.
13018 (unaligned_storehi): Likewise.
13019 (extv_reg): Likewise.
13020 (extzv_t2): Likewise.
13021 (divsi3): Likewise.
13022 (udivsi3): Likewise.
13023 (arm_zero_extendhisi2addsi): Likewise.
13024 (arm_zero_extendqisi2addsi): Likewise.
13025 (compareqi_eq0): Likewise.
13026 (arm_extendhisi2_v6): Likewise.
13027 (arm_extendqisi2addsi): Likewise.
13028 (arm_movt): Likewise.
13029 (thumb2_ldrd): Likewise.
13030 (thumb2_ldrd_base): Likewise.
13031 (thumb2_ldrd_base_neg): Likewise.
13032 (thumb2_strd): Likewise.
13033 (thumb2_strd_base): Likewise.
13034 (thumb2_strd_base_neg): Likewise.
13035 (arm_negsi2): Add alternative for 16-bit encoding.
13036 (arm_one_cmplsi2): Likewise.
13037
1572e697
KT		 130382013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13039
13040 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13041 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13042 (movdfcc): Likewise.
13043 * config/arm/vfp.md (*thumb2_movsf_vfp):
13044 Disable predication for arm_restrict_it.
13045 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13046 (*thumb2_movdfcc_vfp): Likewise.
13047 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13048 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13049 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13050 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13051 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13052 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13053 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13054 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13055 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13056 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13057 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13058 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13059 Disable predication for arm_restrict_it.
13060
ebf8f0ea
KY		 130612013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13062
13063 * config/i386/bmiintrin.h (_bextr_u32): New.
13064 (_bextr_u64): Ditto.
13065
a290fcda
RS		 130662013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13067
13068 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13069 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13070 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13071 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13072 * config/mips/n32-elf.h: ...this new file.
13073
ea169996
MG		 130742013-06-27 Marc Glisse <marc.glisse@inria.fr>
13075
13076 PR target/57224
13077 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13078 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13079
12211b99 130802013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 13081
13082 * config/mips/mips-tables.opt: Regenerate.
13083 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 13084 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 13085 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 13086
2444244e
JJ		 130872013-06-27 Jakub Jelinek <jakub@redhat.com>
13088
cd06d2a2
JJ		 13089 PR target/57623
13090 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13091 constraints of operand 1 and 2.
13092
2444244e
JJ		 13093 PR target/57623
13094 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13095 to match RTL canonicalization. Swap predicates and
13096 constraints of operand 1 and 2.
13097
2f259720
VM		 130982013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13099
0ba6aff3
UB		 13100 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13101 Process OP_INOUT regs for splitting too.
2f259720 13102
fef4d2b3
JJ		 131032013-06-27 Jakub Jelinek <jakub@redhat.com>
13104
13105 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13106 decl before the loop, initialize to NULL.
13107 (vectorizable_load): Initialize ptr_incr to NULL.
13108
79ee9826
MJ		 131092013-06-27 Martin Jambor <mjambor@suse.cz>
13110
13111 PR lto/57208
13112 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13113 * ipa-ref.c (ipa_maybe_record_reference): New function.
13114 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13115 * ipa-cp.c (create_specialized_node): Record potential references from
13116 aggvals.
13117 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13118
e18b4a81
YZ		 131192013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13120
13121 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13122 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13123 'mode' to force_reg.
13124 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13125 (aarch64_expand_mov_immediate): Likewise.
13126
9c023bf0
YZ		 131272013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13128
13129 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13130 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13131
5a3fe9b6
AK		 131322013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13133
13134 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13135 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13136 (struct machine_function): Add tbegin_p.
13137 (s390_canonicalize_comparison): Fold CC mode compares to
13138 conditional jump if possible.
13139 (s390_emit_jump): Return the emitted jump.
13140 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13141 Handle CCRAWmode compares.
13142 (s390_option_override): Default to -mhtm if available.
13143 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13144 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13145 FPRs instead of df_regs_ever_live_p.
13146 (s390_optimize_nonescaping_tx): New function.
13147 (s390_init_frame_layout): Extend clobbered_regs array to cover
13148 FPRs as well.
13149 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13150 (s390_expand_tbegin): New function.
13151 (enum s390_builtin): New enum definition.
13152 (code_for_builtin): New array definition.
13153 (s390_init_builtins): New function.
13154 (s390_expand_builtin): New function.
13155 (TARGET_INIT_BUILTINS): Define.
13156 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 13157 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 13158 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13159 (s390_alc_comparison): Likewise.
13160 * config/s390/s390-modes.def: Add CCRAWmode.
13161 * config/s390/s390.h (processor_flags): Add PF_TX.
13162 (TARGET_CPU_HTM): Define macro.
13163 (TARGET_HTM): Define macro.
13164 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13165 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13166 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13167 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 13168 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 13169 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13170 ("*cc_to_int"): Move up.
13171 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13172 constants other than 0.
13173 ("*ccraw_to_int"): New insn and splitter definition.
13174 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13175 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13176 ("tx_assist"): New expander.
13177 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13178 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13179 * config/s390/s390.opt: Add -mhtm option.
13180 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13181 * config/s390/htmxlintrin.h: New file.
13182 * config/s390/htmintrin.h: New file.
13183 * config/s390/s390intrin.h: New file.
13184 * doc/extend.texi: Document htm builtins.
13185 * config.gcc: Add the new header files to extra_headers.
13186
9e216629
TS		 131872013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13188
13189 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13190 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13191
fd8c65e7 131922013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 13193 Pat Haugen <pthaugen@us.ibm.com>
13194 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 13195
13196 * config/rs6000/power8.md: New.
13197 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13198 setting for power8 entry.
13199 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13200 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13201 test for Power4/Power5 only.
13202 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13203 support.
13204 (force_new_group): Adjust comment.
13205 * config/rs6000/rs6000.md: Include power8.md.
13206
b24a2ce5
GY		 132072013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13208
13209 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13210 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13211 declaration.
13212 (tune_params): Update comment.
13213 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13214 (arm_max_conditional_execute): New function.
13215 (thumb2_final_prescan_insn): Use max_insn_skipped and
13216 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13217
f6fdeb62
JJ		 132182013-06-25 Jakub Jelinek <jakub@redhat.com>
13219
13220 PR tree-optimization/57705
13221 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13222 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13223 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 13224 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13225
c13bc3d9
MJ		 132262013-06-25 Martin Jambor <mjambor@suse.cz>
13227
13228 PR middle-end/57670
13229 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13230 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13231 calls in the dump.
13232 (ipa_note_param_call): Initialize member_ptr flag.
13233 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13234 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13235 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13236 (ipa_read_indirect_edge_info): Likewise.
13237
8a845901
RB		 132382013-06-25 Richard Biener <rguenther@suse.de>
13239
13240 PR middle-end/56977
13241 * passes.c (init_optimization_passes): Move pass_fold_builtins
13242 and pass_dce earlier with -Og.
13243
33be0bec
EB		 132442013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13245
13246 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13247 <BIT_FIELD_REF>: Remove trailing TAB.
13248 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13249 remove blank line.
13250
5fe8e757
MJ		 132512013-06-24 Martin Jambor <mjambor@suse.cz>
13252
13253 PR tree-optimization/57358
13254 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13255 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13256 (ipa_analyze_params_uses): Generate pessimistic info when true.
13257
44a60244
MJ		 132582013-06-24 Martin Jambor <mjambor@suse.cz>
13259
13260 PR tree-optimization/57539
13261 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13262 global.inlined_to of the new node to it. All callers changed.
13263 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13264 inlining_into, pass it to cgraph_clone_node.
13265 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13266 ipa_free_edge_args_substructures.
13267 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13268 rdesc linked list. Do not assert rdesc edges have inlined caller.
13269 Assert we have found an rdesc in the rdesc list.
13270
7c5848b8
RB		 132712013-06-24 Richard Biener <rguenther@suse.de>
13272
0ba6aff3 13273 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 13274 (pointer_set_lookup): Declare.
13275 (class pointer_map): New template class implementing a
13276 generic pointer to T map.
13277 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13278 pointer_map<T>::contains, pointer_map<T>::insert,
13279 pointer_map<T>::traverse): New functions.
13280 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13281 (pointer_set_lookup): New function.
13282 (pointer_set_contains): Use pointer_set_lookup.
13283 (pointer_set_insert): Likewise.
13284 (insert_aux): Remove.
13285 (struct pointer_map_t): Embed a pointer_set_t.
13286 (pointer_map_create): Adjust.
13287 (pointer_map_destroy): Likewise.
13288 (pointer_map_contains): Likewise.
13289 (pointer_map_insert): Likewise.
13290 (pointer_map_traverse): Likewise.
13291 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13292 pointer_map<unsigned> instead of a pointer_map_t.
13293 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13294 (streamer_tree_cache_lookup): Likewise.
13295 (streamer_tree_cache_create): Likewise.
13296 (streamer_tree_cache_delete): Likewise.
13297 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13298 pointer_map<unsigned> instead of a pointer_map_t.
13299 (lto_init_tree_ref_encoder): Adjust.
13300 (lto_destroy_tree_ref_encoder): Likewise.
13301 * lto-section-out.c (lto_output_decl_index): Likewise.
13302 (lto_record_function_out_decl_state): Likewise.
13303 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13304
3dbc97a9
RB		 133052013-06-24 Richard Biener <rguenther@suse.de>
13306
13307 PR tree-optimization/57488
13308 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13309
8345965a
AM		 133102013-06-24 Alan Modra <amodra@gmail.com>
13311
13312 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13313 (gen_easy_altivec_constant): Likewise.
13314 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13315 easy_vector_constant_msb): Likewise.
13316
b7df379f
JJ		 133172013-06-23 Jakub Jelinek <jakub@redhat.com>
13318
13319 PR target/57688
13320 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13321 add missing return true.
13322
12211b99 133232013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 13324
13325 PR target/52483
13326 * config/sh/predicates.md (general_extend_operand): Invoke
13327 general_movsrc_operand for memory operands.
13328 (general_movsrc_operand): Allow reg+reg addressing, do not use
13329 general_operand for memory operands.
13330
97db2bf7
ST		 133312013-06-23 Sriraman Tallam <tmsriram@google.com>
13332
13333 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13334 when current target options does not apply.
13335 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13336 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 13337 * config/i386/bmiintrin.h: Pass appropriate target
13338 attributes to header.
97db2bf7
ST		 13339 * config/i386/mmintrin.h: Ditto.
13340 * config/i386/nmmintrin.h: Ditto.
13341 * config/i386/avx2intrin.h: Ditto.
13342 * config/i386/fxsrintrin.h: Ditto.
13343 * config/i386/tbmintrin.h: Ditto.
13344 * config/i386/xsaveintrin.h: Ditto.
13345 * config/i386/f16cintrin.h: Ditto.
13346 * config/i386/xtestintrin.h: Ditto.
13347 * config/i386/xsaveoptintrin.h: Ditto.
13348 * config/i386/bmi2intrin.h: Ditto.
13349 * config/i386/lzcntintrin.h: Ditto.
13350 * config/i386/smmintrin.h: Ditto.
13351 * config/i386/wmmintrin.h: Ditto.
13352 * config/i386/x86intrin.h: Remove all header include guards.
13353 * config/i386/prfchwintrin.h: Ditto.
13354 * config/i386/pmmintrin.h: Ditto.
13355 * config/i386/tmmintrin.h: Ditto.
13356 * config/i386/xmmintrin.h: Ditto.
13357 * config/i386/popcntintrin.h: Ditto.
13358 * config/i386/rdseedintrin.h: Ditto.
13359 * config/i386/ammintrin.h: Ditto.
13360 * config/i386/emmintrin.h: Ditto.
13361 * config/i386/immintrin.h: Remove all header include guards.
13362 * config/i386/fma4intrin.h: Ditto.
13363 * config/i386/lwpintrin.h: Ditto.
13364 * config/i386/xopintrin.h: Ditto.
13365 * config/i386/ia32intrin.h: Ditto.
13366 * config/i386/avxintrin.h: Ditto.
13367 * config/i386/rtmintrin.h: Ditto.
13368 * config/i386/fmaintrin.h: Ditto.
13369 * config/i386/mm3dnow.h: Ditto.
13370
495e6879
ST		 133712013-06-22 Sriraman Tallam <tmsriram@google.com>
13372
13373 * common/config/i386/i386-common.c: Handle LZCNT.
13374
11e8f687
AK		 133752013-06-22 Andi Kleen <ak@linux.intel.com>
13376
13377 * doc/extend.texi: Use __atomic_store_n instead of
13378 __atomic_store in HLE example.
13379
12211b99 133802013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 13381
13382 * config/sh/sh.c: Remove <cstdlib> workaround.
13383
927f908b
AK		 133842013-06-21 Andi Kleen <ak@linux.intel.com>
13385
0ba6aff3 13386 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13387
cca410b6
AK		 133882013-06-21 Andi Kleen <ak@linux.intel.com>
13389
13390 * doc/extend.texi: Document that __atomic_clear and
13391 __atomic_test_and_set should only be used with bool.
13392
d84db1ed
JH		 133932013-06-20 Jan Hubicka <jh@suse.cz>
13394
13395 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13396 types_same_for_odr.
13397 * tree.c (decls_same_for_odr): New function.
13398 (same_for_edr): New function.
13399 (types_same_for_odr): New function.
13400 (get_binfo_at_offset): Use it.
13401 * tree.h (types_same_for_odr): Declare.
13402
12211b99 134032013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 13404 Jason Merrill <jason@redhat.com>
13405
13406 * system.h: Include <cstdlib> as well as <stdlib.h>.
13407
49b4ff63
UB		 134082013-06-20 Uros Bizjak <ubizjak@gmail.com>
13409
13410 PR target/57655
13411 * config/i386/i386.c (construct_container): Report error if
13412 long double is used with disabled x87 float returns.
13413
7a89ad00
JH		 134142013-06-20 Jan Hubicka <jh@suse.cz>
13415
13416 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13417
12211b99 134182013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 13419
13420 PR rtl-optimization/57425
13421 PR rtl-optimization/57569
13422 * alias.c (write_dependence_p): Remove parameters mem_mode and
13423 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13424 Changed all callers.
13425 (canon_anti_dependence): Get comments and semantics in sync.
13426 Add parameter mem_canonicalized. Changed all callers.
13427 * rtl.h (canon_anti_dependence): Update prototype.
13428
d16e9a99
RB		 134292013-06-20 Richard Biener <rguenther@suse.de>
13430
13431 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13432 byte case, inline streamer_read_uchar and defer section
13433 overrun check.
13434
c3a9b91b
RB		 134352013-06-20 Richard Biener <rguenther@suse.de>
13436
13437 PR tree-optimization/57584
13438 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13439 SSA names into the expanded expression that take part in
13440 abnormal coalescing.
13441
371e77e3 134422013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 13443
13444 * gcov.c (print_usage): Handle new option.
13445 (process_args): Ditto.
13446 (get_gcov_intermediate_filename): New function.
13447 (output_intermediate_file): New function.
13448 (output_gcov_file): New function
13449 (generate_results): Handle new option.
13450 (release_function): Relase demangled name.
13451 (read_graph_file): Handle demangled name.
13452 (output_lines): Ditto.
13453 * doc/gcov.texi: Document gcov intermediate format.
13454
11f5b71a
VM		 134552013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13456
13457 PR bootstrap/57604
13458 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13459 (lra_emit_add): Use the functions. Add comment about Y as an
13460 address segment.
13461
e1f9340b
DE		 134622013-06-19 David Edelsohn <dje.gcc@gmail.com>
13463
13464 PR driver/57652
13465 * collect2.c (collect_atexit): New.
13466 (collect_exit): Delete.
13467 (main): Register collect_atexit with atexit.
13468 (collect_wait): Change collect_exit to exit.
13469 (do_wait): Same.
13470 * collect2.h (collect_exit): Delete.
13471 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13472
e3f9e0ac
WM		 134732013-06-19 Wei Mi <wmi@google.com>
13474
13475 PR rtl-optimization/57518
13476 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13477 if regno is used in paradoxical subreg.
13478 (update_equiv_regs): Check pdx_subregs[regno] before
13479 set a reg to be equivalent with a mem.
13480
a09f1a76
MK		 134812013-06-19 Matthias Klose <doko@ubuntu.com>
13482
13483 PR driver/57651
13484 * file-find.h (find_a_file): Add a mode parameter.
13485 * file-find.c (find_a_file): Likewise.
13486 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13487 with X_OK for the executables.
13488 * collect2.c (main): Call find_a_file with X_OK.
13489
7d18b0ad
SE		 134902013-06-19 Steve Ellcey <sellcey@mips.com>
13491
13492 PR target/56942
13493 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13494 Use NEXT_INSN instead of next_real_insn.
13495
6a6dac52
JH		 134962013-06-19 Jan Hubicka <jh@suse.cz>
13497
13498 * cgraph.h (const_value_known_p): Replace by ...
13499 (ctor_for_folding): .. this one.
13500 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13501 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13502 * expr.c (expand_expr_real_1): Likewise.
13503 (string_constant): Likewise.
13504 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13505 * ipa.c (process_references): Likewise.
13506 (symtab_remove_unreachable_nodes): Likewise.
13507 * ipa-inline-analysis.c (param_change_prob): Likewise.
13508 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13509 (get_base_constructor): Likwise.
13510 * varpool.c (varpool_remove_node): Likewise.
13511 (varpool_remove_initializer): LIkewise.
13512 (dump_varpool_node): LIkwise.
13513 (const_value_known_p): Rewrite to ...
13514 (ctor_for_folding): ... this one.
13515
216c12ab
JJ		 135162013-06-19 Jakub Jelinek <jakub@redhat.com>
13517
13518 PR driver/57651
13519 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13520 PERSONALITY in $PATH derived prefixes.
13521
d53e2f99
JL		 135222013-06-19 Jeff Law <law@redhat.com>
13523
ecdbf306
JL		 13524 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13525 in comment.
13526
d53e2f99
JL		 13527 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13528 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13529 booleans.
13530
89e7fc04
SN		 135312013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13532
13533 * config/arm/vfp.md: Move VFP instruction classification documentation
13534 to ...
13535 * config/arm/arm.md: ... here. Update instruction classification
13536 documentation.
13537
5e7f6aaa
RE		 135382013-06-19 Richard Earnshaw <rearnsha@arm.com>
13539
13540 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13541 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13542 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13543
5fdc1e5d
SB		 135442013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13545
13546 PR target/57609
13547 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13548 with NEXT_INSN. Use tablejump_p to check for jump table data
13549 insns.
13550
e952afba
PC		 135512013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13552
13553 PR c++/56544
13554 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13555 that now in C++ the value is correct per the C++ standards.
13556
3fd0303a
RB		 135572013-06-19 Richard Biener <rguenther@suse.de>
13558
13559 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13560 for global context.
13561
b04e4b08
AK		 135622013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13563
13564 Revert:
13565 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13566
13567 PR target/57609
13568 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13569 with next_active_insn.
13570
bef8491a
ST		 135712013-06-18 Sriraman Tallam <tmsriram@google.com>
13572
13573 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13574 functions are inlined during failures to flag an error.
13575 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13576 in early inline pass.
22fc7d3d 13577
c21b257d
L		 135782013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13579
13580 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13581 in comments.
13582
33255ae3
JB		 135832013-06-18 Julian Brown <julian@codesourcery.com>
13584
13585 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13586 Permit virtual register pre-reload if !strict.
13587 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13588 change.
13589 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13590 prototype.
13591 * config/arm/neon.md (movmisalign<mode>): Use
13592 neon_perm_struct_or_reg_operand instead of
13593 neon_struct_or_register_operand.
13594 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13595 neon_permissive_struct_operand instead of neon_struct_operand.
13596 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13597 neon_vector_mem_operand.
13598 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13599 neon_vector_mem_operand.
13600 (neon_permissive_struct_operand): New.
13601 (neon_struct_or_register_operand): Rename to...
13602 (neon_perm_struct_or_reg_operand): This. Adjust call to
13603 neon_vector_mem_operand.
13604
d579fcda
RB		 136052013-06-18 Richard Biener <rguenther@suse.de>
13606
13607 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13608 * lto-streamer.h: Include pointer-set.h.
13609 (struct lto_decl_slot): Remove.
13610 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13611 Remove next_index entry.
13612 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13613 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13614 (lto_init_tree_ref_encoder): Adjust.
13615 (lto_destroy_tree_ref_encoder): Likewise.
13616 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13617 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13618 (lto_output_decl_index): Adjust.
13619 (lto_new_out_decl_state): Likewise.
13620 (lto_record_function_out_decl_state): Likewise.
13621 * lto-streamer-out.c (copy_function): Likewise.
13622
31ee20ba
RB		 136232013-06-18 Richard Biener <rguenther@suse.de>
13624
13625 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13626 * cgraphunit.c: Include cfgloop.h.
13627 (init_lowered_empty_function): Initialize the loop tree.
13628 (assemble_thunk): Insert new BBs into loops.
13629
ac0511f2
RB		 136302013-06-18 Richard Biener <rguenther@suse.de>
13631
13632 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13633 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13634 the map from cache entry to cache index optional.
13635 (streamer_tree_cache_replace_tree): Adjust accordingly.
13636 (streamer_tree_cache_append): Likewise.
13637 (streamer_tree_cache_delete): Likewise.
13638 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13639 streamer cache map from cache entry to cache index.
13640 * lto-streamer-out.c (create_output_block): Adjust.
13641
09485a08
SN		 136422013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13643
22fc7d3d
UB		 13644 * config/arm/arm.md (attribute "insn"): Move multiplication and
13645 division attributes to...
09485a08
SN		 13646 (attribute "type"): ... here. Remove mult.
13647 (attribute "mul32"): New attribute.
13648 (attribute "mul64"): Add umaal.
13649 (*arm_mulsi3): Update attributes.
13650 (*arm_mulsi3_v6): Likewise.
13651 (*thumb_mulsi3): Likewise.
13652 (*thumb_mulsi3_v6): Likewise.
13653 (*mulsi3_compare0): Likewise.
13654 (*mulsi3_compare0_v6): Likewise.
13655 (*mulsi_compare0_scratch): Likewise.
13656 (*mulsi_compare0_scratch_v6): Likewise.
13657 (*mulsi3addsi): Likewise.
13658 (*mulsi3addsi_v6): Likewise.
13659 (*mulsi3addsi_compare0): Likewise.
13660 (*mulsi3addsi_compare0_v6): Likewise.
13661 (*mulsi3addsi_compare0_scratch): Likewise.
13662 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13663 (*mulsi3subsi): Likewise.
13664 (*mulsidi3adddi): Likewise.
13665 (*mulsi3addsi_v6): Likewise.
13666 (*mulsidi3adddi_v6): Likewise.
13667 (*mulsidi3_nov6): Likewise.
13668 (*mulsidi3_v6): Likewise.
13669 (*umulsidi3_nov6): Likewise.
13670 (*umulsidi3_v6): Likewise.
13671 (*umulsidi3adddi): Likewise.
13672 (*umulsidi3adddi_v6): Likewise.
13673 (*smulsi3_highpart_nov6): Likewise.
13674 (*smulsi3_highpart_v6): Likewise.
13675 (*umulsi3_highpart_nov6): Likewise.
13676 (*umulsi3_highpart_v6): Likewise.
13677 (mulhisi3): Likewise.
13678 (*mulhisi3tb): Likewise.
13679 (*mulhisi3bt): Likewise.
13680 (*mulhisi3tt): Likewise.
13681 (maddhisi4): Likewise.
13682 (*maddhisi4tb): Likewise.
13683 (*maddhisi4tt): Likewise.
13684 (maddhidi4): Likewise.
13685 (*maddhidi4tb): Likewise.
13686 (*maddhidi4tt): Likewise.
13687 (divsi3): Likewise.
13688 (udivsi3): Likewise.
13689 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13690 (thumb2_mulsi_short_compare0): Likewise.
13691 (thumb2_mulsi_short_compare0_scratch): Likewise.
13692 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13693 (1020mult2): Likewise.
13694 (1020mult3): Likewise.
13695 (1020mult4): Likewise.
13696 (1020mult5): Likewise.
13697 (1020mult6): Likewise.
22fc7d3d
UB		 13698 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13699 change.
09485a08
SN		 13700 (cortex_a15_mult64): Likewise.
13701 (cortex_a15_sdiv): Likewise.
13702 (cortex_a15_udiv): Likewise.
13703 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13704 (mult2): Likewise.
13705 (mult3): Likewise.
13706 (mult4): Likewise.
13707 (mult5): Likewise.
13708 (mult6): Likewise.
13709 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13710 (pj4_ir_div): Likewise.
13711 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13712 (11_mult2): Likewise.
13713 (11_mult3): Likewise.
13714 (11_mult4): Likewise.
13715 (11_mult5): Likewise.
13716 (11_mult6): Likewise.
13717 (11_mult7): Likewise.
13718 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13719 (cortex_a8_mla): Likewise.
13720 (cortex_a8_mull): Likewise.
13721 (cortex_a8_smulwy): Likewise.
13722 (cortex_a8_smlald): Likewise.
13723 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13724 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13725 (cortex_r4_mul_3): Likewise.
13726 (cortex_r4_mla_4): Likewise.
13727 (cortex_r4_mla_3): Likewise.
13728 (cortex_r4_smlald): Likewise.
13729 (cortex_r4_mull): Likewise.
13730 (cortex_r4_sdiv): Likewise.
13731 (cortex_r4_udiv): Likewise.
13732 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13733 (cortex_a7_idiv): Likewise.
13734 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13735 (9_mult2): Likewise.
13736 (9_mult3): Likewise.
13737 (9_mult4): Likewise.
13738 (9_mult5): Likewise.
13739 (9_mult6): Likewise.
13740 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13741 (cortex_a53_sdiv): Likewise.
13742 (cortex_a53_udiv): Likewise.
13743 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13744 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13745 (mp626_mult2): Likewise.
13746 (mp626_mult3): Likewise.
13747 (mp626_mult4): Likewise.
13748 * config/arm/fa526.md (526_mult1): Update attribute change.
13749 (526_mult2): Likewise.
13750 * config/arm/arm-generic.md (mult): Update attribute change.
13751 (mult_ldsched_strongarm): Likewise.
13752 (mult_ldsched): Likewise.
13753 (multi_cycle): Likewise.
13754 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13755 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13756 (606te_mult2): Likewise.
13757 (606te_mult3): Likewise.
13758 (606te_mult4): Likewise.
13759 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13760 (cortex_a9_mac16): Likewise.
13761 (cortex_a9_multiply): Likewise.
13762 (cortex_a9_mac): Likewise.
13763 (cortex_a9_multiply_long): Likewise.
13764 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13765 (626te_mult2): Likewise.
13766 (626te_mult3): Likewise.
13767 (626te_mult4): Likewise.
13768
8951f345
RB		 137692013-06-18 Richard Biener <rguenther@suse.de>
13770
13771 PR lto/57334
13772 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13773
d8101d05
AK		 137742013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13775
13776 PR target/57609
13777 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13778 with next_active_insn.
13779
69eff9da
AM		 137802013-06-18 Alan Modra <amodra@gmail.com>
13781
13782 * config/rs6000/rs6000.h (enum data_align): New.
13783 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13784 (DATA_ABI_ALIGNMENT): Define.
13785 (CONSTANT_ALIGNMENT): Correct comment.
13786 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13787 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13788
8ed00d76
DM		 137892013-06-17 David Malcolm <dmalcolm@redhat.com>
13790
13791 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13792 ATTRIBUTE_UNUSED marking.
13793
3d33d06b
SN		 137942013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13795
13796 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13797 alternative and update.
13798 (aarch64_dup_lanedi): Delete.
13799 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13800 * config/aarch64/aarch64-simd-builtins.def: Update.
13801
ee03e71d
RB		 138022013-06-17 Richard Biener <rguenther@suse.de>
13803
13804 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13805 (lto_input_scc): Declare.
13806 (lto_input_tree_1): Likewise.
13807 (struct lto_stats_d): Add num_tree_bodies_output and
13808 num_pickle_refs_output.
13809 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13810 (lto_read_tree_1): Split out from ...
13811 (lto_read_tree): ... this.
13812 (lto_input_scc): New function.
13813 (lto_input_tree_1): Split out from ...
13814 (lto_input_tree): ... this. Handle LTO_tree_scc.
13815 (lto_data_in_create): Create the streamer cache without hashes.
13816 * lto-streamer-out.c (create_output_block): Create the streamer
13817 cache with hashes when not doing WPA.
13818 (lto_write_tree_1): Split out from ...
13819 (lto_write_tree): ... this.
13820 (get_symbol_initial_value): New function.
13821 (lto_output_tree_1): Split out from ...
13822 (lto_output_tree): ... this. Write trees as series of SCCs
13823 using a DFS walk via DFS_write_tree.
13824 (struct sccs, struct scc_entry): New types.
13825 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13826 (DFS_write_tree_body): New function.
13827 (DFS_write_tree): Likewise.
13828 (hash_tree): Likewise.
13829 (scc_entry_compare): Likewise.
13830 (hash_scc): Likewise.
13831 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13832 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13833 TREE_CHAIN as regular reference.
13834 (streamer_read_integer_cst): Remove.
13835 (streamer_get_pickled_tree): Adjust.
13836 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13837 of DECL_EXTERNALs in BLOCK_VARS for now.
13838 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13839 reference.
13840 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13841 Add hash value argument and record that if hashes are recorded
13842 in the cache.
13843 (streamer_tree_cache_insert_1): Adjust.
13844 (streamer_tree_cache_insert): Likewise.
13845 (streamer_tree_cache_insert_at): Rename to ...
13846 (streamer_tree_cache_replace_tree): ... this and adjust.
13847 (streamer_tree_cache_append): Adjust.
13848 (record_common_node): Likewise.
13849 (streamer_tree_cache_create): Add argument whether to
13850 record hash values together with trees.
13851 (streamer_tree_cache_delete): Adjust.
13852 * tree-streamer.h (struct streamer_tree_cache_d): Add
13853 vector of hashes.
13854 (streamer_read_integer_cst): Remove.
13855 (streamer_tree_cache_insert): Adjust.
13856 (streamer_tree_cache_append): Likewise.
13857 (streamer_tree_cache_insert_at): Rename to ...
13858 (streamer_tree_cache_replace_tree): ... this and adjust.
13859 (streamer_tree_cache_create): Add argument whether to record hashes.
13860 (streamer_tree_cache_get): Rename to ...
13861 (streamer_tree_cache_get_tree): ... this.
13862 (streamer_tree_cache_get_hash): New function.
13863 * tree.c (cache_integer_cst): New function.
13864 * tree.h (cache_integer_cst): Declare.
13865 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13866 * lto-symtab.c (lto_varpool_replace_node): Only release
13867 DECL_INITIAL of non-prevailing decls.
13868 * varpool.c (varpool_remove_initializer): Do not release
13869 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13870
0de86a92
JU		 138712013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13872
13873 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13874 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13875 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13876 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13877 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13878 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13879 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13880 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13881 instead of TARGET_64BIT.
13882 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13883 Require ISA_HAS_<D>DIV.
13884
74240413
RS		 138852013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13886
13887 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13888 (mips*-*-linux*): Move default with_llsc setting to where other
13889 defaults are set.
13890 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13891 with_arch block.
13892 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13893 Likewise. Remove default with_tune setting. Move default float
13894 setting to its own block. Handle with_llsc in the same block as above.
13895
12211b99 138962013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 13897
13898 PR rtl-optimization/57425
13899 PR rtl-optimization/57569
299a5f6a 13900 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 13901 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13902 Changed all callers.
13903 (canon_anti_dependence): New function.
13904 * cse.c (check_dependence): Use canon_anti_dependence.
13905 * cselib.c (cselib_invalidate_mem): Likewise.
13906 * rtl.h (canon_anti_dependence): Declare.
13907
b259d352
JU		 139082013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13909
13910 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13911 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13912 ".set mips3" for 64-bit targets.
13913
4de09b85
DC		 139142013-06-15 Dehao Chen <dehao@google.com>
13915
13916 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13917 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13918 (gimple_check_call_args): Likewise.
13919 * value-prof.c (check_ic_target): Likewise.
13920 * ipa-inline.c (early_inliner): Likewise.
13921 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13922 * cgraph.c (cgraph_create_edge_1): Likewise.
13923 (cgraph_make_edge_direct): Likewise.
13924
d1aa4a76
MM		 139252013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
13926
13927 PR target/57615
13928 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
13929 rs6000_output_move_128bit to handle emitting quad memory
13930 operations. Set attribute length to 8 bytes.
13931
12211b99 139322013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 13933
13934 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
13935 New pattern.
13936 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
13937 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
13938 (aarch64_<su>mlsl<mode>): Likewise.
13939
f5273574
MS		 139402013-06-14 Mike Stump <mikestump@comcast.net>
13941
13942 * Makefile.in (TARGET_H): Add insn-codes.h.
13943
ef7e7ab2
AM		 139442013-06-14 Alan Modra <amodra@gmail.com>
13945
13946 PR middle-end/57134
13947 PR middle-end/57586
13948 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
13949 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
13950 bitfield expansion when EXPAND_MEMORY.
13951 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
13952
135faab6
MM		 139532013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
13954
13955 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
13956 test for clearing quad memory on 32-bit later.
13957
948a1fd9
MG		 139582013-06-13 Marc Glisse <marc.glisse@inria.fr>
13959
13960 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
13961 (fold_negate_expr): Likewise.
13962 (fold_real_zero_addition_p): Handle vectors.
13963 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
13964
0597fb9c
AM		 139652013-06-14 Alan Modra <amodra@gmail.com>
13966
13967 * varasm.c (force_const_mem): Revert 2013-06-07 change.
13968
e5b962d0
JH		 139692013-06-13 Jan Hubicka <jh@suse.cz>
13970
13971 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13972 Local comdats are not externally visible.
13973 * symtab.c (dump_symtab_base): Dump externally visible.
13974 (verify_symtab_base): Verify back links in the symtab hash.
13975
a8355e51
BC		 139762013-06-13 Bin Cheng <bin.cheng@arm.com>
13977
13978 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
13979 CONVERT_EXPR as equal nodes.
13980
8f7e6e33
BC		 139812013-06-13 Bin Cheng <bin.cheng@arm.com>
13982
22fc7d3d 13983 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 13984
5d418483
MG		 139852013-06-13 Marc Glisse <marc.glisse@inria.fr>
13986
13987 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
13988 Generalize to complex and vector.
13989 * tree.c (build_all_ones_cst): New function.
13990 * tree.h (build_all_ones_cst): Declare it.
13991
50751417
AM		 139922013-06-13 Alan Modra <amodra@gmail.com>
13993
13994 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
13995 * config/rs6000/rs6000.md (signbittf2): New insn.
13996 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
13997 (abstf2_internal, cmptf_internal2): Likewise.
13998 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
13999
4b02c962 140002013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14001 Pat Haugen <pthaugen@us.ibm.com>
14002 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 14003
14004 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14005 power8 byte, half-word, and quad-word atomic instructions.
14006 (emit_store_conditional): Likewise.
14007 (rs6000_expand_atomic_compare_and_swap): Likewise.
14008 (rs6000_expand_atomic_op): Likewise.
14009
14010 * config/rs6000/sync.md (larx): Add new modes for power8.
14011 (stcx): Likewise.
14012 (AINT): New mode iterator to include TImode as well as normal
14013 integer modes on power8.
14014 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14015 that VSX registers are not considered. Use AINT mode iterator
14016 instead of INT1 to allow inclusion of quad word atomic operations
14017 on power8.
14018 (load_locked<mode>): Likewise.
14019 (store_conditional<mode>): Likewise.
14020 (atomic_compare_and_swap<mode>): Likewise.
14021 (atomic_exchange<mode>): Likewise.
14022 (atomic_nand<mode>): Likewise.
14023 (atomic_fetch_<fetchop_name><mode>): Likewise.
14024 (atomic_nand_fetch<mode>): Likewise.
14025 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14026 each type.
14027 (ATOMIC): On power8, add QImode, HImode modes.
14028 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14029 modes that promote to SImode.
14030 (load_lockedti): Convert TImode arguments to PTImode, so that we
14031 get a guaranteed even/odd register pair.
14032 (load_lockedpti): Likewise.
14033 (store_conditionalti): Likewise.
14034 (store_conditionalpti): Likewise.
14035
14036 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14037 atomic load/store instructions.
14038 (HSI): Likewise.
14039
710b6cc1
RS		 140402013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14041
14042 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14043 loads.
14044 (insn_count): New attribute, with most cases extracted from...
14045 (length): ...here. Redefine most cases in terms of insn_count.
14046 (single_insn): Delete.
14047 (can_delay): Use insn_count to check for single instructions.
14048 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14049 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14050 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14051 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14052 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14053 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14054 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14055 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14056 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14057 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14058 rather than "length".
14059 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14060 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14061 Use "insn_count" rather than "length".
14062 * config/mips/mips-dsp.md
14063 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14064 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14065 length attributes.
14066
1951f101
MG		 140672013-06-12 Marc Glisse <marc.glisse@inria.fr>
14068
14069 PR tree-optimization/57361
14070 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14071
8b033a8a
SN		 140722013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14073
22fc7d3d
UB		 14074 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14075 to split.
8b033a8a
SN		 14076 (aarch64_simd_combine<mode>): New instruction expansion.
14077 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14078 function prototype.
14079 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14080 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14081
ca0f62a8
JH		 140822013-06-12 Jan Hubicka <jh@suse.cz>
14083
14084 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14085 decl has when in streaming stage.
14086 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14087 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14088
12211b99 140892013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 14090
14091 PR target/57578
14092 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14093
bfdeda2c
JJ		 140942013-06-12 Jakub Jelinek <jakub@redhat.com>
14095
14096 PR tree-optimization/57537
14097 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14098 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14099
a4fa02d1
RB		 141002013-06-12 Richard Biener <rguenther@suse.de>
14101
14102 * data-streamer.h (streamer_write_char_stream): CSE
14103 obs->current_pointer.
14104 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14105 streamer_write_char_stream manually and optimize the resulting loop.
14106 (streamer_write_hwi_stream): Likewise.
14107
bbf9ad07
JH		 141082013-06-12 Jan Hubicka <jh@suse.cz>
14109
14110 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14111 * cgraph.h (varpool_create_empty_node): Declare.
14112 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14113 duplicated nodes.
14114 * symtab.c (symtab_unregister_node): Be lax about missin entries
14115 in node hash.
14116 (symtab_get_node): Update comment.
14117 * varpool.c (varpool_create_empty_node): Break out from ...
14118 (varpool_node_for_decl): ... here.
14119 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14120
a53f90ad
EB		 141212013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14122
14123 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14124 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14125 part. Use straight-line flow at the end.
14126 <COMPONENT_REF>: Remove superfluous else.
14127 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14128
946f9306
JJ		 141292013-06-12 Jakub Jelinek <jakub@redhat.com>
14130
14131 PR target/56564
14132 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14133 target hook even for !TREE_PUBLIC decls. If no resolution info
14134 is available, return false for common and external decls.
14135
a7fca6f0
KP		 141362013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14137
14138 * config/rl78/constraints.md (U): New constraint.
14139 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14140 valloc attribute.
14141
c7d8f446
MM		 141422013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14143
14144 PR target/57589
14145 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14146 to allow returning address to AT_PLATFORM name.
14147
08346abd
JH		 141482013-06-11 Jan Hubicka <jh@suse.cz>
14149
14150 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14151 * cgraph.h (symtab_node_base): Add weakref flag.
14152 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14153 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14154 (output_weakrefs): Use weakref flag.
14155 * fold-const.c (simple_operand_p): Handle WEAK.
14156 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14157 * ipa.c (varpool_externally_visible_p): Drop weakref.
14158 (function_and_variable_visibility): Update comment; fix weakref
14159 sanity checks; do not clear DECL_WEAK on them.
14160 * lto-cgraph.c (lto_output_node): update.
14161 (lto_output_varpool_node): Update.
14162 (input_overwrite_node): Update.
14163 (input_node): Update.
14164 (input_varpool_node): Update.
14165 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14166 (lto_symtab_merge_symbols): Add sanity check.
14167 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14168 * passes.c (rest_of_decl_compilation): Set static flag, too.
14169 * symtab.c (dump_symtab_base): Dump weakref.
14170 (verify_symtab_base): Sanity check weakrefs.
14171 (symtab_make_decl_local): Remove duplicated code.
14172 (symtab_alias_ultimate_target): Simplify.
14173 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 14174
2e6a2f88
TV		 141752013-06-11 Tom de Vries <tom@codesourcery.com>
14176
14177 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14178 sequence_vect == NULL.
14179
d9af3652
DD		 141802013-06-11 DJ Delorie <dj@redhat.com>
14181
14182 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14183 (rl78_unwind_word_mode): New.
14184
ffa66012
DM		 141852013-06-11 David Malcolm <dmalcolm@redhat.com>
14186
14187 * final.c (debug_prefix_maps): Make static.
14188
783f2d35
DM		 141892013-06-11 David Malcolm <dmalcolm@redhat.com>
14190
14191 * function.c (initial_trampoline): Remove stray copy.
14192
7af26def
SN		 141932013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14194
14195 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14196
97756c0e
MJ		 141972013-06-11 Martin Jambor <mjambor@suse.cz>
14198
14199 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14200 within bounds at the beginning of the function.
14201
2c7b8bf6
AM		 142022013-06-11 Alan Modra <amodra@gmail.com>
14203
14204 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14205 reporting.
14206 (get_named_section): Don't NULL !DECL_P decl.
14207
77cea46e
IZ		 142082013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14209
14210 * doc/invoke.texi (core-avx2): Document.
14211 (slm): Likewise.
14212 (atom): Updated with MOVBE.
14213
888f0920
RB		 142142013-06-11 Richard Biener <rguenther@suse.de>
14215
22fc7d3d 14216 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14217
6b438d58
AB		 142182013-06-11 Anton Blanchard <anton@samba.org>
14219
14220 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14221 correct shift value in little-endian mode.
14222
3bc66938
JJ		 142232013-06-11 Jakub Jelinek <jakub@redhat.com>
14224
14225 PR target/56564
14226 * varasm.c (get_variable_align): Move #endif to the right place.
14227
a5dc7f88
CC		 142282013-06-10 Cary Coutant <ccoutant@google.com>
14229
14230 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14231 for hash so that hash table traversal order is deterministic.
14232
19be72ab 142332013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14234 Pat Haugen <pthaugen@us.ibm.com>
14235 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 14236
14237 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14238 of vectors in GPRS if they are direct moves or quad word load or
14239 store moves.
14240
14241 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14242 declaration.
14243 (direct_move_p): Likewise.
14244 (quad_load_store_p): Likewise.
14245
14246 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14247 classes into bins based on the physical register type.
14248 (reg_class_to_reg_type): Likewise.
14249 (IS_STD_REG_TYPE): Likewise.
14250 (IS_FP_VECT_REG_TYPE): Likewise.
14251 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14252 use direct move instructions.
14253 (reload_gpr_vsx): Likewise.
14254 (reload_vsx_gpr): Likewise.
14255 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14256 information that is a simplification of register classes. Also
14257 precalculate direct move reload helpers.
14258 (direct_move_p): New function to return true if the operation can
14259 be done as a direct move instruciton.
14260 (quad_load_store_p): New function to return true if the operation
14261 is a quad memory operation.
14262 (rs6000_legitimize_address): If quad memory, only allow register
14263 indirect for TImode addresses.
14264 (rs6000_legitimate_address_p): Likewise.
14265 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14266 (rs6000_reload_register_type): Likewise.
14267 (register_to_reg_type): Return register type.
14268 (rs6000_secondary_reload_simple_move): New helper function for
14269 secondary reload and secondary memory needed to identify anything
14270 that is a simple move, and does not need reloading.
14271 (rs6000_secondary_reload_direct_move): New helper function for
14272 secondary reload to identify cases that can be done with several
14273 instructions via the direct move instructions.
14274 (rs6000_secondary_reload_move): New helper function for secondary
14275 reload to identify moves between register types that can be done.
14276 (rs6000_secondary_reload): Add support for quad memory operations
14277 and for direct move.
14278 (rs6000_secondary_memory_needed): Likewise.
14279 (rs6000_debug_secondary_memory_needed): Change argument names.
14280 (rs6000_output_move_128bit): New function to return the move to
14281 use for 128-bit moves, including knowing about the various
14282 limitations of quad memory operations.
14283
14284 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14285 memory operations. call rs6000_output_move_128bit for the actual
14286 instruciton(s) to generate.
14287 (vsx_movti_64bit): Likewise.
14288
14289 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14290 (UNSPEC_P8V_MTVSRWZ): Likewise.
14291 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14292 (UNSPEC_P8V_MTVSRD): Likewise.
14293 (UNSPEC_P8V_XXPERMDI): Likewise.
14294 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14295 (UNSPEC_FUSION_GPR): Likewise.
14296 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14297 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 14298 (f32_sv): Likewise.
14299 (f32_dm): Likewise.
14300 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14301 loads and direct move instructions.
14302 (zero_extendsidi2_lfiwzx): Likewise.
14303 (extendsidi2_lfiwax): Likewise.
14304 (extendsidi2_nocell): Likewise.
14305 (floatsi<mode>2_lfiwax): Likewise.
14306 (lfiwax): Likewise.
14307 (floatunssi<mode>2_lfiwzx): Likewise.
14308 (lfiwzx): Likewise.
14309 (fix_trunc<mode>_stfiwx): Likewise.
14310 (fixuns_trunc<mode>_stfiwx): Likewise.
14311 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14312 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14313 (parity<mode>2_cmpb): Set length/type attr.
14314 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14315 for 'mr.' to fast_compare.
14316 (bpermd_<mode>): Change type attr to popcnt.
14317 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14318 (p8_mtvsrwz_1): Likewise.
14319 (p8_mtvsrwz_2): Likewise.
14320 (reload_fpr_from_gpr<mode>): Likewise.
14321 (p8_mtvsrd_1): Likewise.
14322 (p8_mtvsrd_2): Likewise.
14323 (p8_xxpermdi_<mode>): Likewise.
14324 (reload_vsx_from_gpr<mode>): Likewise.
14325 (reload_vsx_from_gprsf): Likewise.
14326 (p8_mfvsrd_3_<mode>): LIkewise.
14327 (reload_gpr_from_vsx<mode>): Likewise.
14328 (reload_gpr_from_vsxsf): Likewise.
14329 (p8_mfvsrd_4_disf): Likewise.
14330 (multi-word GPR splits): Do not split direct moves or quad memory
14331 operations.
14332
16876bdc
DM		 143332013-06-10 David Malcolm <dmalcolm@redhat.com>
14334
14335 * tree-into-ssa.c (interesting_blocks): Make static.
14336
df8a1d28
JJ		 143372013-06-10 Jakub Jelinek <jakub@redhat.com>
14338
14339 PR target/56564
14340 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14341 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14342 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14343 (get_variable_align): New function.
14344 (get_variable_section, emit_bss, emit_common,
14345 assemble_variable_contents, place_block_symbol): Use
14346 get_variable_align instead of DECL_ALIGN.
14347 (assemble_noswitch_variable): Add align argument, use it
14348 instead of DECL_ALIGN.
14349 (assemble_variable): Adjust caller. Use get_variable_align
14350 instead of DECL_ALIGN.
14351 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14352 caller.
14353 (DATA_ABI_ALIGNMENT): Define.
14354 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14355 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14356 opt is false, only return the psABI mandated alignment increase.
14357 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14358 (DATA_ABI_ALIGNMENT): ... this.
14359 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14360 (DATA_ABI_ALIGNMENT): ... this.
14361 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14362 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14363 (DATA_ABI_ALIGNMENT): ... this.
14364 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14365 * doc/tm.texi: Regenerated.
14366
57b29ca6
UB		 143672013-06-10 Uros Bizjak <ubizjak@gmail.com>
14368
14369 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14370 cmp_code to construct REG_EQUAL note.
14371
3e56ed50
JJ		 143722013-06-09 Jakub Jelinek <jakub@redhat.com>
14373
14374 PR target/57568
14375 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14376 that operands[2] doesn't overlap with operands[0].
14377
74fe2a1d
DE		 143782013-06-09 David Edelsohn <dje.gcc@gmail.com>
14379 Jan Hubicka <jh@suse.cz>
14380
14381 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14382 hack to mark symbols as used.
14383
ed52a84e
VM		 143842013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14385
14386 PR rtl-optimization/57559
14387 * lra-constraints.c (process_alt_operands): Don't discourage
14388 memory with known offset for offsetable memory constraint.
14389 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14390
299316ed
EB		 143912013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14392
14393 * varasm.c (struct oc_local_state): Reorder fields.
14394 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14395 and adjust accordingly.
14396 (output_constructor): Reorder initialization code and adjust call to
14397 output_constructor_bitfield.
14398
88ac60d3
JH		 143992013-06-07 Jan Hubicka <jh@suse.cz>
14400
14401 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14402
4cdce1a8
DM		 144032013-06-07 David Malcolm <dmalcolm@redhat.com>
14404
14405 * tree-object-size.c (unknown): Make const.
14406
b5e0425c
AK		 144072013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14408
14409 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14410 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14411 for last alternative in the cpu_facility attribute.
14412
79678d04
KT		 144132013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14414
14415 PR target/56315
14416 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14417 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14418 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14419 * config/arm/constraints.md (Dg): New constraint.
14420 * config/arm/neon.md (xordi3_neon): Remove.
14421 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14422 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14423
b31ddbdb
KT		 144242013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14425
14426 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14427 Clean up alternatives.
14428
8f90b7d4
AM		 144292013-06-07 Alan Modra <amodra@gmail.com>
14430
14431 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14432 va_list_gpr_size.
14433
961ce119
AM		 144342013-06-07 Alan Modra <amodra@gmail.com>
14435
14436 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14437
33e49835
KT		 144382013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14439
14440 * config/arm/constraints.md (Df): New constraint.
14441 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14442 Correct length attribute for last two alternatives.
14443
acd17ae6
AM		 144442013-06-07 Alan Modra <amodra@gmail.com>
14445
14446 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14447 override user -mfp-in-toc.
14448 (offsettable_ok_by_alignment): Consider just the current access
14449 rather than the whole object, unless BLKmode. Handle
14450 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14451 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14452 for -mcmodel=medium.
14453 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14454 override user -mfp-in-toc or -msum-in-toc. Default to
14455 -mno-fp-in-toc for -mcmodel=medium.
14456
73310b0e
DD		 144572013-06-06 DJ Delorie <dj@redhat.com>
14458
14459 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14460 TARGET_VALID_POINTER_MODE.
14461
0bd62dca 144622013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14463 Pat Haugen <pthaugen@us.ibm.com>
14464 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 14465
14466 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14467 Document new power8 builtins.
14468
14469 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14470 condition code register, to allow 128-bit logical operations to be
14471 done in the VSX or GPR registers.
14472 (nor<mode>3): Use the canonical form for nor.
14473 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14474 vclz*, and vpopcnt* vector instructions.
14475 (nand<mode>3): Likewise.
14476 (orc<mode>3): Likewise.
14477 (clz<mode>2): LIkewise.
14478 (popcount<mode>2): Likewise.
14479
14480 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14481 that only the GPRs are recognized.
14482
14483 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14484 support for new power8 builtins.
14485
14486 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14487 builtin functions.
14488 (xscvdpspn): Likewise.
14489 (vclz): Likewise.
14490 (vclzb): Likewise.
14491 (vclzh): Likewise.
14492 (vclzw): Likewise.
14493 (vclzd): Likewise.
14494 (vpopcnt): Likewise.
14495 (vpopcntb): Likewise.
14496 (vpopcnth): Likewise.
14497 (vpopcntw): Likewise.
14498 (vpopcntd): Likewise.
14499 (vgbbd): Likewise.
14500 (vmrgew): Likewise.
14501 (vmrgow): Likewise.
14502 (eqv): Likewise.
14503 (eqv_v16qi3): Likewise.
14504 (eqv_v8hi3): Likewise.
14505 (eqv_v4si3): Likewise.
14506 (eqv_v2di3): Likewise.
14507 (eqv_v4sf3): Likewise.
14508 (eqv_v2df3): Likewise.
14509 (nand): Likewise.
14510 (nand_v16qi3): Likewise.
14511 (nand_v8hi3): Likewise.
14512 (nand_v4si3): Likewise.
14513 (nand_v2di3): Likewise.
14514 (nand_v4sf3): Likewise.
14515 (nand_v2df3): Likewise.
14516 (orc): Likewise.
14517 (orc_v16qi3): Likewise.
14518 (orc_v8hi3): Likewise.
14519 (orc_v4si3): Likewise.
14520 (orc_v2di3): Likewise.
14521 (orc_v4sf3): Likewise.
14522 (orc_v2df3): Likewise.
14523
14524 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14525 allow power8 quad mode in 64-bit.
14526 (rs6000_builtin_vectorized_function): Add support to vectorize
14527 ISA 2.07 count leading zeros, population count builtins.
14528 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14529 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14530 (builtin_function_type): Add vgbbd builtin function which takes an
14531 unsigned argument.
14532 (altivec_expand_vec_perm_const): Add support for new power8 merge
14533 instructions.
14534
14535 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14536 that does not include TImdoe for use with 32-bit.
14537 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14538 instructions.
14539 (UNSPEC_VSX_CVDPSPN): Likewise.
14540 (vsx_xscvdpspn): Likewise.
14541 (vsx_xscvspdpn): Likewise.
14542 (vsx_xscvdpspn_scalar): Likewise.
14543 (vsx_xscvspdpn_directmove): Likewise.
14544 (vsx_and<mode>3): Split logical operations into 32-bit and
14545 64-bit. Add support to do logical operations on TImode as well as
14546 VSX vector types. Allow logical operations to be done in either
14547 VSX registers or in general purpose registers in 64-bit mode. Add
14548 splitters if GPRs were used. For AND, add clobber of CCmode to
14549 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14550 encoding.
14551 (vsx_and<mode>3_32bit): Likewise.
14552 (vsx_and<mode>3_64bit): Likewise.
14553 (vsx_ior<mode>3): Likewise.
14554 (vsx_ior<mode>3_32bit): Likewise.
14555 (vsx_ior<mode>3_64bit): Likewise.
14556 (vsx_xor<mode>3): Likewise.
14557 (vsx_xor<mode>3_32bit): Likewise.
14558 (vsx_xor<mode>3_64bit): Likewise.
14559 (vsx_one_cmpl<mode>2): Likewise.
14560 (vsx_one_cmpl<mode>2_32bit): Likewise.
14561 (vsx_one_cmpl<mode>2_64bit): Likewise.
14562 (vsx_nor<mode>3): Likewise.
14563 (vsx_nor<mode>3_32bit): Likewise.
14564 (vsx_nor<mode>3_64bit): Likewise.
14565 (vsx_andc<mode>3): Likewise.
14566 (vsx_andc<mode>3_32bit): Likewise.
14567 (vsx_andc<mode>3_64bit): Likewise.
14568 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14569 and xxlorc instructions.
14570 (vsx_eqv<mode>3_64bit): Likewise.
14571 (vsx_nand<mode>3_32bit): Likewise.
14572 (vsx_nand<mode>3_64bit): Likewise.
14573 (vsx_orc<mode>3_32bit): Likewise.
14574 (vsx_orc<mode>3_64bit): Likewise.
14575
14576 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14577
14578 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14579 instruction.
14580 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14581 (p8_vmrgow): Likewise.
14582 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14583 GPRs to be split under VSX.
14584 (p8v_clz<mode>2): Add power8 count leading zero support.
14585 (p8v_popcount<mode>2): Add power8 population count support.
14586 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14587 support.
14588
14589 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14590 instruction.
14591
14592 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14593 builtin functions.
14594 (vec_nand): Likewise.
14595 (vec_vclz): Likewise.
14596 (vec_vclzb): Likewise.
14597 (vec_vclzd): Likewise.
14598 (vec_vclzh): Likewise.
14599 (vec_vclzw): Likewise.
14600 (vec_vgbbd): Likewise.
14601 (vec_vmrgew): Likewise.
14602 (vec_vmrgow): Likewise.
14603 (vec_vpopcnt): Likewise.
14604 (vec_vpopcntb): Likewise.
14605 (vec_vpopcntd): Likewise.
14606 (vec_vpopcnth): Likewise.
14607 (vec_vpopcntw): Likewise.
14608
37684c46
VM		 146092013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14610
14611 PR rtl-optimization/57468
14612 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14613 spilled pseudos.
14614
d89ae27c
VM		 146152013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14616
14617 PR rtl-optimization/57459
14618 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14619 type when setting live regs.
14620
3597e113
VM		 146212013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14622
14623 * config/s390/s390.opt (mlra): New option.
14624 * config/s390/s390.c (s390_decompose_address): Check displacement
14625 for all registers for LRA.
14626 (s390_secondary_reload): Don't used secondary reloads for LRA.
14627 (s390_lra_p): New function.
14628 (TARGET_LRA_P): Define.
14629 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14630 of attribute cpu_facility to zarch for the last alternative.
14631 (*cmpmem_short): Ditto.
14632
01037aeb
EB		 146332013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14634
14635 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14636 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14637 (arm_expand_prologue): Likewise.
14638
3371a64f
TJ		 146392013-06-06 Teresa Johnson <tejohnson@google.com>
14640
14641 PR c++/53743
14642 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14643 as this is now done by redirect_edge_and_branch_force.
14644 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14645 barriers, and fix interaction with splitting.
14646 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14647 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14648 reflect changes made in the routine.
14649 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14650 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14651 since this is called in cfglayout mode, and replace partition fixup
14652 with assert as that is now done by force_nonfallthru_and_redirect.
14653 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14654 already be marked with region crossing note.
14655 (insert_section_boundary_note): Make non-static, gate on flag
14656 has_bb_partition, rewrite to also check for multiple partitions.
14657 (rest_of_handle_reorder_blocks): Remove call to
14658 insert_section_boundary_note, now done later during free_cfg.
14659 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14660 * bb-reorder.h (insert_section_boundary_note): Declare.
14661 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14662 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14663 invoke insert_section_boundary_note.
14664 (try_redirect_by_replacing_jump): Remove unnecessary
14665 check for region crossing note.
14666 (fixup_partition_crossing): New function.
14667 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14668 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14669 in non-cfglayout mode.
14670 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14671 remove old code that tried to do this. Emit barrier correctly
14672 when we are in cfglayout mode.
14673 (last_bb_in_partition): New function.
14674 (rtl_split_edge): Correctly fixup partition boundaries.
14675 (commit_one_edge_insertion): Remove old code that tried to
14676 fixup region crossing edge since this is now handled in
14677 split_block, and set up insertion point correctly since
14678 block may now end in a jump.
14679 (verify_hot_cold_block_grouping): Guard against checking when not in
14680 linearized RTL mode.
14681 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14682 notes.
14683 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14684 rtl_verify_flow_info, so not called in cfglayout mode.
14685 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14686 (fixup_reorder_chain): Remove old code that attempted to fixup region
14687 crossing note as this is now handled in force_nonfallthru_and_redirect.
14688 (duplicate_insn_chain): Don't duplicate switch section notes.
14689 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14690 note.
14691 * basic-block.h (emit_barrier_after_bb): Declare.
14692
66071e10
KT		 146932013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14694
14695 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14696 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14697 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14698
0e26bf3d
KT		 146992013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14700
14701 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14702 where appropriate.
14703 * config/arm/ldmstm.md: Regenerate.
14704
12b4e7ef
KT		 147052013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14706
14707 * config/arm/sync.md (atomic_loaddi_1):
14708 Disable predication for arm_restrict_it.
14709 (arm_load_exclusive<mode>): Likewise.
14710 (arm_load_exclusivesi): Likewise.
14711 (arm_load_exclusivedi): Likewise.
14712 (arm_load_acquire_exclusive<mode>): Likewise.
14713 (arm_load_acquire_exclusivesi): Likewise.
14714 (arm_load_acquire_exclusivedi): Likewise.
14715 (arm_store_exclusive<mode>): Likewise.
14716 (arm_store_exclusive<mode>): Likewise.
14717 (arm_store_release_exclusivedi): Likewise.
14718 (arm_store_release_exclusive<mode>): Likewise.
14719
15d16c8a
RB		 147202013-06-06 Richard Biener <rguenther@suse.de>
14721
14722 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14723 after LTO_null.
14724 (lto_tag_is_tree_code_p): Adjust.
14725 (lto_tag_is_gimple_code_p): Likewise.
14726 (lto_gimple_code_to_tag): Likewise.
14727 (lto_tag_to_gimple_code): Likewise.
14728 (lto_tree_code_to_tag): Likewise.
14729 (lto_tag_to_tree_code): Likewise.
14730 * data-streamer.h (streamer_write_hwi_in_range): Use
14731 uhwi streaming to stream the normalized range.
14732 (streamer_read_hwi_in_range): Likewise.
14733
17a7fc37
KT		 147342013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14735
14736 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14737 (predicable_short_it): Likewise.
14738 (predicated): Likewise.
14739 (enabled): Handle above.
14740 (define_cond_exec): Set predicated attribute to yes.
14741
b57ca59b
MS		 147422013-06-05 Mike Stump <mikestump@comcast.net>
14743
14744 * gdbinit.in (__FUNCTION__): Add.
14745
c1e183a9
UB		 147462013-06-05 Uros Bizjak <ubizjak@gmail.com>
14747
14748 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14749 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14750
0378bddb
JH		 147512013-06-05 Jan Hubicka <jh@suse.cz>
14752
14753 * varasm.c (mark_decl_referenced): Revert the removal until targets
14754 are fixed.
14755
8f8a2057
DE		 147562013-06-05 David Edelsohn <dje.gcc@gmail.com>
14757
14758 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14759 instead of mark_decl_referenced.
14760
edb983b2
JH		 147612013-06-05 Jan Hubicka <jh@suse.cz>
14762
14763 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14764 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14765 and symtab_used_from_object_file_p.
14766 (cgraph_make_node_local_1): Clear forced_by_abi.
14767 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14768 * cgraph.h (symtab_node_base): Add forced_by_abi.
14769 (decide_is_variable_needed): Remove.
14770 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14771 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14772 (decide_is_symbol_needed): ... this one; handle symbols in general;
14773 always analyze virtuals; honnor forced_by_abi.
14774 (cgraph_finalize_function): Update.
14775 (varpool_finalize_decl): Update.
14776 (symbol_defined_and_needed): Remove.
14777 (analyze_functions): Update.
14778 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14779 output_refs, input_overwrite_node): Handle forced_by_abi.
14780 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14781 (address_taken_from_non_vtable_p): ... this one.
14782 (comdat_can_be_unshared_p_1): New function.
14783 (cgraph_comdat_can_be_unshared_p): Rename to ...
14784 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14785 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14786 (function_and_variable_visibility): Clear forced_by_abi as needed.
14787 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14788 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 14789 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14790 * varpool.c (decide_is_variable_needed): Remove.
14791
9912dbe5
KT		 147922013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14793
14794 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14795 (arm_option_override): Override arm_restrict_it where appropriate.
14796 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14797 * config/arm/arm.opt (mrestrict-it): New command-line option.
14798 * doc/invoke.texi: Document -mrestrict-it.
14799
34c136b6
DM		 148002013-06-05 David Malcolm <dmalcolm@redhat.com>
14801
14802 * tsan.c (tsan_atomic_table): Make const.
14803
a367df53
RB		 148042013-06-05 Richard Biener <rguenther@suse.de>
14805
14806 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14807 index associated with the tree we are supposed to replace.
14808 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14809 TREE_ASM_WRITTEN as zero for everything but SSA names.
14810
70d51a19
DM		 148112013-06-05 David Malcolm <dmalcolm@redhat.com>
14812
14813 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14814
b9bd2075
JH		 148152013-06-04 Jan Hubicka <jh@suse.cz>
14816
14817 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14818 (input_node, input_varpool_node): Handle correctly external same
14819 body aliases.
14820 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14821 nodes at ltrans stage.
14822
aaae719d
JH		 148232013-06-04 Jan Hubicka <jh@suse.cz>
14824
14825 * ipa-inline.c (update_caller_keys): Fix availability test.
14826 (update_callee_keys): Likewise.
14827 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14828 to follow ELF standard.
14829
107eea2c
JU		 148302013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14831
14832 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14833 (mips64r5900el-*-elf*): New configurations.
14834 * config/mips/mips-cpus.def (r5900): New processor.
14835 * config/mips/mips-tables.opt: Regenerate.
14836 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14837 (mips_issue_rate): Handle PROCESSOR_R5900.
14838 (mips_reorg_process_insns): Force reorder mode for the R5900.
14839 * config/mips/mips.h (TARGET_MIPS5900): Define.
14840 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14841 TARGET_MIPS5900.
14842 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14843 TARGET_MIPS5900.
14844 * config/mips/mips.md (processor): Add r5900.
14845 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14846
b7342d25
IB		 148472013-06-04 Ian Bolton <ian.bolton@arm.com>
14848
14849 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14850 into function to generate MOVI instruction.
c1e183a9 14851 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 14852 (aarch64_preferred_simd_mode): Turn into wrapper.
14853 (aarch64_output_scalar_simd_mov_immediate): New function.
14854 * config/aarch64/aarch64-protos.h: Add prototype for above.
14855
81c2dfb9
IB		 148562013-06-04 Ian Bolton <ian.bolton@arm.com>
14857
14858 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14859 element_char member.
14860 (sizetochar): Return signed char.
14861 (aarch64_simd_valid_immediate): Remove elchar and other
14862 unnecessary variables.
14863 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14864 Calculate element_char as required.
14865 * config/aarch64/aarch64-protos.h: Update and move prototype
14866 for aarch64_output_simd_mov_immediate.
14867 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14868 Update arguments.
14869
48063b9d
IB		 148702013-06-04 Ian Bolton <ian.bolton@arm.com>
14871
14872 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14873 information completed by aarch64_simd_valid_immediate.
14874 (aarch64_legitimate_constant_p): Update arguments.
14875 (aarch64_simd_valid_immediate): Work with struct rather than many
14876 pointers.
14877 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14878 (aarch64_simd_make_constant): Update arguments.
14879 (aarch64_output_simd_mov_immediate): Work with struct rather than
14880 many pointers. Output immediate directly rather than as operand.
14881 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14882 Update prototype.
14883 * config/aarch64/constraints.md (Dn): Update arguments.
14884
3ea63f60
IB		 148852013-06-04 Ian Bolton <ian.bolton@arm.com>
14886
14887 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14888 longer static.
14889 (aarch64_simd_immediate_valid_for_move): Remove.
14890 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14891 (aarch64_simd_make_constant): Update call.
14892 (aarch64_output_simd_mov_immediate): Update call.
14893 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14894 Add prototype.
14895 * config/aarch64/constraints.md (Dn): Update call.
14896
d8edd899
IB		 148972013-06-04 Ian Bolton <ian.bolton@arm.com>
14898
14899 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14900 return type to bool for prototype.
14901 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14902 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14903 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14904
44b20bb8
CM		 149052013-06-04 Catherine Moore <clm@codesourcery.com>
14906
14907 * config/mips/mips.opt (meva): New.
c1e183a9 14908 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 14909 (ASM_SPEC): Handle -meva.
14910 * doc/invoke.texi (meva): Document.
14911
52befbd8
AM		 149122013-06-04 Alan Modra <amodra@gmail.com>
14913
14914 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14915 constant output.
14916
aadaf24e
KT		 149172013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14918
14919 * rtl.def: Add extra fourth optional field to define_cond_exec.
14920 * gensupport.c (process_one_cond_exec): Process attributes from
14921 define_cond_exec.
14922 * doc/md.texi: Document fourth field in define_cond_exec.
14923
0cd9e9ee
EB		 149242013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14925
14926 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14927 out the processing order as in store_bit_field_1.
14928
cb0f665d
JH		 149292013-06-04 Jan Hubicka <jh@suse.cz>
14930
14931 PR middle-end/57500
14932 * cgraphunit.c (cgraph_process_same_body_aliases): Create
14933 non-VAR_DECL node if it does not exist yet.
14934
53984b9b
RS		 149352013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
14936
14937 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
14938 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
14939 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
14940 target_cpu_default setting.
14941
5d318fd4
TJ		 149422013-06-03 Teresa Johnson <tejohnson@google.com>
14943
14944 * dumpfile.c (opt_info_switch_p): Change -fopt-info
14945 default to -fopt-info=optimized instead of all.
14946 * doc/invoke.texi: Ditto.
14947 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
14948 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
14949 (execute_vect_slp): Emit BB vectorization success under
14950 MSG_OPTIMIZED_LOCATIONS.
14951 * tree-vect-slp.c (vect_slp_transform_bb): Change
14952 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
14953 * tree-vect-loop.c (vect_transform_loop): Ditto.
14954
ec3af349
JM		 149552013-06-03 Jason Merrill <jason@redhat.com>
14956
14957 PR c++/57415
14958 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
14959 Use TARGET_EXPR for C++.
14960
1baf9159
JJ		 149612013-06-03 Jakub Jelinek <jakub@redhat.com>
14962
14963 PR rtl-optimization/57268
0cd9e9ee 14964 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 14965 if DEBUG_INSN_P (insn).
14966
14967 Reapply
14968 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14969
14970 PR rtl-optimization/57268
14971 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14972 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14973
0f1d3965
YR		 149742013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14975
14976 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
14977 (ix86_avoid_lea_for_addr): Likewise.
14978 (exact_dependency_1): Likewise.
14979 (ix86_adjust_cost): Likewise.
14980 (swap_top_of_ready_list): Fix formatting and !reload_completed check
14981 removed.
14982 (do_reorder_for_imul): Fix typo, formatting and
14983 !reload_completed check removed.
14984 (ix86_sched_reorder): Fix typo and formatting.
14985 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
14986 list.
14987
3bc30db4
SN		 149882013-06-03 Sofiane Naci <sofiane.naci@arm.com>
14989
14990 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
14991
19cc6fac
EB		 149922013-06-03 Eric Botcazou <ebotcazou@adacore.com>
14993
14994 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
14995 <STRING_CST>: Likewise.
14996 <VECTOR_CST>: Likewise.
14997
c68a6e08
JW		 149982013-06-01 Janus Weil <janus@gcc.gnu.org>
14999 Mikael Morin <mikael@gcc.gnu.org>
15000
15001 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15002 * config.in: Regenerated.
15003 * configure: Regenerated.
15004
38e55e5c
JH		 150052013-06-01 Jan Hubicka <jh@suse.cz>
15006
15007 PR middle-end/57366
15008 * cgraphunit.c (compile): When weakref is not supported,
15009 set up transparent aliases before final output pass.
15010 * varasm.c (assemble_alias): Do not try to do it here.
15011
eb51d2ff
JH		 150122013-06-01 Jan Hubicka <jh@suse.cz>
15013
15014 PR middle-end/57467
15015 * passes.c (for_per_function): Skip unanalyzed functions.
15016
40a7fe1e
JH		 150172013-06-01 Jan Hubicka <jh@suse.cz>
15018
15019 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15020 (lto_symtab_merge_symbols_1): ... this one.
15021 (lto_symtab_merge_cgraph_nodes): Rename to ...
15022 (lto_symtab_merge_symbols): ... this one; simplify.
15023 * cgraph.c (same_body_aliases_done): Rename to ...
15024 (cpp_implicit_aliases_done): ... this one.
15025 (cgraph_create_function_alias): Update.
15026 (cgraph_same_body_alias): Update.
c1e183a9 15027 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 15028 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 15029 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 15030 (cgraph_node): Remove same_body_alias.
15031 (varpool_node): Remove alias_of and extra_name_alias.
15032 (same_body_aliases_done): Rename to ..
15033 (cpp_implicit_aliases_done): ... this one.
15034 (symtab_alias_ultimate_target): Add default parameter.
15035 (symtab_resolve_alias): New function.
15036 (fixup_same_cpp_alias_visibility): Declare.
15037 (cgraph_function_node): Add default parameter.
15038 (cgraph_node_asm_name): Likewise.
15039 (cgraph_function_or_thunk_node): Add default parameter; do
15040 not ICE when it is NULL.
15041 (varpool_variable_node): Likewise.
15042 * tree-emutls.c (create_emultls_var): Update.
15043 (ipa_lower_emutls): Update.
15044 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15045 (cgraph_reset_node): Reset alias info.
15046 (cgraph_finalize_function): Update.
15047 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15048 (analyze_function): Simplify.
15049 (cgraph_process_same_body_aliases): Simplify.
15050 (analyze_functions): Fixup same body aliases.
15051 (handle_alias_pairs): Simplify.
15052 (assemble_thunk): Update.
15053 (assemble_thunks_and_aliases): Update.
15054 (output_weakrefs): Rewrite.
15055 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15056 (lto_output_varpool_node): Likewise.
15057 (compute_ltrans_boundary): Remve assert.
15058 (get_alias_symbol): New functoin.
15059 (input_node): Rewrite alias handling.
15060 (input_varpool_node): Likewise.
15061 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15062 * ipa.c (process_references): Handle weakrefs correctly.
15063 (symtab_remove_unreachable_nodes): Likewise.
15064 * trans-mem.c (get_cg_data): Update.
15065 (ipa_tm_create_version_alias): Update.
15066 (ipa_tm_execute): Update.
15067 * symtab.c (dump_symtab_base): Dump aliases.
15068 (verify_symtab_base): Verify aliases.
15069 (symtab_node_availability): New function.
15070 (symtab_alias_ultimate_target): Simplify.
15071 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15072 handle all the fixup cases.
15073 (symtab_resolve_alias): New function.
15074 * passes.c (ipa_write_summaries): Handle weakrefs.
15075 * varpool.c (varpool_analyze_node): Simplify.
15076 (assemble_aliases): Update.
15077 (varpool_create_variable_alias): Simplify.
15078 (varpool_extra_name_alias): Simplify.
15079 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15080 (lto_symtab_merge_symbols): ... this one.
15081
21759881
DT		 150822013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15083
15084 Revert
15085 PR rtl-optimization/57268
15086 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15087 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15088
daf4e940
TB		 150892013-06-01 Tobias Burnus <burnus@net-b.de>
15090
15091 Partially reverted:
15092 2013-05-31 Tobias Burnus <burnus@net-b.de>
15093
15094 PR middle-end/57073
15095 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15096 further up.
15097
c359d8d0
DT		 150982013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15099
15100 PR rtl-optimization/57268
15101 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15102 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15103
4172215d
EB		 151042013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15105
15106 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15107 unordered comparison operators when -fno-trapping-math is in effect
15108 on the e500.
15109 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15110 and implement unordered comparison operators properly on the e500.
15111
a8c50132
EB		 151122013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15113
15114 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15115 for constant scalar integers.
15116 (simplify_relational_operation_1): Likewise.
15117
44626634
SB		 151182013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15119
15120 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15121 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15122 Fix comment.
15123
55805e54
YR		 151242013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15125 Igor Zamyatin <igor.zamyatin@intel.com>
15126
15127 Silvermont (SLM) architecture performance tuning.
15128 * config/i386/i386.h (enum ix86_tune_indices): Add
15129 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15130 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15131
15132 * config/i386/i386.c (initial_ix86_tune_features)
15133 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15134 (ix86_lea_outperforms): Handle Silvermont tuning.
15135 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15136 call.
15137 (ix86_use_lea_for_mov): Likewise.
15138 (ix86_avoid_lea_for_addr): Likewise.
15139 (ix86_lea_for_add_ok): Likewise.
15140 (exact_dependency_1): New function.
15141 (exact_store_load_dependency): Likewise.
15142 (ix86_adjust_cost): Handle Silvermont tuning.
15143 (do_reoder_for_imul): Likewise.
15144 (swap_top_of_ready_list): New function.
15145 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15146
15147 * config/i386/i386.md (peepholes that split memory operand in fp
15148 converts): New.
15149
e19c9de2
MS		 151502013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15151
15152 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15153 Remove un-necessary braces.
15154
38e6c9a6
MS		 151552013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15156
15157 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15158 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15159
e3530904
TB		 151602013-05-31 Tobias Burnus <burnus@net-b.de>
15161
15162 PR middle-end/57073
c1e183a9 15163 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 15164
b6af05a9
KT		 151652013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15166
15167 PR target/56315
15168 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15169 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15170 * config/arm/neon.md (iordi3_neon): Remove.
15171 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15172 * config/arm/predicates.md (imm_for_neon_logic_operand):
15173 Move to earlier in the file.
15174 (neon_logic_op2): Likewise.
15175 (arm_iordi_operand_neon): New predicate.
15176
f800c166
RB		 151772013-05-31 Richard Biener <rguenther@suse.de>
15178
15179 PR tree-optimization/57478
15180 PR tree-optimization/57453
15181 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15182 are life as well.
15183
e571fa59
KP		 151842013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15185
15186 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15187 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15188
0fa6e0ef
TB		 151892013-05-30 Tobias Burnus <burnus@net-b.de>
15190 Thomas Koenig <tkoenig@gcc.gnu.org>
15191
15192 PR middle-end/57073
15193 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15194 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15195
4688ddf5
SB		 151962013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15197
0fa6e0ef 15198 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15199
3b9ceb4b
VM		 152002013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15201
15202 * target.def (register_usage_leveling_p): New hook.
15203 * targhooks.c (default_register_usage_leveling_p): New.
15204 * targhooks.h (default_register_usage_leveling_p): New prototype.
15205 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15206 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15207 * doc/tm.texi: Update.
15208 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15209
65f9e789
IB		 152102013-05-30 Ian Bolton <ian.bolton@arm.com>
15211
15212 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15213 (*insv_reg<mode>): New define_insn.
15214
12211b99 152152013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 15216
15217 PR rtl-optimization/57439
15218 * postreload.c (move2add_valid_value_p): Check that we have
15219 a zero subreg_regno_offset when accessing the register in
15220 the requested mode.
15221
0b871ccf
YR		 152222013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15223 Igor Zamyatin <igor.zamyatin@intel.com>
15224
15225 Silvermont (SLM) architecture pipeline model, tuning and
15226 insn selection.
15227 * config.gcc: Add slm config options and target.
15228
15229 * config/i386/slm.md: New.
15230
15231 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15232
f43245d1 15233 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR		 15234 PROCESSOR_SLM.
15235 (ix86_target_macros_internal): Likewise.
15236
f43245d1 15237 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR		 15238 (m_SLM): New macro flag.
15239 (initial_ix86_tune_features): Set m_SLM.
15240 (x86_accumulate_outgoing_args): Likewise.
15241 (x86_arch_always_fancy_math_387): Likewise.
15242 (processor_target_table): Add slm cost.
15243 (cpu_names): Add slm cpu name.
15244 (x86_option_override_internal): Set SLM ISA.
15245 (ix86_issue_rate): New case PROCESSOR_SLM.
15246 (ia32_multipass_dfa_lookahead): Likewise.
15247 (fold_builtin_cpu): Add slm.
15248
15249 * config/i386/i386.h (TARGET_SLM): New target macro.
15250 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15251 (processor_type): Add PROCESSOR_SLM.
15252
15253 * config/i386/i386.md (cpu): Add new value "slm".
15254 (slm.md): Include slm.md.
15255
24d5b097
XG		 152562013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15257 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15258
15259 * config/arm/arm-protos.h: Add and update function protos.
15260 * config/arm/arm.c (use_simple_return_p): New added.
15261 (thumb2_expand_return): Check simple_return flag.
15262 * config/arm/arm.md: Add simple_return and conditional simple_return.
15263 * config/arm/iterators.md: Add iterator for return and simple_return.
15264
c1cccc15
ZC		 152652013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15266
15267 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15268 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15269 (arm_emit_vfp_multi_reg_pop): Likewise.
15270 (thumb2_emit_ldrd_pop): Likewise.
15271 (arm_expand_epilogue): Add misc REG_CFA notes.
15272 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15273
3a4f280b
LC		 152742013-05-29 Lawrence Crowl <crowl@google.com>
15275
15276 * config/arm/t-arm: Update for below.
15277
15278 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15279 Change type to hash_table. Update dependent calls and types.
15280
15281 * config/i386/t-cygming: Update for below.
15282
15283 * config/i386/t-interix: Update for below.
15284
15285 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15286 Change type to hash_table. Update dependent calls and types.
15287 (i386_find_on_wrapper_list::wrappers): Likewise.
15288
15289 * config/ia64/t-ia64: Update for below.
15290
15291 * config/ia64/ia64.c (bundle_state_table):
15292 Change type to hash_table. Update dependent calls and types.
15293
15294 * config/mips/mips.c (mips_reorg_process_insns::htab):
15295 Change type to hash_table. Update dependent calls and types.
15296
15297 * config/sol2.c (solaris_comdat_htab):
15298 Change type to hash_table. Update dependent calls and types.
15299
15300 * config/t-sol2: Update for above.
15301
1388a0e3
TJ		 153022013-05-29 Teresa Johnson <tejohnson@google.com>
15303
15304 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15305 functions are not yet marked as defined.
15306
a5965b52 153072013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15308 Pat Haugen <pthaugen@us.ibm.com>
15309 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 15310
15311 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15312 instructions.
15313 (VEC_A): Likewise.
15314 (VEC_C): Likewise.
15315 (vrotl<mode>3): Likewise.
15316 (vashl<mode>3): Likewise.
15317 (vlshr<mode>3): Likewise.
15318 (vashr<mode>3): Likewise.
15319
15320 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15321 support for power8 V2DI builtins.
15322
15323 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15324 power8 V2DI builtins.
15325 (vupkhsw): Likewise.
15326 (vupklsw): Likewise.
15327 (vaddudm): Likewise.
15328 (vminsd): Likewise.
15329 (vmaxsd): Likewise.
15330 (vminud): Likewise.
15331 (vmaxud): Likewise.
15332 (vpkudum): Likewise.
15333 (vpksdss): Likewise.
15334 (vpkudus): Likewise.
15335 (vpksdus): Likewise.
15336 (vrld): Likewise.
15337 (vsld): Likewise.
15338 (vsrd): Likewise.
15339 (vsrad): Likewise.
15340 (vsubudm): Likewise.
15341 (vcmpequd): Likewise.
15342 (vcmpgtsd): Likewise.
15343 (vcmpgtud): Likewise.
15344 (vcmpequd_p): Likewise.
15345 (vcmpgtsd_p): Likewise.
15346 (vcmpgtud_p): Likewise.
15347 (vupkhsw): Likewise.
15348 (vupklsw): Likewise.
15349 (vaddudm): Likewise.
15350 (vmaxsd): Likewise.
15351 (vmaxud): Likewise.
15352 (vminsd): Likewise.
15353 (vminud): Likewise.
15354 (vpksdss): Likewise.
15355 (vpksdus): Likewise.
15356 (vpkudum): Likewise.
15357 (vpkudus): Likewise.
15358 (vrld): Likewise.
15359 (vsld): Likewise.
15360 (vsrad): Likewise.
15361 (vsrd): Likewise.
15362 (vsubudm): Likewise.
15363
15364 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15365 support for power8 V2DI instructions.
15366
15367 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15368 power8 V2DI instructions. Combine pack and unpack insns to use an
15369 iterator for each mode. Check whether a particular mode supports
15370 Altivec instructions instead of just checking TARGET_ALTIVEC.
15371 (UNSPEC_VPKUWUM): Likewise.
15372 (UNSPEC_VPKSHSS): Likewise.
15373 (UNSPEC_VPKSWSS): Likewise.
15374 (UNSPEC_VPKUHUS): Likewise.
15375 (UNSPEC_VPKSHUS): Likewise.
15376 (UNSPEC_VPKUWUS): Likewise.
15377 (UNSPEC_VPKSWUS): Likewise.
15378 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15379 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15380 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15381 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15382 (UNSPEC_VUPKHSB): Likewise.
15383 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15384 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15385 (UNSPEC_VUPKHSH): Likewise.
15386 (UNSPEC_VUPKLSB): Likewise.
15387 (UNSPEC_VUPKLSH): Likewise.
15388 (VI2): Likewise.
15389 (VI_char): Likewise.
15390 (VI_scalar): Likewise.
15391 (VI_unit): Likewise.
15392 (VP): Likewise.
15393 (VP_small): Likewise.
15394 (VP_small_lc): Likewise.
15395 (VU_char): Likewise.
15396 (add<mode>3): Likewise.
15397 (altivec_vaddcuw): Likewise.
15398 (altivec_vaddu<VI_char>s): Likewise.
15399 (altivec_vadds<VI_char>s): Likewise.
15400 (sub<mode>3): Likewise.
15401 (altivec_vsubcuw): Likewise.
15402 (altivec_vsubu<VI_char>s): Likewise.
15403 (altivec_vsubs<VI_char>s): Likewise.
15404 (altivec_vavgs<VI_char>): Likewise.
15405 (altivec_vcmpbfp): Likewise.
15406 (altivec_eq<mode>): Likewise.
15407 (altivec_gt<mode>): Likewise.
15408 (altivec_gtu<mode>): Likewise.
15409 (umax<mode>3): Likewise.
15410 (smax<mode>3): Likewise.
15411 (umin<mode>3): Likewise.
15412 (smin<mode>3): Likewise.
15413 (altivec_vpkuhum): Likewise.
15414 (altivec_vpkuwum): Likewise.
15415 (altivec_vpkshss): Likewise.
15416 (altivec_vpkswss): Likewise.
15417 (altivec_vpkuhus): Likewise.
15418 (altivec_vpkshus): Likewise.
15419 (altivec_vpkuwus): Likewise.
15420 (altivec_vpkswus): Likewise.
15421 (altivec_vpks<VI_char>ss): Likewise.
15422 (altivec_vpks<VI_char>us): Likewise.
15423 (altivec_vpku<VI_char>us): Likewise.
15424 (altivec_vpku<VI_char>um): Likewise.
15425 (altivec_vrl<VI_char>): Likewise.
15426 (altivec_vsl<VI_char>): Likewise.
15427 (altivec_vsr<VI_char>): Likewise.
15428 (altivec_vsra<VI_char>): Likewise.
15429 (altivec_vsldoi_<mode>): Likewise.
15430 (altivec_vupkhsb): Likewise.
15431 (altivec_vupkhs<VU_char>): Likewise.
15432 (altivec_vupkls<VU_char>): Likewise.
15433 (altivec_vupkhsh): Likewise.
15434 (altivec_vupklsb): Likewise.
15435 (altivec_vupklsh): Likewise.
15436 (altivec_vcmpequ<VI_char>_p): Likewise.
15437 (altivec_vcmpgts<VI_char>_p): Likewise.
15438 (altivec_vcmpgtu<VI_char>_p): Likewise.
15439 (abs<mode>2): Likewise.
15440 (vec_unpacks_hi_v16qi): Likewise.
15441 (vec_unpacks_hi_v8hi): Likewise.
15442 (vec_unpacks_lo_v16qi): Likewise.
15443 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15444 (vec_unpacks_lo_v8hi): Likewise.
15445 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15446 (vec_pack_trunc_v8h): Likewise.
15447 (vec_pack_trunc_v4si): Likewise.
15448 (vec_pack_trunc_<mode>): Likewise.
15449
15450 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15451 V2DI builtins.
15452 (vec_vmaxsd): Likewise.
15453 (vec_vmaxud): Likewise.
15454 (vec_vminsd): Likewise.
15455 (vec_vminud): Likewise.
15456 (vec_vpksdss): Likewise.
15457 (vec_vpksdus): Likewise.
15458 (vec_vpkudum): Likewise.
15459 (vec_vpkudus): Likewise.
15460 (vec_vrld): Likewise.
15461 (vec_vsld): Likewise.
15462 (vec_vsrad): Likewise.
15463 (vec_vsrd): Likewise.
15464 (vec_vsubudm): Likewise.
15465 (vec_vupkhsw): Likewise.
15466 (vec_vupklsw): Likewise.
15467
e70670cf
JH		 154682013-05-29 Jan Hubicka <jh@suse.cz>
15469
15470 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15471 flags; reorder rest of fields in more consistent way.
15472 (varpool_node): Remove analyzed, finalized and alias.
15473 (cgraph_ndoe): Likewise.
15474 (symtab_alias_ultimate_target): New function.
15475 (cgraph_function_node): Move offline.
15476 (cgraph_reset_node): Declare.
15477 (cgraph_comdat_can_be_unshared_p): Remove.
15478 (varpool_remove_initializer): Declare.
15479 (varpool_first_defined_variable, varpool_next_defined_variable
15480 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15481 (cgraph_function_with_gimple_body_p): Update.
15482 (varpool_all_refs_explicit_p): Update.
15483 (symtab_alias_target): New function.
15484 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15485 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 15486 (cgraph_function_or_thunk_node): Simplify using
15487 symtab_alias_ultimate_target.
e70670cf
JH		 15488 (varpool_variable_node): Likewise.
15489 * cgraph.c (cgraph_create_function_alias): Update.
15490 (cgraph_add_thunk): Update.
15491 (cgraph_remove_node): Update.
15492 (dump_cgraph_node): Do not dump removed flags.
15493 (cgraph_function_body_availability): Update.
15494 (cgraph_propagate_frequency): Update.
15495 (verify_cgraph_node): Check sanity of local flag.
15496 (cgraph_function_node): Move here from cgraph.h; revamp for
15497 cgraph_function_or_thunk_node.
15498 * lto-symtab.c (lto_varpool_replace_node): Update.
15499 (lto_symtab_resolve_can_prevail_p): Update.
15500 (lto_symtab_merge_cgraph_nodes): Update.
15501 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15502 propagate_constants_accross_call, devirtualization_time_bonus,
15503 ipcp_propagate_stage): Update.
15504 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 15505 * ipa-inline-transform.c (clone_inlined_nodes,
15506 preserve_function_body_p): Update.
e70670cf
JH		 15507 * ipa-reference.c (propagate): Update.
15508 (write_node_summary_p): Update.
15509 * toplev.c (wrapup_global_declaration_2): Update.
15510 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15511 (analyze_function) ... this one.
15512 (cgraph_process_new_functions): Update.
15513 (cgraph_reset_node): Export.
15514 (cgraph_finalize_function): Update.
15515 (cgraph_add_new_function): Update.
15516 (process_function_and_variable_attributes): Update.
15517 (varpool_finalize_decl): Update.
15518 (symbol_finalized): Remove.
15519 (symbol_finalized_and_needed): Rename to ...
15520 (symbol_defined_and_needed): ... update.
15521 (cgraph_analyze_functions): Update.
15522 (handle_alias_pairs): Update.
15523 (mark_functions_to_output): Update.
15524 (assemble_thunk): Update.
15525 (output_in_order): Update.
15526 (output_weakrefs): Update.
15527 (finalize_compilation_unit): Update.
15528 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15529 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15530 input_node, input_varpool_node): Update.
15531 * dbxout.c (dbxout_expand_expr): Update.
15532 * cgraphclones.c (cgraph_clone_node): Update.
15533 (cgraph_copy_node_for_versioning): Update.
15534 (cgraph_materialize_clone): Update.
15535 (cgraph_materialize_all_clones): Update.
15536 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15537 propagate_pure_const, propagate_nothrow): Update.
15538 * lto-streamer-out.c (lto_output, write_symbol): Update.
15539 * ipa-utils.c (ipa_reverse_postorder): Update.
15540 * ipa-inline.c (can_inline_edge_p): Update.
15541 (update_caller_keys, ipa_inline): Update.
15542 * dwarf2out.c (reference_to_unused,
15543 premark_types_used_by_global_vars_helper): Update.
15544 * tree-eh.c (tree_could_trap_p): Update.
15545 * ipa-split.c (consider_split, execute_split_functions): Update.
15546 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 15547 has_addr_references_p): Update; move ahead in file for better
15548 readability.
e70670cf
JH		 15549 (process_references): Simplify.
15550 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15551 bodies are removed.
15552 (cgraph_comdat_can_be_unshared_p): Make static.
15553 (cgraph_externally_visible_p): Update.
15554 (varpool_externally_visible_p): Update.
15555 (function_and_variable_visibility): Update.
15556 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15557 ipa_tm_mark_force_output_node): Update.
15558 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15559 estimate_edge_devirt_benefit, inline_generate_summary,
15560 inline_write_summary): Update.
15561 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15562 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 15563 (ipa_print_node_params, ipa_prop_read_section,
15564 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 15565 * varasm.c (mark_decl_referenced): Update.
15566 (assemble_alias, dump_tm_clone_pairs): Update.
15567 * tree-inline.c (copy_bb): Update.
15568 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15569 Update.
15570 * symtab.c (dump_symtab_base): Print new flags.
15571 (verify_symtab_base): Verify new flags.
15572 (symtab_alias_ultimate_target): New function.
15573 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15574 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15575 Update.
c1e183a9
UB		 15576 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15577 Update.
e70670cf
JH		 15578 * i386.c (ix86_get_function_versions_dispatcher,
15579 ix86_generate_version_dispatcher_body): Update.
15580 (fold_builtin_cpu): Use varpool_add_new_variable.
15581 * varpool.c (varpool_remove_initializer): Break out from ...
15582 (varpool_remove_node): ... this one.
15583 (dump_varpool_node, varpool_node_for_asm,
15584 cgraph_variable_initializer_availability, varpool_analyze_node,
15585 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 15586 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15587 Update.
e70670cf 15588
182802ad
JH		 155892013-05-29 Jan Hubicka <jh@suse.cz>
15590
15591 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15592
8190b609
ER		 155932013-05-29 Easwaran Raman <eraman@google.com>
15594
15595 PR tree-optimization/57442
15596 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15597 when control exits the main loop.
15598
69f5aa9b
SKS		 155992013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15600
15601 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15602 and RX600.
c1e183a9 15603 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 15604 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15605 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15606 option.
15607
4b847da9
BS		 156082013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15609
15610 PR tree-optimization/57441
15611 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15612 Don't limit size of incr_vec to number of candidates.
15613
4fc43c39
SE		 156142013-05-29 Steve Ellcey <sellcey@imgtec.com>
15615
15616 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15617 and mips16 directories.
c1e183a9 15618 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 15619 (MULTILIB_DIRNAMES): Ditto.
15620 (MULTILIB_EXCEPTIONS): Add new exceptions.
15621 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15622 (MULTILIB_DIRNAMES): Ditto.
15623 (MULTILIB_EXCEPTIONS): Add new exceptions.
15624
12211b99 156252012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 15626 Marcus Shawcroft <marcus.shawcroft@arm.com>
15627
15628 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15629 SYMBOL_TINY_ABSOLUTE.
15630 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15631 SYMBOL_TINY_ABSOLUTE.
15632 (aarch64_expand_mov_immediate): Likewise.
15633 (aarch64_classify_symbol): Likewise.
15634 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15635 Permit SYMBOL_TINY_ABSOLUTE.
15636 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15637
12211b99 156382013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 15639 Marcus Shawcroft <marcus.shawcroft@arm.com>
15640
15641 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15642 Refactor if/switch. Replace gcc_assert with if.
15643
c0186656
GG		 156442013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15645
15646 * config/i386/i386.c (initial_ix86_tune_features): Enable
15647 FP Reassociation for AMD bdver1 and bdver2.
15648
d20188f3
MJ		 156492013-05-29 Martin Jambor <mjambor@suse.cz>
15650
15651 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15652 and IMAGPART_EXPR do not occur within other handled_components.
15653
292cba13
RB		 156542013-05-29 Richard Biener <rguenther@suse.de>
15655
15656 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15657 access on whether the use is in the BB we currently try to
15658 vectorize.
15659 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15660 vectorize to vect_bb_slp_scalar_cost.
15661
6eddf228
RB		 156622013-05-29 Richard Biener <rguenther@suse.de>
15663
15664 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15665 computing scalar cost offsetted by stmts that are kept live
15666 by scalar uses.
15667 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15668 for computation of scalar cost.
15669
7df36117
SE		 156702013-05-28 Steve Ellcey <sellcey@mips.com>
15671
15672 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15673
36536d79
BI		 156742013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15675
15676 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15677 array notation built-in reduction functions.
15678 * doc/passes.texi (Passes): Added documentation about changes done
15679 for Cilk Plus.
15680 * doc/invoke.texi (C Dialect Options): Added documentation about
15681 the -fcilkplus flag.
15682 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15683 (BUILTINS_DEF): Depend on cilkplus.def.
15684 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15685 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15686 * cilkplus.def: New file.
15687
12211b99 156882013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 15689
15690 PR rtl-optimization/57439
15691 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15692
9daf6dbb
ER		 156932013-05-28 Easwaran Raman <eraman@google.com>
15694
15695 PR tree-optimization/57337
15696 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15697 (find_insert_point): Correctly identify the insertion point
15698 when two statements with the same UID is compared.
15699
fbd7e877
RB		 157002013-05-28 Richard Biener <rguenther@suse.de>
15701
15702 PR tree-optimization/56787
15703 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15704 from the list of data references.
15705 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15706 clobbers.
15707 (vect_analyze_loop_operations): Likewise.
15708 (vect_transform_loop): Remove clobbers.
15709
bbba1117
MJ		 157102013-05-28 Martin Jambor <mjambor@suse.cz>
15711
15712 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15713 and REALPART_EXPRs have scalar type.
15714
bd388c2a
RB		 157152013-05-28 Richard Biener <rguenther@suse.de>
15716
15717 PR tree-optimization/57411
15718 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15719 virtual operands.
15720 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15721 virtual operand propagation.
15722
2f56a311
EB		 157232013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15724
15725 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15726 destination register for bmasksi_vis.
15727 (vector_init_bshuffle): Likewise.
15728 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15729
5c3eacbb
EB		 157302013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15731
15732 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15733 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15734 mode if the instruction isn't available in the original mode.
15735 * config/sparc/sparc.opt (mfix-ut699): New option.
15736 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15737 (divdf3): Turn into expander.
15738 (divdf3_nofix): New insn.
15739 (divdf3_fix): Likewise.
15740 (divsf3): Disable if -mfix-ut699.
15741 (sqrtdf2): Turn into expander.
15742 (sqrtdf2_nofix): New insn.
15743 (sqrtdf2_fix): Likewise.
15744 (sqrtsf2): Disable if -mfix-ut699.
15745
a1756c0a
RB		 157462013-05-27 Richard Biener <rguenther@suse.de>
15747
15748 PR middle-end/57412
15749 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15750 block for the new loop.
15751
5a892248
RB		 157522013-05-27 Richard Biener <rguenther@suse.de>
15753
15754 PR tree-optimization/57343
15755 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15756 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15757 (number_of_iterations_cond): Do not build the folded tree.
15758
d1de852b
RB		 157592013-05-27 Richard Biener <rguenther@suse.de>
15760
15761 Revert
15762 PR middle-end/57381
15763 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15764 OEP_CONSTANT_ADDRESS_OF retained.
15765
15766 PR tree-optimization/57417
15767 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15768 for unchanged base.
15769 (set_ssa_val_to): Compare addresses using
15770 get_addr_base_and_unit_offset.
15771
12211b99 157722013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 15773
15774 PR rtl-optimization/56833
15775 * postreload.c (move2add_record_mode): New function.
15776 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15777 (move2add_use_add2_insn): Use move2add_record_sym_value.
15778 (move2add_use_add3_insn): Likewise.
15779 (reload_cse_move2add): Use move2add_valid_value_p and
15780 move2add_record_mode. Invalidate call-clobbered and REG_INC
15781 affected regs by setting reg_mode to VOIDmode.
15782 (move2add_note_store): Don't pretend the inside of a SUBREG is
15783 the actual destination. Invalidate single/leading registers by
15784 setting reg_mode to VOIDmode.
15785 Use move2add_record_sym_value, move2add_valid_value_p and
15786 move2add_record_mode.
15787
b03be25f
RB		 157882013-05-27 Richard Biener <rguenther@suse.de>
15789
15790 PR tree-optimization/57396
15791 * tree-affine.c (double_int_constant_multiple_p): Properly
15792 return false for val == 0 and div != 0.
15793
44e88db2
RS		 157942013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15795
15796 * config/mips/mips.h: Use #elif in preprocessor conditions.
15797
3b859704
RS		 157982013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15799
15800 PR target/53916
15801 * config/mips/constraints.md (kl): New constraint.
15802 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15803 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15804 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15805 from LO for MIPS16.
15806 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15807
c979d5f5
RS		 158082013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15809
15810 PR target/55777
15811 * config/mips/mips.c (mips_can_inline_p): New function.
15812 (TARGET_CAN_INLINE_P): Define.
15813
8e90de43
SB		 158142013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15815
15816 * sched-int.h (ds_t, dw_t): Make unsigned int.
15817 Fix documentation that describes how all the ds_t bits are used.
15818 Reserve the last bit for delayed-branch scheduling.
15819 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15820 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15821 (gen_dep_weak_1): Remove prototype.
15822 * sched-deps.c (get_dep_weak_1): Make static.
15823 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15824 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15825 * doc/tm.texi: Regenerate.
15826 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15827
cb5cb194
SB		 158282013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15829
15830 PR debug/56950
15831 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15832
5bd93ff6
NS		 158332013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15834 Sandra Loosemore <sandra@codesourcery.com>
15835
15836 * config.gcc (powerpc-*): Allow native for with-cpu.
15837
2343af65
JL		 158382013-05-24 Jeff Law <law@redhat.com>
15839
15840 PR tree-optimization/57124
15841 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15842 conversion feeding a condition if the range has an overflow
15843 if -fstrict-overflow. Add warnings for when we do make the
15844 transformation.
15845
3ad695b9
DC		 158462013-05-24 Dehao Chen <dehao@google.com>
15847
04960246 15848 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 15849 (locus_discrim_hasher): Likewise.
15850 (locus_discrim_hasher::hash): Likewise.
15851 (locus_discrim_hasher::equal): Likewise.
15852
82338059
MJ		 158532013-05-24 Martin Jambor <mjambor@suse.cz>
15854
15855 PR tree-optimization/57294
15856 * cgraph.h (ipa_record_stmt_references): Declare.
15857 * cgraphbuild.c (ipa_record_stmt_references): New function.
15858 (build_cgraph_edges): Use ipa_record_stmt_references.
15859 (rebuild_cgraph_edges): Likewise.
15860 (cgraph_rebuild_references): Likewise.
15861 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15862 associated with the old statement and build references from the
15863 newly built statements.
15864 * ipa-ref.c (ipa_remove_stmt_references): New function.
15865 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15866
1ccd4874
VM		 158672013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15868
55805e54 15869 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 15870 mem-mem moves.
55805e54 15871 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 15872 too.
55805e54 15873 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 15874 secondary memory moves.
15875 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15876 reg set up in the current insn.
15877
25e25c73
DC		 158782013-05-24 Dehao Chen <dehao@google.com>
15879
04960246 15880 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 15881 hash function.
15882 (locus_descrim_hasher::equal): Likewise.
04960246 15883 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 15884 (make_edges): Likewise.
15885 (next_discriminator_for_locus): Likewise.
15886 (same_line_p): Likewise.
15887 (assign_discriminators): Likewise.
15888 (make_cond_expr_edges): Likewise.
15889 (make_gimple_switch_edges): Likewise.
15890 (make_goto_expr_edges): Likewise.
15891 (make_gimple_asm_edges): Likewise.
15892
50d38551
IB		 158932013-05-24 Ian Bolton <ian.bolton@arm.com>
15894
15895 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15896 X format specifier to only display bottom 16 bits.
15897 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15898 immediate to match for operand 2, since it will be masked.
15899
aea0101d
RB		 159002013-05-24 Richard Biener <rguenther@suse.de>
15901
15902 PR tree-optimization/57287
15903 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15904 all SSA names that occur in abnormal PHIs.
15905
634e03d3
AI		 159062013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15907
15908 PR tree-ssa/57385
15909 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15910 that index is not negative.
15911
b17c024f
EB		 159122013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15913
15914 PR rtl-optimization/55177
15915 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15916 (simplify_byte_swapping_operation): New.
15917 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15918 (simplify_relational_operation_1): Deal with BSWAP.
15919
46aeac1b
RH		 159202013-05-23 Richard Henderson <rth@redhat.com>
15921
15922 PR target/56742
15923 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15924 (ix86_reorg): Call it.
15925
70cc1536
UB		 159262013-05-23 Uros Bizjak <ubizjak@gmail.com>
15927
15928 PR target/57379
15929 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
15930 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
15931 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
15932
b3851501
CB		 159332013-05-23 Christian Bruel <christian.bruel@st.com>
15934
15935 PR debug/57351
15936 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
15937
12211b99 159382013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 15939 Marcus Shawcroft <marcus.shawcroft@arm.com>
15940
15941 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
15942 * config/aarch64/constraints.md (Usa): Remove.
15943 * doc/md.texi (AArch64 Usa): Remove.
15944
12211b99 159452013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 15946 Marcus Shawcroft <marcus.shawcroft@arm.com>
15947
15948 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
15949 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
15950 * config/aarch64/predicates.md (aarch64_const_address): Remove.
15951 (aarch64_mov_operand): Use aarch64_mov_operand_p.
15952
12211b99 159532013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 15954
15955 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
15956 instruction (AdvSIMD).
15957 * config/aarch64/aarch64-builtins.c
15958 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
15959 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
15960
76c36cb1 159612013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 15962
15963 PR middle-end/57347
15964 * tree.h (contains_bitfld_component_ref_p): Declare.
15965 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 15966 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
15967 caller.
7d2fb524
MJ		 15968 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
15969 not access a bit-field. Assert all final offsets are byte-aligned.
15970
ce521ae6
RB		 159712013-05-23 Richard Biener <rguenther@suse.de>
15972
15973 PR tree-optimization/57380
15974 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
15975 least one invariant or re-used load.
15976 * passes.c (init_optimization_passes): Move pass_phiprop before
15977 pass_forwprop.
15978
75c7257f
JG		 159792013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
15980
15981 * config/aarch64/aarch64-simd.md
15982 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
15983
af6d44b5
RB		 159842013-05-23 Richard Biener <rguenther@suse.de>
15985
15986 PR middle-end/57381
15987 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15988 OEP_CONSTANT_ADDRESS_OF retained.
15989
bd3647bf
JJ		 159902013-05-23 Jakub Jelinek <jakub@redhat.com>
15991
15992 PR middle-end/57344
70cc1536
UB		 15993 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
15994 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 15995
a55757ea
RB		 159962013-05-23 Richard Biener <rguenther@suse.de>
15997
15998 PR rtl-optimization/57341
15999 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16000 instead of true_dependence.
16001
cfb00b41
DM		 160022013-05-22 David Malcolm <dmalcolm@redhat.com>
16003
dd1c676f
DM		 16004 * bb-reorder.c (branch_threshold): Make const.
16005 (exec_threshold): Ditto.
cfb00b41 16006
f62511da 160072013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16008 Pat Haugen <pthaugen@us.ibm.com>
16009 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16010
16011 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16012 documentation for the power8 crypto builtins.
16013
16014 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16015
16016 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16017 macros for defining power8 builtin functions.
16018 (BU_P8V_AV_2): Likewise.
16019 (BU_P8V_AV_P): Likewise.
16020 (BU_P8V_VSX_1): Likewise.
16021 (BU_P8V_OVERLOAD_1): Likewise.
16022 (BU_P8V_OVERLOAD_2): Likewise.
16023 (BU_CRYPTO_1): Likewise.
16024 (BU_CRYPTO_2): Likewise.
16025 (BU_CRYPTO_3): Likewise.
16026 (BU_CRYPTO_OVERLOAD_1): Likewise.
16027 (BU_CRYPTO_OVERLOAD_2): Likewise.
16028 (XSCVSPDP): Fix typo, point to the correct instruction.
16029 (VCIPHER): Add power8 crypto builtins.
16030 (VCIPHERLAST): Likewise.
16031 (VNCIPHER): Likewise.
16032 (VNCIPHERLAST): Likewise.
16033 (VPMSUMB): Likewise.
16034 (VPMSUMH): Likewise.
16035 (VPMSUMW): Likewise.
16036 (VPERMXOR_V2DI): Likewise.
16037 (VPERMXOR_V4SI: Likewise.
16038 (VPERMXOR_V8HI: Likewise.
16039 (VPERMXOR_V16QI: Likewise.
16040 (VSHASIGMAW): Likewise.
16041 (VSHASIGMAD): Likewise.
16042 (VPMSUM): Likewise.
16043 (VPERMXOR): Likewise.
16044 (VSHASIGMA): Likewise.
16045
16046 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16047 __CRYPTO__ if the crypto instructions are available.
16048 (altivec_overloaded_builtins): Add support for overloaded power8
16049 builtins.
16050
16051 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16052 support for power8 crypto builtins.
16053 (builtin_function_type): Likewise.
16054 (altivec_init_builtins): Add support for builtins that take vector
16055 long long (V2DI) arguments.
16056
16057 * config/rs6000/crypto.md: New file, define power8 crypto
16058 instructions.
16059
160602013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16061 Pat Haugen <pthaugen@us.ibm.com>
16062 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16063
16064 * doc/invoke.texi (Option Summary): Add power8 options.
16065 (RS/6000 and PowerPC Options): Likewise.
16066
16067 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16068 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16069 wm, wn, wr documentation.
16070
f43245d1 16071 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM		 16072 registers if direct move instructions are enabled.
16073 (wn): New constraint for no registers.
16074 (wq): New constraint for quad word even GPR registers.
16075 (wr): New constraint if 64-bit instructions are enabled.
16076 (wv): New constraint if power8 vector instructions are enabled.
16077 (wQ): New constraint for quad word memory locations.
16078
f43245d1 16079 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM		 16080 constraint for 0..15 for crypto instructions.
16081 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16082 as GPR and floating point registers.
16083 (int_reg_operand): New predicate to match only GPR registers.
16084 (base_reg_operand): New predicate to match base registers.
16085 (quad_int_reg_operand): New predicate to match even GPR registers
16086 for quad memory operations.
16087 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16088 operations in both GPR and VSX registers.
16089 (quad_memory_operand): New predicate for quad memory operations.
16090 (reg_or_indexed_operand): New predicate for direct move support.
16091
f43245d1 16092 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM		 16093 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16094 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16095 (POWERPC_MASKS): Add power8 options.
16096 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16097 various options.
16098
f43245d1 16099 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM		 16100 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16101
f43245d1 16102 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM		 16103 (-mpower8-fusion): New power8 options.
16104 (-mpower8-fusion-sign): Likewise.
16105 (-mpower8-vector): Likewise.
16106 (-mcrypto): Likewise.
16107 (-mdirect-move): Likewise.
16108 (-mquad-memory): Likewise.
16109
f43245d1 16110 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM		 16111 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16112 registers.
70cc1536 16113 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 16114 (rs6000_debug_vector_unit): Add p8_vector.
16115 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16116 definitions. Also print fusion state.
16117 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16118 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16119 (rs6000_option_override_internal): Add support for power8.
16120 (rs6000_common_init_builtins): Add debugging for skipped builtins
16121 if -mdebug=builtin.
16122 (rs6000_adjust_cost): Add power8 support.
16123 (rs6000_issue_rate): Likewise.
16124 (insn_must_be_first_in_group): Likewise.
16125 (insn_must_be_last_in_group): Likewise.
16126 (force_new_group): Likewise.
16127 (rs6000_register_move_cost): Likewise.
16128 (rs6000_opt_masks): Likewise.
16129
16130 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16131 power8 capable assembler, default to power7 options.
16132 (TARGET_DIRECT_MOVE): Likewise.
16133 (TARGET_CRYPTO): Likewise.
16134 (TARGET_P8_VECTOR): Likewise.
16135 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16136 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16137 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16138 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16139 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16140 (TARGET_XSCVDPSPN): Likewise.
16141 (TARGET_XSCVSPDPN): Likewsie.
16142 (TARGET_SYNC_HI_QI): Likewise.
16143 (TARGET_SYNC_TI): Likewise.
16144 (MASK_CRYPTO): Likewise.
16145 (MASK_DIRECT_MOVE): Likewise.
16146 (MASK_P8_FUSION): Likewise.
16147 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 16148 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16149 temporary used by some of the direct move instructions to get two FP
16150 temporary registers does not force creation of a stack frame.
f62511da
MM		 16151 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16152 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16153 that any VSX registers are tieable, even if they are also an
16154 Altivec vector mode.
16155 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16156 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16157 (RS6000_BTM_CRYPTO): Likewise.
16158 (RS6000_BTM_COMMON): Likewise.
16159
16160 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16161 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16162 (enum rs6000_vector): Add power8 vector support.
16163
73a1a707
RR		 161642013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16165
16166 PR target/19599
16167 PR target/57340
16168 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16169 (any_sibcall_could_use_r3): this and handle indirect calls.
16170 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16171
d28073d4
BS		 161722013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16173
16174 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16175
bee0b10c
RB		 161762013-05-22 Richard Biener <rguenther@suse.de>
16177
16178 PR middle-end/57349
16179 * profile.c (branch_prob): Do not split blocks that are
16180 abnormally receiving from ECF_RETURNS_TWICE functions.
16181
98409b51
RS		 161822013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16183
16184 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16185 address mode. Move pointer mode initialization to the same place.
16186
c0602ab8
MZ		 161872013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16188
16189 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16190 while it has any effect.
16191
4f7a634e
ER		 161922013-05-21 Easwaran Raman <eraman@google.com>
16193
16194 PR tree-optimization/57322
9daf6dbb
ER		 16195 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16196 UID of the statement added to the BB to be 1.
4f7a634e 16197
a4ce1258
JJ		 161982013-05-21 Jakub Jelinek <jakub@redhat.com>
16199
16200 PR tree-optimization/57331
70cc1536
UB		 16201 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16202 of conversion from pointer type to integral type with integer.
a4ce1258 16203
1b14621a
MJ		 162042013-05-21 Martin Jambor <mjambor@suse.cz>
16205
16206 PR lto/57289
16207 * ipa-prop.c (ipa_read_node_info): Process param_used and
16208 controlled_uses in the same order as when writing.
16209
e60661f0
MG		 162102013-05-21 Magnus Granberg <baldrick@free.fr>
16211
16212 PR plugins/56754
ee49aa34 16213 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16214
4a61be9a
RB		 162152013-05-21 Richard Biener <rguenther@suse.de>
16216
16217 PR tree-optimization/57318
16218 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16219 estimate stmts with side-effects as likely eliminated.
16220
c52da5f7
RB		 162212013-05-21 Richard Biener <rguenther@suse.de>
16222
16223 PR tree-optimization/57330
16224 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16225 preserve the call stmts fntype.
16226
7ec67e2a
RB		 162272013-05-21 Richard Biener <rguenther@suse.de>
16228
16229 PR tree-optimization/57303
16230 * tree-ssa-sink.c (statement_sink_location): Improve killing
16231 stmt detection and properly handle self-assignments.
16232
b112d513
CB		 162332013-05-21 Christian Bruel <christian.bruel@st.com>
16234
55805e54
YR		 16235 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16236 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 16237 registers. Set register size out of the PARALLEL loop.
16238
14c2ec26
OE		 162392013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16240
16241 PR target/56547
16242 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16243 (*fmasf4, *fmasf4_media): New insns.
16244
da734fa1
RS		 162452013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16246
16247 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16248 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16249 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16250 (mips_idiv_insns): Update the comments to say that the returned
16251 instruction counts are in units of BASE_INSN_LENGTH.
16252 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16253 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16254 using 2 rather than 4 as the length of indirect MIPS16 and
16255 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16256 length of a NOP. Don't divide MIPS16 lengths by 2.
16257 (mips16_split_long_branches): Assume a branch is long if the
16258 length is greater than 4 rather than 8.
16259 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16260 rather than multiplying them by 2. Multiply instruction counts
16261 by BASE_INSN_LENGTH rather than 4.
16262 (*jump_mips16, tls_get_tp_mips16_<mode>)
16263 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16264
13719e8b
RS		 162652013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16266
16267 * config/mips/mips.md (extended_mips16): Remove branch case.
16268 (length): Remove duplicated extended_mips16 test.
16269
c3850d14
RS		 162702013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16271
16272 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16273
8da2e059
RS		 162742013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16275
16276 * recog.h (Recog_data): Rename to...
16277 (recog_data_d): ...this.
16278 (recog_data): Update accordingly.
16279 * recog.c (recog_data): Likewise.
16280 * reload.c (save_recog_data): Likewise.
16281 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16282 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16283
2df013f3
JB		 162842013-05-17 Julian Brown <julian@codesourcery.com>
16285
16286 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16287 found in a REG_EQUAL note, invalidate it.
16288
371e77e3 162892013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 16290
16291 * tree-ssa-reassoc.c (find_insert_point): New function.
16292 (insert_stmt_after): Likewise.
16293 (get_def_stmt): Likewise.
16294 (ensure_ops_are_available): Likewise.
16295 (not_dominated_by): Likewise.
16296 (rewrite_expr_tree): Do not move statements beyond what is
16297 necessary. Remove call to swap_ops_for_binary_stmt...
16298 (reassociate_bb): ... and move it here.
16299 (build_and_add_sum): Assign UIDs for new statements.
16300 (linearize_expr): Likewise.
16301 (do_reassoc): Renumber gimple statement UIDs.
16302
e01c7cca
JH		 163032013-05-17 Jan Hubicka <jh@suse.cz>
16304
16305 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16306 weakrefs.
16307 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 16308 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16309 weakrefs.
e01c7cca
JH		 16310 (output_weakrefs): Update.
16311
c3272a92
PCC		 163122013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16313 Martin Jambor <mjambor@suse.cz>
16314
16315 PR middle-end/57276
16316 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16317 value that corresponds to the given aggval is found in values vector.
16318
11c2aa39
UB		 163192013-05-17 Uros Bizjak <ubizjak@gmail.com>
16320
16321 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16322 sse, sse2, sse3, ssse3 and sse4a flags to options.
16323
abecc8c6
DM		 163242013-05-17 David Malcolm <dmalcolm@redhat.com>
16325
16326 * gengtype-state.c: (s_expr_writer): New class, to handle
16327 prettifying of output layout of s-expressions.
16328 (state_writer): New class, to write out gtype.state.
16329 (state_written_type_count): Move this variable into member data of
16330 state_writer.
16331 (s_expr_writer::s_expr_writer): New code: constructor for new class
16332 (state_writer::state_writer(): ditto
16333 (s_expr_writer::write_new_line): New function
16334 (s_expr_writer::write_any_indent): ditto
16335 (s_expr_writer::begin_s_expr): ditto
16336 (s_expr_writer::end_s_expr): ditto
16337 (write_state_fileloc): convert to method of state_writer...
16338 (state_writer:: write_state_fileloc): ...and use methods of
16339 s_expr_writer to write indentation into the gtype.state output file
16340 to visually represent the hierarchical structure of the list
16341 structures
16342 (write_state_fields): ditto, renaming to...
16343 (state_writer::write_state_fields)
16344 (write_state_a_string): ditto, renaming to...
16345 (state_writer::write_state_a_string)
16346 (write_state_string_option): ditto, renaming to...
16347 (state_writer::write_state_string_option)
16348 (write_state_type_option): ditto, renaming to...
16349 (state_writer::write_state_type_option)
16350 (write_state_nested_option): ditto, renaming to...
16351 (state_writer::write_state_nested_option)
16352 (write_state_option): ditto, renaming to...
16353 (state_writer::write_state_option)
16354 (write_state_options): ditto, renaming to...
16355 (state_writer::write_state_options)
16356 (write_state_lang_bitmap): ditto, renaming to...
16357 (state_writer::write_state_lang_bitmap)
16358 (write_state_version): ditto, renaming to...
16359 (state_writer::write_state_version)
16360 (write_state_scalar_type): ditto, renaming to...
16361 (state_writer::write_state_scalar_type)
16362 (write_state_string_type): ditto, renaming to...
16363 (state_writer::write_state_string_type)
16364 (write_state_undefined_type): ditto, renaming to...
16365 (state_writer::write_state_undefined_type)
16366 (write_state_struct_union_type): ditto, renaming to...
16367 (state_writer::write_state_struct_union_type)
16368 (write_state_struct_type): ditto, renaming to...
16369 (state_writer::write_state_struct_type)
16370 (write_state_user_struct_type): ditto, renaming to...
16371 (state_writer::write_state_user_struct_type)
16372 (write_state_lang_struct_type): ditto, renaming to...
16373 (state_writer::write_state_lang_struct_type)
16374 (write_state_param_struct_type): ditto, renaming to...
16375 (state_writer::write_state_param_struct_type)
16376 (write_state_pointer_type): ditto, renaming to...
16377 (state_writer::write_state_pointer_type)
16378 (write_state_array_type): ditto, renaming to...
16379 (state_writer::write_state_array_type)
16380 (write_state_gc_used): ditto, renaming to...
16381 (state_writer::write_state_gc_used)
16382 (write_state_common_type_content): ditto, renaming to...
16383 (state_writer::write_state_common_type_content)
16384 (write_state_type): ditto, renaming to...
16385 (state_writer::write_state_type)
16386 (write_state_pair_list): ditto, renaming to...
16387 (state_writer::write_state_pair_list)
16388 (write_state_pair): ditto, renaming to...
16389 (state_writer::write_state_pair)
16390 (write_state_typedefs): ditto, renaming to...
16391 (state_writer::write_state_typedefs)
16392 (write_state_structures): ditto, renaming to...
16393 (state_writer::write_state_structures)
16394 (write_state_param_structs): ditto, renaming to...
16395 (state_writer::write_state_param_structs)
16396 (write_state_variables): ditto, renaming to...
16397 (state_writer::write_state_variables)
16398 (write_state_srcdir): ditto, renaming to...
16399 (state_writer::write_state_srcdir)
16400 (write_state_files_list): ditto, renaming to...
16401 (state_writer::write_state_files_list)
16402 (write_state_languages): ditto, renaming to...
16403 (state_writer::write_state_languages)
16404 (write_state): create a state_writer instance and use it when
16405 writing out the state file
16406
d6545f29
MS		 164072013-05-17 Mike Stump <mikestump@comcast.net>
16408
816a3f73 16409 PR rtl-optimization/57304
d6545f29
MS		 16410 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16411 accessing DF_REF_REAL_LOC.
16412
38c821cf
JJ		 164132013-05-17 Jakub Jelinek <jakub@redhat.com>
16414
16415 PR rtl-optimization/57281
16416 PR rtl-optimization/57300
16417 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16418 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16419 what the other splitter did if the registers are dead.
16420
2a293391
RB		 164212013-05-17 Richard Biener <rguenther@suse.de>
16422
16423 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16424 MEM_REF offsets.
16425
17042d2b
JJ		 164262013-05-17 Jakub Jelinek <jakub@redhat.com>
16427
16428 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16429 linking.
16430
5b115c1f
MP		 164312013-05-17 Marek Polacek <polacek@redhat.com>
16432
11c2aa39
UB		 16433 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16434 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 16435
52d84413
JJ		 164362013-05-17 Jakub Jelinek <jakub@redhat.com>
16437
68119618
JJ		 16438 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16439 vect_external_def oprnd1 with loop_vinfo, try to emit
16440 optional cast, negation and and stmts on the loop preheader
16441 edge instead of into the pattern def seq.
16442
52d84413
JJ		 16443 PR tree-optimization/57051
16444 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16445 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16446
17b962bd
NC		 164472013-05-16 Nick Clifton <nickc@redhat.com>
16448
16449 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16450 (rl78_is_naked_func): New function.
56aefbf7
UB		 16451 (rl78_expand_prologue): Skip prologue generation for naked functions.
16452 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 16453 * doc/extend.texi (naked): Add RL78 to the list of processors
16454 that supports this attribute.
16455
b1a0f84e
JL		 164562013-05-16 Jeff Law <law@redhat.com>
16457
16458 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16459
af0e415b
UB		 164602013-05-16 Uros Bizjak <ubizjak@gmail.com>
16461
16462 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16463 cache parameters using detect_caches_amd also for CYRIX,
16464 NSC and TM2 signatures.
16465
19db293a
UB		 164662013-05-16 Uros Bizjak <ubizjak@gmail.com>
16467 Dzianis Kahanovich <mahatma@eu.by>
16468
16469 PR target/45359
16470 PR target/46396
16471 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16472 VIA/Centaur processors and determine their cache parameters
16473 using detect_caches_amd.
16474
251a41b9
TJ		 164752013-05-16 Teresa Johnson <tejohnson@google.com>
16476
16477 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16478 (rtl_verify_edges): New function.
16479 (rtl_verify_bb_insns): Ditto.
16480 (rtl_verify_bb_pointers): Ditto.
16481 (rtl_verify_bb_insn_chain): Ditto.
16482 (rtl_verify_fallthru): Ditto.
16483 (rtl_verify_bb_layout): Ditto.
16484 (rtl_verify_flow_info_1): Outline checks into new functions.
16485 (rtl_verify_flow_info): Ditto.
16486
f14540b6
SE		 164872013-05-16 Steve Ellcey <sellcey@imgtec.com>
16488
16489 * cfghooks.c (copy_bbs): Add update_dominance argument.
16490 * cfghooks.h (copy_bbs): Update prototype.
16491 * tree-cfg.c (gimple_duplicate_sese_region):
16492 Add update_dominance argument.
16493 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16494 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16495 gimple_duplicate_sese_region call.
16496 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16497 Update copy_bbs call.
16498 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16499 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16500
7e9a3abb
JJ		 165012013-05-16 Jakub Jelinek <jakub@redhat.com>
16502
16503 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16504 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16505 vect_recog_rotate_pattern.
16506 (vect_recog_rotate_pattern): New function.
16507
427b248d
JM		 165082013-05-16 Jason Merrill <jason@redhat.com>
16509
16510 * Makefile.in (LLINKER): New variable.
16511 (mostlyclean): Remove link mutex.
16512 * configure.ac: Handle --enable-link-mutex.
16513 * lock-and-run.sh: New script.
16514
b871e3d2
RR		 165152013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16516
16517 PR target/19599
16518 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16519 for NULL decl.
16520
ce858126
RO		 165212013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16522
16523 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16524
798d3d04
GY		 165252013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16526
16527 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16528 * config/arm/arm.c (next_consecutive_mem): New function.
16529 (gen_movmem_ldrd_strd): Likewise.
16530 * config/arm/arm.md (movmemqi): Update condition and code.
16531 (unaligned_loaddi, unaligned_storedi): New patterns.
16532
0baddc45
RO		 165332013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16534
16535 * config.gcc: Obsolete *-*-solaris2.9*.
16536 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16537
aa9d5bcf
RB		 165382013-05-16 Richard Biener <rguenther@suse.de>
16539
16540 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16541 earlier, after GRAPHITE transforms and IV canonicalization.
16542
5a0f4dd3
JJ		 165432013-05-16 Jakub Jelinek <jakub@redhat.com>
16544
16545 * omp-low.c (extract_omp_for_data): For collapsed loops,
16546 if at least one of the loops is known at compile time to
16547 iterate zero times, set count to 0.
16548 (expand_omp_regimplify_p): New function.
16549 (expand_omp_for_generic): For collapsed loops, if at least
16550 one of the loops isn't known to iterate at least once,
16551 add runtime check with setting count to 0.
16552 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16553 For unsigned types if it isn't known at compile time that
16554 the loop will iterate at least once, add runtime check to bypass
16555 the whole loop if initial condition isn't true.
16556
e3753785
NS		 165572013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16558
16559 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16560
43bb4dd1
MG		 165612013-05-16 Marc Glisse <marc.glisse@inria.fr>
16562
16563 PR middle-end/57286
16564 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16565 transformations to avoid an infinite loop.
16566
3571dde6
MP		 165672013-05-16 Marek Polacek <polacek@redhat.com>
16568
16569 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16570
b8b3f0ca
LE		 165712013-05-15 Leif Ekblad <leif@rdos.net>
16572
16573 * config/i386/i386.c (ix86_decompose_address): Use
16574 DEFAULT_TLS_SEG_REG to access TLS segment register.
16575 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16576 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16577 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16578
e299a383
RS		 165792013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16580
16581 PR target/57260
16582 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16583 sibling calls to functions that would normally be lazily bound,
16584 unless $gp is call-clobbered.
16585
e7208ea3
UB		 165862013-05-15 Uros Bizjak <ubizjak@gmail.com>
16587
19db293a 16588 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 16589 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 16590 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16591 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 16592 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16593 of TARGET_3DNOW.
16594 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16595
e7413f3d
AS		 165962013-05-15 Andreas Schwab <schwab@suse.de>
16597
16598 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16599 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16600 third operand.
16601
ad4db775
TJ		 166022013-05-15 Teresa Johnson <tejohnson@google.com>
16603
16604 * loop-unroll.c (report_unroll_peel): Check decision before
16605 emitting unroll/peel message.
16606
af205f67
TJ		 166072013-05-15 Teresa Johnson <tejohnson@google.com>
16608
16609 * function.h (has_bb_partition): New rtl_data flag.
16610 (bb_reorder_complete): Ditto.
16611 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16612 instead of flag_reorder_blocks_and_partition.
16613 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16614 with some enhancements.
16615 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16616 * bb-reorder.c (connect_traces): Check for has_bb_partition
16617 instead of flag_reorder_blocks_and_partition.
16618 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16619 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16620 verify_hot_cold_block_grouping.
16621 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16622
9adcfa3c
RR		 166232013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16624
16625 PR target/19599
16626 * config/arm/predicates.md (call_insn_operand): New predicate.
16627 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16628 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16629 if insn is not a tail call.
16630 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16631 registers.
16632 * config/arm/arm.h (enum reg_class): New caller save register class.
16633 (REG_CLASS_NAMES): Likewise.
16634 (REG_CLASS_CONTENTS): Likewise.
16635 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16636 without decls.
16637
ccb3ad87
RB		 166382013-05-15 Richard Biener <rguenther@suse.de>
16639
16640 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16641 of MSG_OPTIMIZED_LOCATIONS.
16642 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16643 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16644 message.
16645 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16646 of MSG_OPTIMIZED_LOCATIONS.
16647 (execute_vect_slp): Likewise.
16648 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16649 (vect_create_cond_for_alias_checks): Likewise.
16650 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16651 (vect_recog_widen_mult_pattern): Likewise.
16652 (vect_recog_widen_sum_pattern): Likewise.
16653 (vect_recog_over_widening_pattern): Likewise.
16654 (vect_recog_widen_shift_pattern): Likewise.
16655 (vect_recog_vector_vector_shift_pattern): Likewise.
16656 (vect_recog_divmod_pattern): Likewise.
16657 (vect_recog_mixed_size_cond_pattern): Likewise.
16658 (vect_recog_bool_pattern): Likewise.
16659 (vect_pattern_recog_1): Likewise.
16660
48b1474e
MJ		 166612013-05-15 Martin Jambor <mjambor@suse.cz>
16662
16663 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16664 non-functions to builtin_unreachable.
16665 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16666 correct when new direct edges were discovered.
16667
9de04252
MJ		 166682013-05-15 Martin Jambor <mjambor@suse.cz>
16669
16670 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16671 header, print symbol order instead of node uid, print more information
16672 about indirect edge targets.
16673 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16674 uids.
16675 (ipa_make_edge_direct_to_target): Likewise.
16676 (remove_described_reference): Likewise.
16677 (propagate_controlled_uses): Likewise.
16678 (ipa_print_node_params): Also print symbol order.
16679 (ipcp_transform_function): Print symbol order instead of node uids.
16680 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16681 (cgraph_get_create_real_symbol_node): Likewise.
16682 * ipa-cp.c (print_lattice): Likewise.
16683 (print_all_lattices): Likewise.
16684 (determine_versionability): Likewise.
16685 (initialize_node_lattices): Likewise.
16686 (estimate_local_effects): Likewise.
16687 (update_profiling_info): Likewise.
16688 (create_specialized_node): Likewise.
16689 (perhaps_add_new_callers): Likewise.
16690 (decide_about_value): Likewise.
16691 (decide_whether_version_node): Likewise.
16692 (identify_dead_nodes): Likewise.
16693 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16694 (dump_inline_summary): Likewise.
16695 (estimate_node_size_and_time): Likewise.
16696 (inline_analyze_function): Likewise.
16697 * ipa-inline.c (report_inline_failed_reason): Likewise.
16698 (want_early_inline_function_p): Likewise.
16699 (edge_badness): Likewise.
16700 (update_edge_key): Likewise.
16701 (inline_small_functions): Likewise. Add dumping of order to two other
16702 dumps.
16703 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16704 instead of node uids.
16705 (propagate_pure_const): Likewise.
16706 (propagate_pure_const): Likewise.
16707 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16708 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16709 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16710 of node uids.
16711 * tree-pretty-print.c (dump_function_header): Likewise.
16712 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16713 Print symbol order instead of node uids.
16714
1dd03b91
AK		 167152013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16716
16717 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16718 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16719
fdf6a7b9
RB		 167202013-05-15 Richard Biener <rguenther@suse.de>
16721
16722 PR tree-optimization/57275
16723 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16724 return value for fail to do runtime alias checks for gather loads.
16725
2d6e4603
JH		 167262013-05-15 Jan Hubicka <jh@suse.cz>
16727
16728 PR lto/57038
16729 PR lto/47375
e7208ea3
UB		 16730 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16731 weakrefs are not external.
16732 (lto_symtab_merge_decls): Fix thinko when dealing with
16733 non-lto_symtab decls.
2d6e4603
JH		 16734 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16735 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16736 * varpool.c (dump_varpool_node): Dump more flags.
16737
83f44b39
GG		 167382013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16739
16740 * config/i386/i386.c (processor_alias_table): Add instruction
16741 FSGSBASE for AMD bdver3 architecture.
16742
2e55d062
JJ		 167432013-05-14 Jakub Jelinek <jakub@redhat.com>
16744
16745 * tree.c (warn_deprecated_use): Print file:line using locus color.
16746 * diagnostic.c (diagnostic_report_current_module): Print file:line
16747 and file:line:column using locus color.
16748
0bfdb81e
MS		 167492013-05-14 Mike Stump <mikestump@comcast.net>
16750
16751 * gdbinit.in: Add __null.
16752
a508ef22
MS		 167532013-05-14 Mike Stump <mikestump@comcast.net>
16754
16755 * recog.h: Rename struct recog_data to Recog_data.
16756 * recog.c: Likewise.
16757 * reload.c (can_reload_into): Likewise.
16758 * config/picochip/picochip.c: Likewise.
16759
e7180acb
MS		 167602013-05-14 Mike Stump <mikestump@comcast.net>
16761
16762 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16763
61216c26
SB		 167642013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16765
fd6beed4
SB		 16766 * resource.h (struct resources): Remove unch_memory member.
16767 (CLEAR_RESOURCE): Don't clear unch_memory.
16768 * resource.c (mark_referenced_resources): Don't set it.
16769 (mark_set_resources): Likewise.
16770 (mark_target_live_regs): Don't clear it.
16771 (init_resource_info): Likewise.
16772 * reorg.c (resource_conflicts_p): Don't compare it.
16773 (redundant_insn): Don't set it.
16774
61216c26
SB		 16775 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16776 Remove prototypes.
16777 * emit-rtl.c (next_label): Remove unused function.
16778 (skip_consecutive_labels, link_cc0_insns): Move to ...
16779 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16780 only place where these functions are used, and make them static.
16781
418d1b87
MG		 167822013-05-14 Marc Glisse <marc.glisse@inria.fr>
16783
16784 * fold-const.c (fold_negate_expr): Handle vectors.
16785 (fold_truth_not_expr): Make it static.
16786 (fold_invert_truthvalue): New static function.
16787 (invert_truthvalue_loc): Handle vectors. Do not call
16788 fold_truth_not_expr directly.
16789 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16790 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16791 (fold_comparison): Handle vector constants.
16792 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16793 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16794 * tree.h (fold_truth_not_expr): Remove declaration.
16795
fc21784d
JG		 167962013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16797
16798 * config/aarch64/aarch64-simd.md
16799 (aarch64_vcond_internal<mode>): Rename to...
16800 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16801 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16802 float modes. Clarify all iterator modes.
16803 (vcond<mode><mode>): Use new name for vcond expanders.
16804 (vcond<v_cmp_result><mode>): Likewise.
16805 (vcondu<mode><mode>: Likewise.
16806 * config/aarch64/iterators.md (VDQF_COND): New.
16807
d4c52634
MG		 168082013-05-14 Marc Glisse <marc.glisse@inria.fr>
16809
16810 PR bootstrap/57266
16811 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16812 variable for the shift amount. Check that we shift by non-negative
16813 amounts.
16814
2b261262
CLT		 168152013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16816
16817 PR target/42017
16818 * config/arm/arm.h (EPILOGUE_USES): Only return true
16819 for LR_REGNUM after epilogue_completed.
16820
12211b99 168212013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 16822
16823 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 16824 is error_mark_node.
6e022d7b 16825
56cf7859
RO		 168262013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16827
16828 PR target/57261
16829 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16830 and Solaris 11+/x86 with gld.
16831 * configure: Regenerate.
16832
75776c6d
JJ		 168332013-05-14 Jakub Jelinek <jakub@redhat.com>
16834
16835 * expmed.c (expand_shift_1): Canonicalize rotates by
16836 constant bitsize / 2 to bitsize - 1.
7f998021 16837 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 16838 case ROTATERT>: Likewise.
16839
16840 Revert:
16841 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16842
16843 * config/i386/i386.md (rotateinv): New code attr.
16844 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16845 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16846 roll $31, %eax, etc.
16847
df35498a
RB		 168482013-05-14 Richard Biener <rguenther@suse.de>
16849
16850 PR middle-end/57235
16851 * tree-eh.c (sink_clobbers): Give up for successors with
16852 multiple predecessors and no virtual uses.
16853
cc6e7ece
EB		 168542013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16855
16856 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16857 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16858
f778dd4d
JJ		 168592013-05-14 Jakub Jelinek <jakub@redhat.com>
16860
16861 PR middle-end/57251
16862 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16863 the case when both op0 and op1 have VOIDmode.
16864
bad4df9b
KP		 168652013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16866
16867 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16868 in multiply-accumulate mode.
16869
56f3e9ac
GW		 168702013-05-13 Guozhi Wei <carrot@google.com>
16871
16872 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16873
a3d7ab92
KT		 168742013-05-13 Kai Tietz <ktietz@redhat.com>
16875
16876 PR target/56975
16877 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 16878 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 16879 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 16880 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 16881 (ix86_expand_prologue): Likewise.
16882 (ix86_expand_split_stack_prologue): Likewise.
16883 (legitimate_pic_address_disp_p): Likewise.
16884 (legitimize_pic_address): Likewise.
16885 (legitimize_tls_address): Likewise.
16886 (legitimize_pe_coff_symbol): Likewise.
16887 (output_pic_addr_const): Likewise.
16888 (construct_plt_address): Likewise.
16889 (ix86_expand_call): Likewise.
16890 (x86_output_mi_thunk): Likewise.
16891 (x86_function_profiler): Likewise.
16892
c59b7e28
SN		 168932013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16894
16895 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16896 similar switch cases.
16897 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16898 (aarch64_simd_mov_to_<mode>low): Delete.
16899 (aarch64_simd_mov_to_<mode>high): Delete.
16900 (move_lo_quad_<mode>): Add w<-r alternative.
16901 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16902 (aarch64_simd_mov_from_*): Update type attribute.
16903 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16904 statement.
16905
5f28524a
JH		 169062013-05-13 Jan Hubicka <jh@suse.cz>
16907
16908 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16909 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 16910 ix86_expand_epilogue, emit_i387_cw_initialization,
16911 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16912 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 16913
ae6fa899
JJ		 169142013-05-13 Jakub Jelinek <jakub@redhat.com>
16915
16916 PR tree-optimization/45216
16917 PR tree-optimization/57157
16918 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16919 the (-Y) & (B - 1) variant if OP is |.
16920 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16921 return shifted. Use (-op1) & (prec - 1) as other_amount
16922 instead of prec - op1.
16923
4502fe8d
MJ		 169242013-05-13 Martin Jambor <mjambor@suse.cz>
16925
16926 PR middle-end/42371
16927 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
16928 (ipa_constant_data): New type.
16929 (ipa_jump_func): Use ipa_constant_data to hold information about
16930 constant jump functions.
16931 (ipa_get_jf_constant): Adjust to jump function type changes.
16932 (ipa_get_jf_constant_rdesc): New function.
16933 (ipa_param_descriptor): New field controlled_uses.
16934 (ipa_get_controlled_uses): New function.
16935 (ipa_set_controlled_uses): Likewise.
16936 * ipa-ref.h (ipa_find_reference): Declare.
16937 * ipa-prop.c (ipa_cst_ref_desc): New type.
16938 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
16939 changes.
16940 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
16941 New parameter cs. Adjust all callers.
16942 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
16943 (remove_described_reference): New function.
16944 (jfunc_rdesc_usable): Likewise.
16945 (try_make_edge_direct_simple_call): Decrement controlled use count,
16946 attempt to remove reference if it hits zero.
16947 (combine_controlled_uses_counters): New function.
16948 (propagate_controlled_uses): Likewise.
16949 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
16950 (ipa_edge_duplication_hook): Duplicate reference descriptions.
16951 (ipa_print_node_params): Print described use counter.
16952 (ipa_write_jump_function): Adjust to jump function type changes.
16953 (ipa_read_jump_function): New parameter CS, pass it to
16954 ipa_set_jf_constant. Adjust caller.
16955 (ipa_write_node_info): Stream controlled use count
16956 (ipa_read_node_info): Likewise.
16957 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
16958 asserting.
16959 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
16960 count. Remove cloning-added reference if it reaches zero.
16961 * ipa-ref.c (ipa_find_reference): New function.
16962
0864bfc2
GG		 169632013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16964
e7208ea3 16965 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 16966 alignment values for AMD BD and BT architectures.
16967
640bfeb2
MG		 169682013-05-13 Marc Glisse <marc.glisse@inria.fr>
16969
16970 * tree-vect-generic.c (uniform_vector_p): Move ...
16971 * tree.c (uniform_vector_p): ... here.
16972 * tree.h (uniform_vector_p): Declare it.
16973 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
16974 into a scalar.
16975
3a60f32b
JJ		 169762013-05-13 Jakub Jelinek <jakub@redhat.com>
16977
198fe1bf
JJ		 16978 PR tree-optimization/57230
16979 * tree-ssa-strlen.c (handle_char_store): Record length for
16980 array store from STRING_CST.
16981
3a60f32b
JJ		 16982 PR tree-optimization/57230
16983 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
16984 check.
16985
566be57c
JR		 169862013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
16987
16988 * config/epiphany/epiphany.c (epiphany_init): Check size of
16989 NUM_MODES_FOR_MODE_SWITCHING.
16990 (epiphany_expand_prologue):
16991 Remove CONFIG_REGNUM initial value handling code.
16992 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
16993 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 16994 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 16995 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
16996 Don't return 1 for FP_MODE_NONE.
16997 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
16998 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
16999 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17000 * config/epiphany/epiphany.md (save_config): New pattern.
17001
0f2c2331
UB		 170022013-05-12 Uros Bizjak <ubizjak@gmail.com>
17003
17004 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17005
5b3f93c7
UB		 170062013-05-10 Uros Bizjak <ubizjak@gmail.com>
17007
17008 * config/i386/i386.md (memory): Handle sseishft1.
17009 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17010 (*vec_extractv2di_1): Ditto.
17011
1f873f0e
VM		 170122013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17013
17014 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17015 saved registers.
17016
341427fa 170172013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 17018
17019 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17020 Add mthumb/march=armv7-a multilib.
17021 Add mthumb/march=armv7-r multilib.
17022 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17023
9e69bdde
RC		 170242013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17025
17026 * config/v850/t-rtems: Add more multilibs.
17027
9ff09a22
RB		 170282013-05-10 Richard Biener <rguenther@suse.de>
17029
17030 PR tree-optimization/57214
17031 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17032 not propagate from SSA names that occur in abnormal PHI nodes.
17033
a5e0cd1d
MG		 170342013-05-10 Marc Glisse <marc.glisse@inria.fr>
17035
17036 * stor-layout.c (element_precision): New function.
17037 * machmode.h (element_precision): Declare it.
17038 * tree.c (build_minus_one_cst): New function.
17039 (element_precision): Likewise.
17040 * tree.h (build_minus_one_cst): Declare new function.
17041 (element_precision): Likewise.
17042 * fold-const.c (operand_equal_p): Use element_precision.
17043 (fold_binary_loc): Handle vector types.
17044 * convert.c (convert_to_integer): Use element_precision.
17045 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17046 separately.
17047
cb2558bc
RS		 170482013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17049
17050 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17051 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17052 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17053 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17054 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17055 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17056 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17057 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17058 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17059 (Uuw8): New constraints.
17060 (Usb4): Move into alphabetical order.
17061 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17062 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17063 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17064 previously unnamed patterns.
17065 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17066 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17067 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17068 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17069 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17070 of specific lengths.
17071
cb3b8d33
JJ		 170722013-05-10 Jakub Jelinek <jakub@redhat.com>
17073
6f93c008
JJ		 17074 * config/i386/i386.md (rotateinv): New code attr.
17075 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17076 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17077 roll $31, %eax, etc.
17078
cb3b8d33
JJ		 17079 PR tree-optimization/45216
17080 PR tree-optimization/57157
17081 * tree-ssa-forwprop.c (simplify_rotate): New function.
17082 (ssa_forward_propagate_and_combine): Call it.
17083
afb119be
RB		 170842013-05-10 Richard Biener <rguenther@suse.de>
17085
17086 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17087 disable peeling when we version for aliasing.
17088 (vector_alignment_reachable_p): Honor explicit user alignment.
17089 (vect_supportable_dr_alignment): Likewise.
17090 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17091 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17092 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17093 then peeling to arrange for the cost-model check to come first.
17094
01ae4861
AM		 170952013-05-10 Alan Modra <amodra@gmail.com>
17096
17097 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17098 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17099 * configure: Regenerate.
17100
ebc9a431
AM		 171012013-05-10 Alan Modra <amodra@gmail.com>
17102
17103 PR target/55033
17104 * varasm.c (default_elf_select_section): Move !DECL_P check..
17105 (get_named_section): ..to here before calling get_section_name.
17106 Adjust assertion.
17107 (default_section_type_flags): Add DECL_P check.
17108 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17109 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17110
d4bca93c
JR		 171112013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17112
17113 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17114 When using gen_stack_adjust_str with a register offset, add a
17115 REG_FRAME_RELATED_EXPR note.
17116
60ca9a65
UB		 171172013-05-09 Uros Bizjak <ubizjak@gmail.com>
17118
17119 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17120 (*vec_extractv4si_zext_mem): Ditto.
17121 (*vec_extractv2di): Add 0->x and x->x alternatives.
17122 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17123 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17124
a3409c02
JM		 171252013-05-09 Jason Merrill <jason@redhat.com>
17126
0138d6b2
JM		 17127 N3639 C++1y VLA support
17128 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17129 DECL_VALUE_EXPR.
17130
a3409c02
JM		 17131 * tree.c (build_constructor_va): New.
17132 * tree.h: Declare it.
17133
66e6b990
MJ		 171342013-05-09 Martin Jambor <mjambor@suse.cz>
17135
17136 PR lto/57084
17137 * gimple-fold.c (canonicalize_constructor_val): Call
17138 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17139
64cfa6c0
JH		 171402013-05-09 Jan Hubicka <jh@suse.cz>
17141 Richard Biener <rguenther@suse.de>
17142
17143 PR lto/54095
17144 * symtab.c (symtab_make_decl_local): Do not add private names.
17145
c3167b00
JH		 171462013-05-09 Jan Hubicka <jh@suse.cz>
17147
17148 PR lto/54095
17149 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17150 (unlink_from_assembler_name_hash): Likewise.
17151 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17152 symtab_unregister_node, symtab_initialize_asm_name_hash,
17153 change_decl_assembler_name): Update.
17154
12dc6974
SN		 171552013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17156
17157 * config/aarch64/aarch64.md: New movtf split.
17158 (*movtf_aarch64): Update.
17159 (aarch64_movdi_tilow): Handle TF modes and rename to
17160 aarch64_movdi_<mode>low.
17161 (aarch64_movdi_tihigh): Handle TF modes and rename to
17162 aarch64_movdi_<mode>high
17163 (aarch64_movtihigh_di): Handle TF modes and rename to
17164 aarch64_mov<mode>high_di
17165 (aarch64_movtilow_di): Handle TF modes and rename to
17166 aarch64_mov<mode>low_di
17167 (aarch64_movtilow_tilow): Remove spurious whitespace.
17168 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17169 splits.
17170 (aarch64_print_operand): Update.
17171
227eb343
AM		 171722013-05-09 Alan Modra <amodra@gmail.com>
17173
17174 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17175 powerpc64le.
17176 * configure: Regenerate.
17177
0b013847
UB		 171782013-05-08 Uros Bizjak <ubizjak@gmail.com>
17179
17180 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17181 splitter preparation statements.
17182 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17183 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17184 adjust_address_nv.
17185
1dc3d6e9
BS		 171862013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17187
17188 * gimple-ssa-strength-reduction.c (count_candidates): Change
17189 return value to int.
17190 (analyze_candidates_and_replace): Change type of length to int.
17191
e61e7d28
UB		 171922013-05-08 Uros Bizjak <ubizjak@gmail.com>
17193
17194 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17195 (*vec_extract<mode>): Use VI12_128 mode iterator.
17196 (*vec_extract<mode>_mem): Ditto.
17197 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17198 attribute.
17199
4fbfcf44
DN		 172002013-05-08 Diego Novillo <dnovillo@google.com>
17201
17202 PR bootstrap/54659
17203
17204 Revert:
4fbfcf44
DN		 17205 2012-08-17 Diego Novillo <dnovillo@google.com>
17206
e61e7d28
UB		 17207 PR bootstrap/54281
17208 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17209 * config.in: Regenerate.
17210 * configure: Regenerate.
17211 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17212
702d8703
JH		 172132013-05-08 Jan Hubicka <jh@suse.cz>
17214
17215 PR lto/54095
17216 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17217 * cgraph.h (symtab_node_base): Add unique_name.
17218 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17219 input_overwrite_node, input_varpool_node): Stream unique_name.
17220 * cgraphclones.c (cgraph_create_virtual_clone,
17221 cgraph_function_versioning): Set unique_name.
17222 * ipa.c (function_and_variable_visibility): Set unique_name.
17223
8b28cf47
BS		 172242013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17225
17226 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17227 (alloc_cand_and_find_basis): Restrict conditional candidate
17228 processing to CAND_MULTs.
17229
e86074fd
JH		 172302013-05-08 Jan Hubicka <jh@suse.cz>
17231
17232 PR lto/54095
17233 lto-symtab.c (lto_symtab_symbol_p): New function.
17234 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17235 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17236 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17237 Skip static symbols.
17238
44398cbe
PC		 172392013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17240
17241 PR tree-optimization/57200
17242 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17243 Only call inform if the preceding warning_at returns true.
17244
f6bc1c4a
HS		 172452013-05-07 Han Shen <shenhan@google.com>
17246
17247 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17248 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17249 * common.opt (fstack-protector-strong): New option.
17250 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17251 * doc/invoke.texi (Optimization Options): Document
17252 "-fstack-protector-strong".
17253 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17254
4ffecb1f
SB		 172552013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17256
17257 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17258
5a107a0f
VM		 172592013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17260
17261 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17262 it is not lhs of the insn.
17263 (setup_reg_equiv): Remove insn having equiv memory which it is not
17264 lhs of the insn.
17265 * lra-constraints.c (process_address): Try to improve generation
17266 code for address base + disp.
17267 (lra_constraints): Make correct the code for checking insn setting
17268 up backward equivalence. Remove insn only if it is in the init
17269 insn list.
17270 * lra-eliminations.c (update_reg_eliminate): Change return value.
17271 (lra_eliminate): Use the result.
17272
3f5783ea
UB		 172732013-05-07 Uros Bizjak <ubizjak@gmail.com>
17274
17275 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17276 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17277 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17278 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17279 register target operands.
17280 (*vec_extractv8hi_sse2): New pattern.
17281 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17282 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17283 (*vec_extract<mode>_mem): New insn and split pattern.
17284
8a5800b8
CL		 172852013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17286
17287 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17288 (TARGET_ASAN_SHADOW_OFFSET): Define.
17289 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17290 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17291
7bf55a70
BS		 172922013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17293
17294 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17295 (incr_vec_index): Return -1 if increment not found.
17296 (create_add_on_incoming_edge): Assert if increment not found.
17297 (record_increment): Limit number of increments recorded.
17298 (all_phi_incrs_profitable): Return false if an increment not found.
17299 (replace_profitable_candidates): Don't process increments that were
17300 not recorded.
17301 (analyze_candidates_and_replace): Limit size of incr_vec.
17302
3f8825c0
RB		 173032013-05-07 Richard Biener <rguenther@suse.de>
17304
17305 * calls.c (special_function_p): setjmp-like functions are leaf.
17306 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17307 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17308
fd4842cd
SN		 173092013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17310
17311 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17312 (aarch64_simd_mov<mode>): New expander.
17313 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17314 (aarch64_simd_mov_to_<mode>high): Likewise.
17315 (aarch64_simd_mov_from_<mode>low): Likewise.
17316 (aarch64_simd_mov_from_<mode>high): Likewise.
17317 (aarch64_dup_lane<mode>): Update.
17318 (aarch64_dup_lanedi): New instruction pattern.
17319 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17320 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17321
a7a7d10e
BS		 173222013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17323
17324 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17325 (replace_mult_candidate): Remove unnecessary argument; remove
17326 unnecessary parameter from call to introduce_cast_before_cand.
17327 (replace_unconditional_candidate): Remove unnecessary parameter
17328 from call to replace_mult_candidate.
17329 (replace_conditional_candidate): Likewise.
17330 (insert_initializers): Use make_temp_ssa_name.
17331 (introduce_cast_before_cand): Remove unnecessary argument; use
17332 make_temp_ssa_name.
17333 (replace_one_candidate): Remove unnecessary argument; remove
17334 unnecessary parameter from calls to introduce_cast_before_cand.
17335 (replace_profitable_candidates): Remove unnecessary parameters
17336 from calls to replace_one_candidate.
17337
29105868
BS		 173382013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17339
17340 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17341 phi def as possibly hiding a basis for a CAND_ADD whose operands
17342 have been commuted in the analysis.
17343 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17344
4095f9fa
N		 173452013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17346
17347 * config/aarch64/aarch64.md
17348 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17349 shift value between 0-4.
17350
39e843e8
RB		 173512013-05-07 Richard Biener <rguenther@suse.de>
17352
17353 * double-int.h (rshift): New overload.
17354 * double-int.c (rshift): New function.
17355 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17356 (create_reference_ops_from_ref): Remove.
17357 (vn_reference_insert): Use shared ops for constructing the
17358 reference and copy it.
17359
0a1a83cb
RB		 173602013-05-07 Richard Biener <rguenther@suse.de>
17361
17362 PR middle-end/57190
17363 * tree-eh.c (sink_clobbers): Properly propagate
17364 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17365
43303d6f
JJ		 173662013-05-07 Jakub Jelinek <jakub@redhat.com>
17367
ba7e83f8
JJ		 17368 PR tree-optimization/57149
17369 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17370 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17371 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17372 uninit_undefined_value_p instead of ssa_undefined_value_p.
17373
43303d6f
JJ		 17374 PR debug/57184
17375 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17376 for modifier == EXPAND_INITIALIZER.
17377
14523c25
AB		 173782013-05-07 Anton Blanchard <anton@samba.org>
17379
17380 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17381 for powerpc64 little endian.
17382 * configure: Regenerate.
17383
cb7c8be9
GS		 173842013-05-06 Graham Stott <grahams@btinternet.com>
17385
17386 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17387 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17388 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17389 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17390
173912013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 17392
17393 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17394 codes which allow non-lvalues.
17395
9a0ee7b0
MG		 173962013-05-06 Marc Glisse <marc.glisse@inria.fr>
17397
17398 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17399 components are all 1s.
17400 (integer_minus_onep): New function.
17401 * tree.h (integer_minus_onep): Declare it.
17402 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17403 integer_minus_onep instead of integer_all_onesp.
17404
f2c17ea9
OE		 174052013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17406
17407 PR target/52933
17408 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17409 variations of these patterns.
17410
f75e6a51
UB		 174112013-05-06 Uros Bizjak <ubizjak@gmail.com>
17412
17413 * config/i386/i386.md (isa): Add x64_sse4 member.
17414 (enabled): Handle x64_sse4.
17415 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17416 instruction for 64bit SSE4_1 targets. Update insn attributes.
17417 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17418 instruction for SSE4_1 targets. Update insn attributes.
17419 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17420 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17421 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17422 const_1 selector.
17423 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17424 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17425 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17426
a986d468
OE		 174272013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17428
17429 PR target/57108
17430 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17431
382522cb
MK		 174322013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17433
17434 * final.c (do_assembler_dialects): Don't handle curly braces and
17435 vertical bar escaped by % as dialect delimiters.
17436 (output_asm_insn): Print curly braces and vertical bar if escaped
17437 by % and ASSEMBLER_DIALECT defined.
17438 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17439 * doc/tm.texi: Regenerated.
17440
fb0d5c60
SB		 174412013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17442
fb0d5c60
SB		 17443 * config/mips/mips.c: Include tree-pass.h.
17444 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17445 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17446 (pass_mips_machine_reorg2): New machine specific pass.
17447 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17448 (mips_option_override): Register the new pass.
17449 * rtl.h (cleanup_barriers): Remove prototype.
17450 (dbr_schedule): Likewise.
17451 * jump.c (cleanup_barriers): Make static.
17452 * reorg.c (dbr_schedule): Likewise.
17453
aa06a978
RB		 174542013-05-06 Richard Biener <rguenther@suse.de>
17455
17456 PR tree-optimization/57185
17457 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17458 SSA names properly.
17459
0b953bec
UB		 174602013-05-06 Uros Bizjak <ubizjak@gmail.com>
17461
17462 PR target/57106
17463 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17464 between "first_arg" and "insn", not "last" and "insn".
17465
28708525
WS		 174662013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17467
17468 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17469 (find_candidates_in_block): Re-enable slsr_process_phi.
17470 (create_phi_basis): Fix double counting of candidate adjustment.
17471
0107dca2
RB		 174722013-05-06 Richard Biener <rguenther@suse.de>
17473
17474 PR middle-end/57147
17475 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17476 the edge is also fallthru, preserve it and just clear the
17477 abnormal flag.
17478 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17479 also complex, preserve that and just clear the fallthru flag.
17480 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17481 update virtual operands.
17482
470d4d13
AM		 174832013-05-06 Alan Modra <amodra@gmail.com>
17484
17485 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17486 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17487 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17488 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17489 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17490 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17491
8f1dbf8d
AM		 174922013-05-06 Alan Modra <amodra@gmail.com>
17493
17494 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17495 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17496 (DEFAULT_ASM_ENDIAN): Define.
17497 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17498 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17499 Update -K PIC clause from sysv4.h.
17500 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17501 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17502
54890767
AM		 175032013-05-06 Alan Modra <amodra@gmail.com>
17504
17505 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17506 twice for little-endian.
17507 (ashrdi3_no_power, ashrdi3): Support little-endian.
17508
2353515d
OE		 175092013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17510
17511 PR target/55303
17512 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17513 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17514 related expanders.
17515 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17516 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17517 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17518 New predicates.
17519
33e67557
SB		 175202013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17521 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17522
17523 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17524 * config/pa/pa.opt: Make mbig-switch a no-op.
17525 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17526 (CASE_VECTOR_MODE): Always return SImode.
17527 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17528 for the !TARGET_BIG_SWITCH case.
17529 * config/pa/pa-linux.h: Likewise.
17530 * config/pa/pa-openbsd.h: Likewise.
17531 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17532 * config/pa/pa.md (short_jump): Remove define_insn.
17533 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17534 (casesi0): Remove define_insn.
17535 (type): Remove btable_branch.
17536 (pa_combine_type): Likewise.
17537 (in_nullified_branch_delay): Likewise.
17538 (in_call_delay): Likewise.
17539 (define_delay): Likewise.
17540 (define_insn_reservation "Z3"): Likewise.
17541 (define_insn_reservation "Z4"): Likewise.
17542 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17543 (pa_adjust_insn_length): Remove adjustment for btable branches.
17544 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17545 and mno-big-switch
17546
3095685e
UB		 175472013-05-05 Uros Bizjak <ubizjak@gmail.com>
17548
17549 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17550 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17551 Add m->r,x alternatives.
17552 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17553 splitters using SWI48x mode iterator.
17554 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17555 TARGET_64BIT. Add m->x alternative.
17556 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17557 Add o->x alternative. Enable for TARGET_SSE.
17558 (sse_storeq): Remove expander.
17559 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17560 with memory input operand.
17561 (*vec_extractv2di_1 splitter): New.
17562 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17563 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17564
4b36ae28
SB		 175652013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17566
17567 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17568 (INT_LOWPART): Delete.
17569 (extract_MB): Adjust.
17570 (extract_ME): Adjust.
17571 (print_operand): Adjust.
17572
da226db2
SB		 175732013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17574
17575 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17576 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17577 (reg_or_logical_cint_operand, easy_fp_constant,
17578 logical_const_operand): Delete "CONST_DOUBLE" case.
17579 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17580 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 17581 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 17582 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17583 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17584 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17585 test.
17586 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17587 CONST_DOUBLE DImode/VOIDmode case.
17588 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17589 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17590 CONST_DOUBLE VOIDmode case.
17591 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17592 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17593 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17594 Delete CONST_DOUBLE case.
17595 (splitters for mov FMOVE64 const_double): Delete
17596 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17597 "HOST_BITS_PER_WIDE_INT >= 64" test.
17598 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17599 case.
17600 (mov DI const_double): Delete.
17601
40de22d6
JJ		 176022013-05-04 Jakub Jelinek <jakub@redhat.com>
17603
17604 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17605 on op shows all bits zero in mode of a lowpart subreg, return zero.
17606
5ec6aff2
MM		 176072013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17608
17609 PR target/57150
17610 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17611 to save TFmode registers and DImode to save TImode registers for
17612 caller save operations.
17613 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17614 mark being partially clobbered since they only use the first
17615 double word.
17616
17617 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17618 and TDmode only use the upper 64-bits of each VSX register.
17619
2cefad90
BS		 176202013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17621
17622 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17623 (find_candidates_in_block): Disable slsr_process_phi.
17624
d6d7eee1
GW		 176252013-05-03 Guozhi Wei <carrot@google.com>
17626
17627 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 17628 constructor to ...
17629 (build_init_ctor): ... here.
d6d7eee1 17630
9b92d12b
BS		 176312013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17632
17633 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17634 (slsr_cand_d): Redefine def_phi.
17635 (stride_status, phi_adjust_status, count_phis_status): New enums.
17636 (find_phi_def): New.
17637 (find_basis_for_base_expr): New.
17638 (find_basis_for_candidate): Handle hidden bases.
17639 (alloc_cand_and_find_basis): Handle phi candidates.
17640 (slsr_process_phi): New.
17641 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17642 (create_mul_imm_cand): Likewise.
17643 (create_add_ssa_cand): Exclude phi base candidates.
17644 (create_add_imm_cand): Likewise.
17645 (slsr_process_cast): Likewise.
17646 (slsr_process_copy): Likewise.
17647 (find_candidates_in_block): Handle phi candidates.
17648 (dump_candidate): Likewise.
17649 (unconditional_cands): Delete.
17650 (unconditional_cands_with_known_stride_p): Delete.
17651 (phi_dependent_cand_p): New.
17652 (cand_increment): Handle phi-dependent candidates.
17653 (replace_dependent): Delete.
17654 (replace_mult_candidate): New.
17655 (replace_unconditional_candidate): New.
17656 (incr_vec_index): Move to avoid forward reference.
17657 (create_add_on_incoming_edge): New.
17658 (create_phi_basis): New.
17659 (replace_dependents): Delete.
17660 (replace_conditional_candidate): New.
17661 (phi_add_costs): New.
17662 (replace_uncond_cands_and_profitable_phis): New.
17663 (record_increment): Handle phi adjustments.
17664 (record_phi_increments): New.
17665 (record_increments): Handle phi adjustments.
17666 (phi_incr_cost): New.
17667 (lowest_cost_path): Handle phis.
17668 (total_savings): Likewise.
17669 (analyze_increments): Likewise.
17670 (ncd_with_phi): New.
17671 (ncd_of_cand_and_phis): New.
17672 (nearest_common_dominator_for_cands): Handle phi increments.
17673 (all_phi_incrs_profitable): New.
17674 (replace_profitable_candidates): Handle phi-dependent candidates.
17675 (analyze_candidates_and_replace): Likewise.
17676
68f073d4
TJ		 176772013-05-03 Teresa Johnson <tejohnson@google.com>
17678
17679 PR bootstrap/57154
17680 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17681 do not exceed REG_BR_PROB_BASE.
17682
a4ee7cb9
JL		 176832013-05-03 Jeff Law <law@redhat.com>
17684
ade67f70 17685 PR tree-optimization/57144
a4ee7cb9
JL		 17686 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17687 operand of the condition will bit into the new type when eliminating
17688 a cast feeding a condition.
17689
47954c4e
JJ		 176902013-05-03 Jakub Jelinek <jakub@redhat.com>
17691
17692 PR rtl-optimization/57130
3095685e
UB		 17693 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17694 of COMPARE as in_code to the recursive call if needed.
47954c4e 17695
3c21604f
UB		 176962013-05-03 Uros Bizjak <ubizjak@gmail.com>
17697
17698 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17699 (enabled): Handle new members.
17700 * config/i386/sse.md (*vec_concatv2si): Merge from
17701 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17702 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17703
12211b99 177042013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 17705
17706 PR tree-optimization/57027
17707 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17708 for fnms opportunity, check we got the prerequisite kind
17709 of tree / gimple before using accessor functions.
17710
07bfc9ec
RB		 177112013-05-03 Richard Biener <rguenther@suse.de>
17712
17713 * double-int.h (lshift): New overload without precision
17714 and arith argument.
17715 (operator *=, operator +=, operator -=): Move ...
17716 * double-int.c (operator *=, operator +=, operator -=): ... here
17717 and implement more efficiently.
17718 (mul_double_with_sign): Remove.
17719 (lshift_double): Adjust to take unsinged shift argument, push
17720 dispatching code to callers.
17721 (mul_double_wide_with_sign): Add early out for callers that
17722 are not interested in high parts or overflow.
17723 (lshift): New function.
17724 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17725 dispatch code here.
17726 (lrotate, rrotate): Use logical shifts.
17727 * expr.c (get_inner_reference): Use lshift.
17728 * fixed-value.c (do_fixed_divide): Likewise.
17729 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17730 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17731 (indirect_refs_may_alias_p): Likewise.
17732 (stmt_kills_ref_p_1): Likewise.
17733
7769bb64
VP		 177342013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17735
17736 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17737
1d0c8e5c
VP		 177382013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17739
17740 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17741 scalar form of FABD instruction.
17742
f15feaf9
VM		 177432013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17744
17745 * lra-constraints.c (process_alt_operands): Add checking alt
17746 number to choose the best alternative.
17747
d90e76d4
RB		 177482013-05-02 Richard Biener <rguenther@suse.de>
17749
17750 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17751 bitmap and its handling.
17752 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17753
9f8e7a96
RB		 177542013-05-02 Richard Biener <rguenther@suse.de>
17755
17756 PR middle-end/57140
17757 * tree-inline.c (copy_loops): Properly handle removed loops.
17758 (copy_cfg_body): Mark destination loops for fixup if source
17759 loops needed fixup.
17760
f3a81b39
GY		 177612013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17762
17763 PR target/56732
17764 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17765 generating simple_return for naked functions.
17766
7b920a9a
MJ		 177672013-05-02 Martin Jambor <mjambor@suse.cz>
17768
17769 PR middle-end/56988
17770 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17771 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17772 flags match.
17773 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17774 ipa_agg_replacement_value structures.
17775 (known_aggs_to_agg_replacement_list): Likewise.
17776 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17777 (read_agg_replacement_chain): Likewise.
17778 (ipcp_transform_function): Also check that by_ref flags match.
17779
2c41c19d
RB		 177802013-05-02 Richard Biener <rguenther@suse.de>
17781
17782 * graphds.h (struct graph): Add obstack member.
17783 * graphds.c (new_graph): Initialize obstack and allocate
17784 vertices from it.
17785 (add_edge): Allocate edge from the obstack.
3c21604f 17786 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 17787
8b47039c
TJ		 177882013-05-02 Teresa Johnson <tejohnson@google.com>
17789
17790 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17791 divides.
17792 * cfg.c (update_bb_profile_for_threading): Ditto.
17793 * tree-inline.c (copy_bb): Ditto.
17794 (copy_edges_for_bb): Ditto.
17795 (initialize_cfun): Ditto.
17796 (copy_cfg_body): Ditto.
17797 (expand_call_inline): Ditto.
17798 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17799 (estimate_node_size_and_time): Ditto.
17800 (inline_merge_summary): Ditto.
17801 * cgraphclones.c (cgraph_clone_edge): Ditto.
17802 (cgraph_clone_node): Ditto.
17803 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17804 (compute_trg_info): Ditto.
17805
da65928c
IB		 178062013-05-02 Ian Bolton <ian.bolton@arm.com>
17807
17808 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17809 S reg when fp attribute set.
17810 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17811
add3c965
IB		 178122013-05-02 Ian Bolton <ian.bolton@arm.com>
17813
17814 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17815 New pattern.
17816 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17817 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17818 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17819
fdd43ac4
RB		 178202013-05-02 Richard Biener <rguenther@suse.de>
17821
17822 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17823 (struct instantiate_cache_entry): New type.
17824 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17825 (struct instantiate_cache_type): New type.
17826 (set_instantiated_value, get_instantiated_value): Remove.
17827 (get_instantiated_value_entry): New function.
17828 (instantiate_scev_name): Use the new cache and adjust.
17829 (instantiate_scev_poly): Adjust.
17830 (instantiate_scev_binary): Likewise.
17831 (instantiate_array_ref): Likewise.
17832 (instantiate_scev_convert): Likewise.
17833 (instantiate_scev_not): Likewise.
17834 (instantiate_scev_3): Likewise.
17835 (instantiate_scev_2): Likewise.
17836 (instantiate_scev_r): Likewise.
17837 (instantiate_scev): Likewise.
17838 (resolve_mixers): Likewise.
17839
36ff9dfb
VM		 178402013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17841
17842 PR target/57091
17843 * lra-constraints.c (best_small_class_operands_num): Remove.
17844 (process_alt_operands): Remove small_class_operands_num. Take
17845 small classes operands into losers and only if the operand is not
17846 matched. Modify debugging output.
17847 (curr_insn_transform): Remove best_small_class_operands_num.
17848 Print insn name.
17849
36054fab
JG		 178502013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17851
17852 * config/aarch64/aarch64-builtins.c
17853 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17854 * config/aarch64/aarch64-simd-builtins.def
17855 (reduc_splus_): Add new modes.
17856 (reduc_uplus_): New.
17857 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17858 (reduc_uplus_v4sf): Likewise.
17859 (reduc_splus_v4sf): Likewise.
17860 (aarch64_addv<mode>): Likewise.
17861 (reduc_uplus_<mode>): Likewise.
17862 (reduc_splus_<mode>): Likewise.
17863 (aarch64_addvv2di): Likewise.
17864 (reduc_uplus_v2di): Likewise.
17865 (reduc_splus_v2di): Likewise.
17866 (aarch64_addvv2si): Likewise.
17867 (reduc_uplus_v2si): Likewise.
17868 (reduc_splus_v2si): Likewise.
17869 (reduc_<sur>plus_<mode>): New.
17870 (reduc_<sur>plus_v2di): Likewise.
17871 (reduc_<sur>plus_v2si): Likewise.
17872 (reduc_<sur>plus_v4sf): Likewise.
17873 (aarch64_addpv4sf): Likewise.
17874 * config/aarch64/arm_neon.h
17875 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17876 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17877 add UNSPEC_SADDV, UNSPEC_UADDV.
17878 (SUADDV): New.
17879 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17880
6dce23a8
JG		 178812013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17882
17883 * config/aarch64/arm_neon.h
17884 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17885
1598945b
JG		 178862013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17887
17888 * config/aarch64/aarch64-builtins
17889 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17890
998eaf97
JG		 178912013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17892
17893 * config/aarch64/aarch64-simd-builtins.def
17894 (reduc_smax_): New.
17895 (reduc_smin_): Likewise.
17896 (reduc_umax_): Likewise.
17897 (reduc_umin_): Likewise.
17898 (reduc_smax_nan_): Likewise.
17899 (reduc_smin_nan_): Likewise.
17900 (fmax): Remove.
17901 (fmin): Likewise.
17902 (smax): Update for V2SF, V4SF and V2DF modes.
17903 (smin): Likewise.
17904 (smax_nan): New.
17905 (smin_nan): Likewise.
17906 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17907 (<su><maxmin><mode>3): ...This, refactor.
17908 (s<maxmin><mode>3): New.
17909 (<maxmin_uns><mode>3): Likewise.
17910 (reduc_<maxmin_uns>_<mode>): Refactor.
17911 (reduc_<maxmin_uns>_v4sf): Likewise.
17912 (reduc_<maxmin_uns>_v2si): Likewise.
17913 (aarch64_<fmaxmin><mode>: Remove.
17914 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17915 new builtin names.
17916 (vmin<q>_f<32,64>): Likewise.
17917 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17918 (FMAXMIN): New.
17919 (su): Add mappings for smax, smin, umax, umin.
17920 (maxmin): New.
17921 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17922 (FMAXMIN): Rename as...
17923 (FMAXMIN_UNS): ...This.
17924 (maxminv): Remove.
17925 (fmaxminv): Likewise.
17926 (fmaxmin): Likewise.
17927 (maxmin_uns): New.
17928 (maxmin_uns_op): Likewise.
17929
bd11644e
JG		 179302013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17931
17932 * config/aarch64/arm_neon.h
17933 (vac<ge, gt><sd>_f<32, 64>): Rename to...
17934 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
17935 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
17936
75dd5ace
JG		 179372013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17938
17939 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
17940 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
17941
7c19979f
JG		 179422013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17943
17944 * config/aarch64/aarch64-simd.md
17945 (vcond<mode>_internal): Handle special cases for constant masks.
17946 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
17947 (vcondu<mode><mode>): Likewise.
17948 (vcond<v_cmp_result><mode>): New.
17949
bb60efd9
JG		 179502013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17951
17952 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
17953 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
17954 * config/aarch64/aarch64-simd-builtins.def
17955 (cmeq): Update to BUILTIN_VALLDI.
17956 (cmgt): Likewise.
17957 (cmge): Likewise.
17958 (cmle): Likewise.
17959 (cmlt): Likewise.
17960 * config/aarch64/arm_neon.h
17961 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
17962 to builtins or C as appropriate.
17963
889b9412
JG		 179642013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17965
17966 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
17967 (cmgeu): ...This.
17968 (cmhi): Rename to...
17969 (cmgtu): ...This.
17970 * config/aarch64/aarch64-simd.md
17971 (simd_mode): Add SF.
17972 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
17973 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
17974 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
17975 (cstore<mode>_neg): ...This.
17976 * config/aarch64/iterators.md
17977 (VALLF): new.
17978 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
17979 (COMPARISONS): New.
17980 (UCOMPARISONS): Likewise.
17981 (optab): Add missing comparisons.
17982 (n_optab): New.
17983 (cmp_1): Likewise.
17984 (cmp_2): Likewise.
17985 (CMP): Likewise.
17986 (cmp): Remove.
17987 (VCMP_S): Likewise.
17988 (VCMP_U): Likewise.
17989 (V_cmp_result): Add DF, SF modes.
17990 (v_cmp_result): Likewise.
17991 (v): Likewise.
17992 (vmtype): Likewise.
17993 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
17994
0a7dbb76
GY		 179952013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
17996
17997 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
17998 define_insn to define_insn_and_split.
17999 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18000 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18001 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18002 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18003 (thumb2_negscc): Likewise.
18004
fb614ca6
GY		 180052013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18006
18007 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18008
9e64a0bf
GY		 180092013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18010
18011 * config/arm/thumb2.md: Remove trailing whitespaces.
18012
d6b28156
RS		 180132013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18014
18015 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18016 Use gen_int_mode rather than GEN_INT.
18017
f91674c3
L		 180182013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18019
b0dec607 18020 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 18021 debug_gimple_stmt.
18022
3551257c
RB		 180232013-04-30 Richard Biener <rguenther@suse.de>
18024
18025 PR middle-end/57122
3c21604f 18026 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 18027
4c1aff1c
RB		 180282013-04-30 Richard Biener <rguenther@suse.de>
18029
18030 PR middle-end/57107
18031 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18032
54b8379a
AB		 180332013-04-30 Andrey Belevantsev <abel@ispras.ru>
18034
18035 PR rtl-optimization/56957
18036 PR rtl-optimization/57105
54b8379a
AB		 18037 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18038 variable. Use just INSN_UID for determining whether an insn
18039 should be only disconnected from the insn stream.
18040 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18041
abc27962
JJ		 180422013-04-30 Jakub Jelinek <jakub@redhat.com>
18043
18044 PR tree-optimization/57104
18045 * tsan.c (instrument_expr): Don't instrument accesses to
18046 DECL_HARD_REGISTER VAR_DECLs.
18047
0fc822d0
RB		 180482013-04-30 Richard Biener <rguenther@suse.de>
18049
18050 * function.h (loops_for_fn): New inline function.
18051 (set_loops_for_fn): Likewise.
18052 * cfgloop.h (place_new_loop): Add struct function parameter.
18053 (get_loop): Likewise.
18054 (get_loops): Likewise.
18055 (number_of_loops): Likewise.
18056 (fel_next): Adjust.
18057 (fel_init): Likewise.
18058 * cfg.c (get_loop_copy): Adjust.
18059 * cfgloop.c (flow_loops_dump): Likewise.
18060 (record_loop_exits): Likewise.
18061 (verify_loop_structure): Likewise.
18062 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18063 (estimate_reg_pressure_cost): Likewise.
18064 (mark_loop_exit_edges): Likewise.
18065 * cfgloopmanip.c (place_new_loop): Likewise.
18066 (add_loop): Likewise.
18067 (duplicate_loop): Likewise.
18068 * graph.c (draw_cfg_nodes): Likewise.
18069 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18070 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18071 (extract_affine_chrec): Likewise.
18072 (build_scop_iteration_domain): Likewise.
18073 * graphite.c (graphite_initialize): Likewise.
18074 * ira-build.c (create_loop_tree_nodes): Likewise.
18075 (more_one_region_p): Likewise.
18076 (rebuild_regno_allocno_maps): Likewise.
18077 (mark_loops_for_removal): Likewise.
18078 (mark_all_loops_for_removal): Likewise.
18079 (remove_unnecessary_regions): Likewise.
18080 (ira_build): Likewise.
18081 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18082 * loop-init.c (fix_loop_structure): Likewise.
18083 (gate_rtl_move_loop_invariants): Likewise.
18084 (gate_rtl_unswitch): Likewise.
18085 (gate_rtl_unroll_and_peel_loops): Likewise.
18086 (rtl_doloop): Likewise.
18087 * lto-streamer-in.c (input_cfg): Likewise.
18088 * lto-streamer-out.c (output_cfg): Likewise.
18089 * modulo-sched.c (sms_schedule): Likewise.
18090 * predict.c (tree_estimate_probability): Likewise.
18091 (tree_estimate_probability_driver): Likewise.
18092 (estimate_loops): Likewise.
18093 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18094 (move_sese_region_to_fn): Likewise.
18095 (debug_loop_num): Likewise.
18096 * tree-chrec.c (chrec_evaluate): Likewise.
18097 (hide_evolution_in_other_loops_than_loop): Likewise.
18098 (chrec_component_in_loop_num): Likewise.
18099 (reset_evolution_in_loop): Likewise.
18100 (evolution_function_is_invariant_rec_p): Likewise.
18101 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18102 * tree-inline.c (copy_loops): Likewise.
18103 (copy_cfg_body): Likewise.
18104 (tree_function_versioning): Likewise.
18105 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18106 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18107 Likewise.
18108 (add_to_evolution_1): Likewise.
18109 (scev_const_prop): Likewise.
18110 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18111 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18112 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18113 (tree_ssa_lim_initialize): Likewise.
18114 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18115 (verify_loop_closed_ssa): Likewise.
18116 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18117 (tree_ssa_loop_im): Likewise.
18118 (tree_ssa_loop_unswitch): Likewise.
18119 (tree_vectorize): Likewise.
18120 (check_data_deps): Likewise.
18121 (tree_ssa_loop_ivcanon): Likewise.
18122 (tree_ssa_loop_bounds): Likewise.
18123 (tree_complete_unroll): Likewise.
18124 (tree_complete_unroll_inner): Likewise.
18125 (tree_parallelize_loops): Likewise.
18126 (tree_ssa_loop_prefetch): Likewise.
18127 (tree_ssa_loop_ivopts): Likewise.
18128 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18129 * tree-vectorizer.c (vectorize_loops): Likewise.
18130
37953bd3
MF		 181312013-04-29 Mike Frysinger <vapier@gentoo.org>
18132
18133 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18134 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18135 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18136 with EABI_LINK_SPEC.
18137
f9ed28db
UB		 181382013-04-29 Uros Bizjak <ubizjak@gmail.com>
18139
18140 PR target/44578
18141 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18142 alternative.
18143
deca73f5
VM		 181442013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18145
18146 PR target/57097
37953bd3 18147 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 18148 using memory for pseudos. Print cost dump for alternatives.
18149 Modify cost values for conflicts with early clobbers.
18150 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18151
fdca7d03
UB		 181522013-04-29 Uros Bizjak <ubizjak@gmail.com>
18153
18154 PR target/57098
18155 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18156
0b064172
IB		 181572013-04-29 Ian Bolton <ian.bolton@arm.com>
18158
18159 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18160 from/to S register.
18161 (movdi_aarch64): Support LDR/STR from/to D register.
18162
473cec55
IB		 181632013-04-29 Ian Bolton <ian.bolton@arm.com>
18164
18165 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18166 or higher by default.
18167
a6f30e66
RB		 181682013-04-29 Richard Biener <rguenther@suse.de>
18169
18170 PR middle-end/57075
18171 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18172 even if not adding abnormal edges for calls that can make
18173 abnormal gotos.
18174
0c2b2040
RB		 181752013-04-29 Richard Biener <rguenther@suse.de>
18176
18177 PR middle-end/57103
18178 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18179 TREE_BLOCK.
18180 (move_stmt_r): Remove redundant checking.
18181
f41f80f9
TJ		 181822013-04-29 Teresa Johnson <tejohnson@google.com>
18183
18184 PR bootstrap/57077
18185 * basic-block.h (apply_scale): New function.
18186 (apply_probability): Use apply_scale.
18187 * gimple-streamer-in.c (input_bb): Ditto.
18188 * lto-streamer-in.c (input_cfg): Ditto.
18189 * lto-cgraph.c (merge_profile_summaries): Ditto.
18190 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18191 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 18192 (copy_edges_for_bb): Ditto.
18193 (copy_cfg_body): Ditto.
18194
315bbd2e
TV		 181952013-04-29 Tom de Vries <tom@codesourcery.com>
18196
18197 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18198 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18199 (tail_merge_optimize): Handle current_loops == NULL.
18200
ebbd90d8
JL		 182012013-04-26 Jeff Law <law@redhat.com>
18202
18203 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18204 (simplify_cond_using_ranges): Generalize code to simplify
18205 COND_EXPRs where one argument is a constant and the other
18206 is an SSA_NAME created by an integral type conversion.
18207
8b9b57eb
KT		 182082013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18209
18210 * config/arm/arm.md (store_minmaxsi): Use only when
18211 optimize_insn_for_size_p.
18212
9498e5dc
CB		 182132013-04-29 Christian Bruel <christian.bruel@st.com>
18214
18215 PR target/57108
18216 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18217
6093bc06
RB		 182182013-04-29 Richard Biener <rguenther@suse.de>
18219
18220 PR middle-end/57089
fdca7d03
UB		 18221 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18222 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 18223 (expand_omp_for_generic): Properly add loops.
18224 (expand_omp_for_static_nochunk): Likewise.
18225 (expand_omp_for_static_chunk): Likewise.
18226 (expand_omp_for): For the degenerate case fixup loops.
18227 (expand_omp_sections): Fix default bb placement in loops.
18228 (expand_omp_atomic_pipeline): Properly add loops.
18229
84aacbfd
KT		 182302013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18231
18232 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18233
8dee4479
TV		 182342013-04-29 Tom de Vries <tom@codesourcery.com>
18235
18236 * tree-ssa-tail-merge.c: Update header comment.
18237
47934dc4
JG		 182382013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18239
18240 * config/aarch64/arm_neon.h
18241 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18242 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18243 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18244 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18245 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18246 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18247 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18248 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18249
384be29f
JG		 182502013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18251
18252 * config/aarch64/aarch64-simd.md
18253 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18254 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18255 fix_trunc, fixuns_trunc.
18256 (ftrunc<VDQF:mode>2): New.
18257 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18258 (fix_trunc_optab): New.
18259
0386b123
JG		 182602013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18261
18262 * config/aarch64/aarch64-builtins.c
18263 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18264 iceilf, lround, iroundf.
18265
00fcb892
UB		 182662013-04-29 Uros Bizjak <ubizjak@gmail.com>
18267
18268 PR target/54349
18269 * config/i386/i386.h (enum ix86_tune_indices)
18270 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18271 New, split from X86_TUNE_INTER_UNIT_MOVES.
18272 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18273 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18274 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18275 (TARGET_INTER_UNIT_MOVES): Remove.
18276 * config/i386/i386.c (initial_ix86_tune_features): Update.
18277 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18278 (ix86_expand_convert_uns_didf_sse): Use
18279 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18280 (ix86_expand_vector_init_one_nonzero): Ditto.
18281 (ix86_expand_vector_init_interleave): Ditto.
18282 (inline_secondary_memory_needed): Return true for moves from SSE class
18283 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18284 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18285 * config/i386/constraints.md (Yi, Ym): Depend on
18286 TARGET_INTER_UNIT_MOVES_TO_VEC.
18287 (Yj, Yn): New constraints.
18288 * config/i386/i386.md (*movdi_internal): Change constraints of
18289 operand 1 from Yi to Yj and from Ym to Yn.
18290 (*movsi_internal): Ditto.
18291 (*movdf_internal): Ditto.
18292 (*movsf_internal): Ditto.
18293 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18294 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18295 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18296 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18297 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18298 * config/i386/sse.md (movdi_to_sse): Ditto.
18299 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18300 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18301 TARGET_INTER_UNIT_MOVES.
18302 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18303 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18304 instead of TARGET_INTER_UNIT_MOVES.
18305 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18306 operand 1 from Yi to Yj and from Ym to Yn.
18307
4c871069
JG		 183082013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18309
18310 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18311 (float_truncate_hi_): Likewise.
18312 (float_extend_lo_): Likewise.
18313 (float_truncate_lo_): Likewise.
18314 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18315 (aarch64_float_extend_lo_v2df): Likewise.
18316 (vec_unpacks_hi_v4sf): Likewise.
18317 (aarch64_float_truncate_lo_v2sf): Likewise.
18318 (aarch64_float_truncate_hi_v4sf): Likewise.
18319 (vec_pack_trunc_v2df): Likewise.
18320 (vec_pack_trunc_df): Likewise.
18321
1709ff9b
JG		 183222013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18323
18324 * config/aarch64/aarch64-builtins.c
18325 (aarch64_fold_builtin): Fold float conversions.
18326 * config/aarch64/aarch64-simd-builtins.def
18327 (floatv2si, floatv4si, floatv2di): New.
18328 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18329 * config/aarch64/aarch64-simd.md
18330 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18331 * config/aarch64/iterators.md (FLOATUORS): New.
18332 (optab): Add float, floatuns.
18333 (su_optab): Likewise.
18334
ce966824
JG		 183352013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18336
18337 * config/aarch64/aarch64-builtins.c
18338 (aarch64_builtin_vectorized_function): Use new names for
18339 fcvt builtins.
18340 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18341 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18342 (fcvtzu): Split as...
18343 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18344 (fcvtas): Split as...
18345 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18346 (fcvtau): Split as...
18347 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18348 (fcvtps): Split as...
18349 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18350 (fcvtpu): Split as...
18351 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18352 (fcvtms): Split as...
18353 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18354 (fcvtmu): Split as...
18355 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18356 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18357 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18358 (lfrintnusf, lfrintnudf): Likewise.
18359 * config/aarch64/aarch64-simd.md
18360 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18361 define_insn.
18362 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18363 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18364 (fcvt_pattern): Likewise.
18365
b9de24fe
JG		 183662013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18367
18368 * config/aarch64/aarch64-simd.md
18369 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18370 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18371
77a205be
JG		 183722013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18373
18374 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18375 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18376 (vrnd<a,m,n,p>_f32): Implement using builtins.
18377 (vrnd<i,x><q>_f<32, 64>): New.
18378
0659ce6f
JG		 183792013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18380
18381 * config/aarch64/aarch64-builtins.c
18382 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18383 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18384 (frintz): Rename to...
18385 (btrunc): ...this.
18386 (frintp): Rename to...
18387 (ceil): ...this.
18388 (frintm): Rename to...
18389 (floor): ...this.
18390 (frinti): Rename to...
18391 (nearbyint): ...this.
18392 (frintx): Rename to...
18393 (rint): ...this.
18394 (frinta): Rename to...
18395 (round): ...this.
18396 * config/aarch64/aarch64-simd.md
18397 (aarch64_frint<frint_suffix><mode>): Delete.
18398 (<frint_pattern><mode>2): Convert to insn.
18399 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18400 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18401 (frint_pattern): Likewise.
18402 (frint_suffix): Likewise.
18403
ea78906a
RB		 184042013-04-29 Richard Biener <rguenther@suse.de>
18405
18406 PR tree-optimization/57081
18407 * loop-init.c: Include tree-flow.h.
18408 (loop_optimizer_finalize): Free number of iteration estimates.
18409 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18410
baee1763
JJ		 184112013-04-29 Jakub Jelinek <jakub@redhat.com>
18412
94dc5332
JJ		 18413 PR tree-optimization/57083
18414 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18415 non-singleton shift count range, zero extend low_bound for uns case.
18416
baee1763
JJ		 18417 * config/i386/predicates.md (general_vector_operand): New predicate.
18418 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18419 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18420 if they aren't nonimmediate operands. If their original values
18421 satisfy const_vector_equal_evenodd_p, don't shift them.
18422 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18423 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18424 if not nonimmediate_operand.
18425 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18426 instead of register_operand.
18427 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18428
a9073727 184292013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 18430
18431 * stor-layout.c (finalize_size_functions): Allocate a structure and
18432 reset cfun before dumping the functions.
18433
ba8011e6
JJ		 184342013-04-27 Jakub Jelinek <jakub@redhat.com>
18435
d6fde69e
JJ		 18436 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18437
ba8011e6
JJ		 18438 PR target/56866
18439 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18440 use xop_pmacsdqh if uns_p.
18441 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18442 the immediate rotate count.
18443
2c62cbaa
VM		 184442013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18445
18446 * rtl.h (struct rtx_def): Add comment for field jump.
18447 (LRA_SUBREG_P): New macro.
18448 * recog.c (register_operand): Check LRA_SUBREG_P.
18449 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18450 stack frame.
18451 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18452 (lra_final_code_change): Skip subreg change for operators.
18453 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18454 if there are no operand changes.
18455 * lra-constraints.c (curr_insn_set): New.
18456 (match_reload): Set LRA_SUBREG_P.
18457 (emit_spill_move): Ditto.
18458 (check_and_process_move): Use curr_insn_set. Process only single
18459 set insns. Don't initialize sec_mem_p and change_p.
18460 (simplify_operand_subreg): Use LRA_SUBREG_P.
18461 (reg_in_class_p): New function.
18462 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18463 of #ifdef. Add code to remove cycling.
18464 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18465 non-null disp. Reload inner instead of disp when base and index
18466 are null. Try to put lo_sum into register.
18467 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18468 (check_and_process_move): Move code for move cost check to
18469 simple_move_p. Remove equiv_substitution.
18470 (simple_move_p): New function.
18471 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18472 curr_insn_set. Call check_and_process_move only for single set
18473 insns. Use the new function. Move call of check_and_process_move
18474 after operand equiv substitution and address process.
18475
e7d764f3
JJ		 184762013-04-26 Jakub Jelinek <jakub@redhat.com>
18477
18478 PR go/57045
18479 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18480 with nonlocal goto receivers or returns twice calls, ignore
18481 unininitialized values from abnormal edges to nl goto receiver
18482 or returns twice call.
18483
41e10689
JJ		 184842013-04-26 Jakub Jelinek <jakub@redhat.com>
18485
18486 PR tree-optimization/57051
18487 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18488 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18489 bitsize.
18490
d7ed20db
RB		 184912013-04-26 Richard Biener <rguenther@suse.de>
18492
18493 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18494 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18495 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18496 (fixup_loop_arrays_after_move): New function.
18497 (move_sese_region_to_fn): Properly outline the loop tree parts
18498 of the SESE region.
18499
df93505e
UB		 185002013-04-26 Uros Bizjak <ubizjak@gmail.com>
18501
18502 * config/i386/i386.md (type, unit): Fix long lines.
18503
dd366ec3
RB		 185042013-04-26 Richard Biener <rguenther@suse.de>
18505
18506 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18507 (lto-streamer-out.o): Likewise.
18508 * cfgloop.c (init_loops_structure): Export, add struct function
18509 argument and adjust.
18510 (flow_loops_find): Adjust.
18511 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18512 (init_loops_structure): Declare.
18513 * lto-streamer-in.c: Include cfgloop.h.
18514 (input_cfg): Input the loop tree.
18515 * lto-streamer-out.c: Include cfgloop.h.
18516 (output_cfg): Output the loop tree.
18517 (output_struct_function_base): Do not drop PROP_loops.
18518
a9e0d843
RB		 185192013-03-26 Richard Biener <rguenther@suse.de>
18520
18521 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18522 (pass_build_cfg): Provide PROP_loops.
18523 (move_sese_region_to_fn): Remove loops that are outlined into fn
18524 for now.
18525 * tree-inline.c: Include cfgloop.h.
18526 (initialize_cfun): Do not drop PROP_loops.
18527 (copy_loops): New function.
18528 (copy_cfg_body): Copy loop structure.
18529 (tree_function_versioning): Initialize destination loop tree.
18530 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18531 (pass_parallelize_loops): Do IL verification.
18532 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18533 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18534 the CFG make sure we fixup loops as well.
18535 * tree-ssa-tail-merge.c: Include cfgloop.h.
18536 (replace_block_by): When merging loop latches mark loops for fixup.
18537 * lto-streamer-out.c (output_struct_function_base): Drop
18538 PROP_loops for now.
18539 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18540 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18541 * ipa-split.c: Include cfgloop.h.
18542 (split_function): Add the new return block to the loop tree root.
18543 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18544 whether we have removed the forwarder block.
18545 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18546 * cfgloop.h (place_new_loop): Declare.
18547 * cfgloopmanip.c (place_new_loop): Export.
18548 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18549 (tree-switch-conversion.o): Likewise.
18550 (tree-complex.o): Likewise.
18551 (tree-inline.o): Likewise.
18552 (tree-ssa-tailmerge.o): Likewise.
18553 (ipa-split.o): Likewise.
18554 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18555 (tree-ssa-copy.o): Likewise.
18556 * tree-switch-conversion.c: Include cfgloop.h
18557 (process_switch): If we emit a bit-test cascade, schedule loops
18558 for fixup.
18559 * tree-complex.c: Include cfgloop.h.
18560 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18561 * asan.c: Include cfgloop.h.
18562 (create_cond_insert_point): Properly add new basic-blocks to
18563 loops, schedule loop fixup.
18564 * cfgloop.c (verify_loop_structure): Check that looks are not
18565 marked for fixup.
18566 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18567 to loops.
18568 (expand_omp_for_generic): Likewise.
18569 (expand_omp_sections): Likewise.
18570 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18571 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18572 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18573 is initialized, not when loops are present.
18574 * tree-parloops.c (parallelize_loops): Remove checking here.
18575 * passes.c (init_optimization_passes): Schedule a copy-propagation
18576 pass before complete unrolling of inner loops.
18577
e78e8a0b
JJ		 185782013-04-26 Jakub Jelinek <jakub@redhat.com>
18579
a2e836b2
JJ		 18580 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18581 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18582 (colorize_init): Add argument to _WIN32 version.
18583 * toplev.c: Include diagnostic-color.h.
18584 (process_options): Default to -fdiagnostics-color=auto if
18585 GCC_COLORS env var is in the environment.
18586 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 18587 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18588 env var is in the environment, the default is auto rather than never.
a2e836b2 18589
e78e8a0b
JJ		 18590 * diagnostic.h (file_name_as_prefix): Add context argument.
18591 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18592 the string as locus.
18593 * langhooks.c (lhd_print_error_function): Adjust caller.
18594
013e5ef9
LC		 185952013-04-25 Lawrence Crowl <crowl@google.com>
18596
18597 * var-tracking.c (shared_hash_def::htab):
18598 Change type to hash_table. Update dependent calls and types.
18599
4a8fb1a1
LC		 186002013-04-25 Lawrence Crowl <crowl@google.com>
18601
18602 * Makefile.in: Update as needed below.
18603
18604 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18605 Move declaration to after the type's method definitons.
18606
18607 * attribs.c (htab_t scoped_attributes::attribute_hash):
18608 Change type to hash_table. Update dependent calls and types.
18609
18610 * bitmap.c (htab_t bitmap_desc_hash):
18611 Change type to hash_table. Update dependent calls and types.
18612
18613 * cselib.c (htab_t cselib_hash_table):
18614 Change type to hash_table. Update dependent calls and types.
18615
18616 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18617 (hash_string_slot_node): Move implementation into lto-streamer.h
18618 struct string_slot_hasher.
18619 (eq_string_slot_node): Likewise.
18620
18621 * data-streamer-out.c: Update output_block::string_hash_table
18622 dependent calls and types.
18623
18624 * dwarf2cfi.c (htab_t trace_index):
18625 Change type to hash_table. Update dependent calls and types.
18626
18627 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18628 Change type to hash_table. Update dependent calls and types.
18629 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18630 (htab_t optimize_external_refs::map): Likewise.
18631 (htab_t output_comp_unit::extern_map): Likewise.
18632 (htab_t output_comdat_type_unit::extern_map): Likewise.
18633 (htab_t output_macinfo::macinfo_htab): Likewise.
18634 (htab_t optimize_location_lists::htab): Likewise.
18635 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18636
18637 * except.c (htab_t ehspec_hash_type):
18638 Change type to hash_table. Update dependent calls and types.
18639 (assign_filter_values::ttypes): Likewise.
18640 (assign_filter_values::ehspec): Likewise.
18641 (sjlj_assign_call_site_values::ar_hash): Likewise.
18642 (convert_to_eh_region_ranges::ar_hash): Likewise.
18643
18644 * gcse.c (htab_t pre_ldst_table):
18645 Change type to hash_table. Update dependent calls and types.
18646
18647 * ggc-common.c (htab_t saving_htab):
18648 Change type to hash_table. Update dependent calls and types.
18649 (htab_t loc_hash): Likewise.
18650 (htab_t ptr_hash): Likewise.
18651 (call_count): Rename ggc_call_count.
18652 (call_alloc): Rename ggc_call_alloc.
18653 (loc_descriptor): Rename make_loc_descriptor.
18654 (add_statistics): Rename ggc_add_statistics.
18655
18656 * ggc-common.c (saving_htab):
18657 Change type to hash_table. Update dependent calls and types.
18658
18659 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18660 (push_gimplify_context): Likewise.
18661 (pop_gimplify_context): Likewise.
18662 (struct gimple_temp_hash_elt): Added.
18663 (struct gimplify_hasher): Likewise.
18664 (struct gimplify_ctx.temp_htab):
18665 Change type to hash_table. Update dependent calls and types.
18666
18667 * gimple-fold.c: Include gimplify-ctx.h.
18668
18669 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18670 Change type to hash_table. Update dependent calls and types.
18671 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18672 avoid potential global name collision.
18673
18674 * gimplify.c: Include gimplify-ctx.h.
18675 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18676 (htab_t gimplify_ctx::temp_htab):
18677 Update dependent calls and types for new type hash_table.
18678 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18679 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18680
18681 * gimplify-ctx.h: New.
18682 (struct gimple_temp_hash_elt): Move from gimplify.c.
18683 (class gimplify_hasher): New.
18684 (struct gimplify_ctx): Move from gimple.h.
18685 (htab_t gimplify_ctx::temp_htab):
18686 Change type to hash_table. Update dependent calls and types.
18687
18688 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18689 (htab_t ivs_params::newivs_index):
18690 Change type to hash_table. Update dependent calls and types.
18691 (htab_t ivs_params::params_index): Likewise.
18692 (htab_t print_generated_program::params_index): Likewise.
18693 (htab_t gloog::newivs_index): Likewise.
18694 (htab_t gloog::params_index): Likewise.
18695
18696 * graphite.c: Include graphite-htab.h.
18697 4htab_t graphite_transform_loops::bb_pbb_mapping):
18698 Change type to hash_table. Update dependent calls and types.
18699
18700 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18701 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18702 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18703
18704 * graphite-dependences.c: Include graphite-htab.h.
18705 (loop_is_parallel_p): Change hash table type of parameter.
18706
18707 * graphite-htab.h: New.
18708 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18709 (extern find_pbb_via_hash): Move from graphite-poly.h.
18710 (extern loop_is_parallel_p): Move from graphite-poly.h.
18711 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18712
18713 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18714 (extern loop_is_parallel_p): Move to graphite-htab.h.
18715 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18716
18717 * haifa-sched.c (htab_t delay_htab):
18718 Change type to hash_table. Update dependent calls and types.
18719 (htab_t delay_htab_i2): Likewise.
18720
18721 * ira-color.c (htab_t allocno_hard_regs_htab):
18722 Change type to hash_table. Update dependent calls and types.
18723
18724 * ira-costs.c (htab_t cost_classes_htab):
18725 Change type to hash_table. Update dependent calls and types.
18726
18727 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18728 Change type to hash_table. Update dependent calls and types.
18729
18730 * loop-iv.c (htab_t bivs):
18731 Change type to hash_table. Update dependent calls and types.
18732
18733 * loop-unroll.c (htab_t opt_info::insns_to_split):
18734 Change type to hash_table. Update dependent calls and types.
18735 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18736
18737 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18738 (struct string_slot_hasher): New.
18739 (htab_t output_block::string_hash_table):
18740 Change type to hash_table. Update dependent calls and types.
18741
18742 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18743 (htab_t file_name_hash_table):
18744 Change type to hash_table. Update dependent calls and types.
18745
18746 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18747 calls and types.
18748
18749 * lto-streamer.c (htab_t tree_htab):
18750 Change type to hash_table. Update dependent calls and types.
18751
18752 * omp-low.c: Include gimplify-ctx.h.
18753
18754 * passes.c (htab_t name_to_pass_map):
18755 Change type to hash_table. Update dependent calls and types.
18756 (pass_traverse): Rename to passes_pass_traverse.
18757
18758 * plugin.c (htab_t event_tab):
18759 Change type to hash_table. Update dependent calls and types.
18760
18761 * postreload-gcse.c (htab_t expr_table):
18762 Change type to hash_table. Update dependent calls and types.
18763 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18764
18765 * sese.c (debug_rename_map_1): Make extern.
18766 (htab_t copy_bb_and_scalar_dependences::rename_map):
18767 Change type to hash_table. Update dependent calls and types.
18768
18769 * sese.h (extern debug_rename_map): Move to .c file.
18770
18771 * store-motion.c (htab_t store_motion_mems_table):
18772 Change type to hash_table. Update dependent calls and types.
18773
18774 * trans-mem.c (htab_t tm_new_mem_hash):
18775 Change type to hash_table. Update dependent calls and types.
18776
18777 * tree-browser.c (htab_t TB_up_ht):
18778 Change type to hash_table. Update dependent calls and types.
18779
18780 * tree-cfg.c (htab_t discriminator_per_locus):
18781 Change type to hash_table. Update dependent calls and types.
18782
18783 * tree-complex.c: Include tree-hasher.h
18784 (htab_t complex_variable_components):
18785 Change type to hash_table. Update dependent calls and types.
18786
18787 * tree-eh.c (htab_t finally_tree):
18788 Change type to hash_table. Update dependent calls and types.
18789
18790 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18791 struct int_tree_hasher.
18792 (extern int_tree_map_eq): Likewise.
18793 (uid_decl_map_hash): Removed.
18794 (extern decl_tree_map_eq): Likewise.
18795
18796 * tree-hasher.h: New.
18797 (struct int_tree_hasher): New.
18798 (typedef int_tree_htab_type): New.
18799
18800 * tree-inline.c: Include gimplify-ctx.h.
18801
18802 * tree-mudflap.c: Include gimplify-ctx.h.
18803
18804 * tree-parloops.c: Include tree-hasher.h.
18805 (htab_t eliminate_local_variables_stmt::decl_address):
18806 Change type to hash_table. Update dependent calls and types.
18807 (htab_t separate_decls_in_region::decl_copies): Likewise.
18808
18809 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18810 Change type to hash_table. Update dependent calls and types.
18811
18812 * tree-sra.c (candidates):
18813 Change type to hash_table. Update dependent calls and types.
18814
18815 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18816 in tree-flow.h.
18817 (int_tree_map_hash): Likewise.
18818
18819 * tree-ssa-dom.c (htab_t avail_exprs):
18820 Change type to hash_table. Update dependent calls and types.
18821
18822 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18823 Change type to hash_table. Update dependent calls and types.
18824
18825 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18826 Change type to hash_table. Update dependent calls and types.
18827
18828 * tree-ssa-phiopt.c (seen_ssa_names):
18829 Change type to hash_table. Update dependent calls and types.
18830
18831 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18832 Change type to hash_table. Update dependent calls and types.
18833
18834 * tree-ssa-uncprop.c (equiv):
18835 Change type to hash_table. Update dependent calls and types.
18836
c5a44004
JJ		 188372013-04-25 Jakub Jelinek <jakub@redhat.com>
18838
18839 PR rtl-optimization/57003
18840 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18841 call note_stores with kill_clobbered_value callback again after
18842 killing regs_invalidated_by_call.
18843
09962a4a
JG		 188442013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18845
18846 * config/aarch64/aarch64-simd.md
18847 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18848 (aarch64_simd_bsl<mode>): Likewise.
18849 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18850
ea28bb0b
MP		 188512013-04-25 Marek Polacek <polacek@redhat.com>
18852
18853 PR tree-optimization/57066
3c21604f 18854 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 18855
96659611
JG		 188562013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18857
18858 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18859
9697e620
JG		 188602013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18861
18862 * config/aarch64/aarch64-builtins.c
18863 (aarch64_fold_builtin): New.
18864 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18865 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18866 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18867 * config/aarch64/arm_neon.h
18868 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18869
0ac198d3
JG		 188702013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18871 Tejas Belagod <tejas.belagod@arm.com>
18872
18873 * config/aarch64/aarch64-builtins.c
18874 (aarch64_gimple_fold_builtin): New.
18875 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18876 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18877 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18878 (addvv4sf): Update.
18879 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18880
58cff58c
N		 188812013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18882
df93505e 18883 * config/aarch64/aarch64.md
58cff58c
N		 18884 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18885
7e0228bf
N		 188862013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18887
18888 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18889 (*ngcsi_uxtw): New pattern.
18890
5819f96f 188912013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 18892 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 18893
18894 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18895 (TB_DREG): Add T_V4HF.
18896 (v4hf_UP): New macro.
18897 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 18898 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 18899 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18900 built-ins.
df93505e 18901 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 18902 (arm_vector_mode_supported_p): Handle V4HF.
18903 (arm_mangle_map): Handle V4HFmode.
18904 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18905 * config/arm/arm_neon_builtins.def: Add entries for
18906 vcvtv4hfv4sf, vcvtv4sfv4hf.
18907 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18908 (neon_vcvtv4hfv4sf): Likewise.
18909 * config/arm/neon-gen.ml: Handle half-precision floating point
18910 features.
18911 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18912 * config/arm/arm_neon.h: Regenerate.
18913 * config/arm/neon.ml (type elts): Add F16.
18914 (type vectype): Add T_float16x4, T_floatHF.
18915 (type vecmode): Add V4HF.
18916 (type features): Add Requires_FP_bit feature.
18917 (elt_width): Handle F16.
18918 (elt_class): Likewise.
18919 (elt_of_class_width): Likewise.
18920 (mode_of_elt): Refactor.
18921 (type_for_elt): Handle F16, fix error messages.
18922 (vectype_size): Handle T_float16x4.
18923 (vcvt_sh): New function.
18924 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18925 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18926 (string_of_mode): Handle V4HF.
18927 * doc/arm-neon-intrinsics.texi: Regenerate.
18928
1ef395e4
JG		 189292013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18930
18931 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
18932 format specifier in 'X' case.
18933
41c34e94
AM		 189342013-04-25 Alan Modra <amodra@gmail.com>
18935
18936 PR target/57052
18937 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
18938 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
18939 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
18940 Repeat for many other rotate/shift and mask patterns using subregs.
18941 Name lshiftrt insns.
18942 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
18943 on WORDS_BIG_ENDIAN.
18944
b9a7eb5d
AM		 189452013-04-25 Alan Modra <amodra@gmail.com>
18946
18947 * config.gcc: Support little-endian powerpc-linux targets.
18948 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
18949 (LINK_OS_LINUX_SPEC): Define.
18950 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
18951 Preserve MASK_LITTLE_ENDIAN.
18952 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
18953 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
18954 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
18955 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
18956 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
18957 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
18958 Correct fp word order for little-endian. Don't shift toc entries
18959 smaller than a word for little-endian.
18960 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
18961 (bswapdi2 splits): Correct low-part subreg for little-endian.
18962 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
18963 low/high where such is correct only for be.
18964 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
18965 little-endian for -mcall-aixdesc.
18966
87f73374
AM		 189672013-04-25 Alan Modra <amodra@gmail.com>
18968
18969 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
18970 replace_equiv_address_nv.
18971
cabf91cd
AM		 189722013-04-25 Alan Modra <amodra@gmail.com>
18973
18974 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
18975
0ae24cc8
VM		 189762013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18977
18978 Revert:
18979 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18980 * rtl.h (struct rtx_def): ...
cabf91cd 18981
77bce07c
VM		 189822013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18983
18984 PR rtl-optimizations/57046
18985 * lra-constraints (split_reg): Set up lra_risky_transformations_p
18986 for multi-reg splits.
18987
0db63e7f
L		 189882013-04-24 H.J. Lu <hongjiu.lu@intel.com>
18989
18990 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
18991
3362b6b6
SA		 189922013-04-24 Sterling Augustine <saugustine@google.com>
18993
18994 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
18995 (comp_dir_string, debug_str_dwo_section): New.
18996 (DEBUG_STR_DWO_SECTION): Rename to ...
18997 (DEBUG_DWO_STR_SECTION): ... this.
18998 (DEBUG_NORM_STR_SECTION): Delete.
18999 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19000 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19001 (find_AT_string): Move most logic to ...
19002 (find_AT_string_in_table): ... here. New.
19003 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19004 add_skeleton_AT_string. Delete logic.
19005 (output_skeleton_debug_sections): Remove call to
19006 add_top_level_skeleton_die_attrs.
19007 (add_comp_dir_attribute): Move logic to comp_dir_string.
19008 (dwarf2out_init): Initialize debug_str_dwo_section.
19009 (output_indirect_string): Call find_string_form.
19010 (output_indirect_strings): Rewrite.
19011 (prune_unused_types): Empty skeleton_debug_str_hash.
19012 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19013 (dwarf2out_finish): Call output_indirect_strings.
19014
e93e18e9
PC		 190152013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19016
19017 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19018
f6ce35ac
VM		 190192013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19020
cabf91cd 19021 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 19022 (LRA_SUBREG_P): New macro.
19023 * recog.c (register_operand): Check LRA_SUBREG_P.
19024 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19025 stack frame.
19026 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19027 (lra_final_code_change): Skip subreg change for operators.
19028 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19029 if there are no operand changes.
19030 * lra-constraints.c (curr_insn_set): New.
19031 (match_reload): Set LRA_SUBREG_P.
19032 (emit_spill_move): Ditto.
19033 (check_and_process_move): Use curr_insn_set. Process only single
19034 set insns. Don't initialize sec_mem_p and change_p.
19035 (simplify_operand_subreg): Use LRA_SUBREG_P.
19036 (reg_in_class_p): New function.
19037 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19038 of #ifdef. Add code to remove cycling.
19039 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19040 non-null disp. Reload inner instead of disp when base and index
19041 are null. Try to put lo_sum into register.
19042 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 19043 (check_and_process_move): Move code for move cost check to
f6ce35ac 19044 simple_move_p. Remove equiv_substitution.
cabf91cd 19045 (simple_move_p): New function.
f6ce35ac
VM		 19046 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19047 curr_insn_set. Call check_and_process_move only for single set
19048 insns. Use the new function. Move call of check_and_process_move
19049 after operand equiv substitution and address process.
19050
38047d90
JG		 190512013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19052
19053 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19054 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19055 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19056
13f39b2e
PC		 190572013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19058
19059 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19060
97c116dc
MP		 190612013-04-24 Marek Polacek <polacek@redhat.com>
19062
19063 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19064 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19065 (select_loops_exit_conditions): Likewise.
19066 (number_of_iterations_for_all_loops): Likewise.
19067 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19068 (scev_analysis): Likewise.
19069
83082391 190702013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 19071 Chao-ying Fu <fu@mips.com>
83082391 19072
cabf91cd
AM		 19073 * config/mips/micromips.md (jraddiusp): New pattern.
19074 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19075 instruction if possible.
83082391 19076
19e34aa2
AM		 190772013-04-24 Alan Modra <amodra@gmail.com>
19078
19079 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19080
fdb6603c
JB		 190812013-04-24 Julian Brown <julian@codesourcery.com>
19082 Chung-Lin Tang <cltang@codesourcery.com>
19083
19084 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 19085 dependency behavior in enumeration type DIE generation. Add TODO note
19086 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 19087
bf190e8d
LC		 190882013-04-23 Lawrence Crowl <crowl@google.com>
19089
19090 * Makefile.in: Update as needed below.
19091
19092 * hash-table.h (class hash_table):
19093 Correct many methods with parameter types compare_type to the correct
19094 value_type. (Correct code was unlikely to notice the change.)
19095 (hash_table::elements_with_deleted) New.
19096 (class hashtable::iterator): New.
19097 (hashtable::begin()): New.
19098 (hashtable::end()): New.
19099 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19100
19101 * statistics.c (statistics_hashes):
19102 Change type to hash_table. Update dependent calls and types.
19103
19104 * tree-into-ssa.c (var_infos):
19105 Change type to hash_table. Update dependent calls and types.
19106
19107 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19108 Change type to hash_table. Update dependent calls and types.
19109
19110 * tree-ssa-loop-im.c (struct mem_ref.refs):
19111 Change type to hash_table. Update dependent calls and types.
19112
19113 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19114 Change type to hash_table. Update dependent calls and types.
19115
19116 * tree-ssa-sccvn.c (vn_tables_s::nary):
19117 Change type to hash_table. Update dependent calls and types.
19118 (vn_tables_s::phis): Likewise.
19119 (vn_tables_s::references): Likewise.
19120
19121 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19122 (vn_reference_eq): Update parameter and return types.
19123
19124 * tree-ssa-structalias.c (pointer_equiv_class_table):
19125 Change type to hash_table. Update dependent calls and types.
19126 (location_equiv_class_table): Likewise.
19127
19128 * tree-vect-data-refs.c: Consequential changes for making
19129 peeling a hash_table.
19130
19131 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19132 (destroy_loop_vec_info): Dependent hash_table update.
19133
19134 * tree-vectorizer.h (peeling_htab):
19135 Change type to hash_table. Update dependent calls and types.
19136
d70a81dd
SC		 191372013-04-23 Shiva Chen <shiva0217@gmail.com>
19138
cabf91cd
AM		 19139 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19140 to check the register content is equal or not.
19141 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19142 to assign register content record.
19143 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 19144 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 19145 * lra-int.h (struct lra_reg): Add offset member.
19146 (lra_reg_val_equal_p): New static inline function.
19147 (lra_update_reg_val_offset): New static inline function.
19148 (lra_assign_reg_val): New static inline function.
19149 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19150 to assign register content record.
19151 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 19152
b894a1f3
CM		 191532013-04-23 Catherine Moore <clm@codesourcery.com>
19154
19155 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19156 operands. Record compression.
19157
ecd14de9
XDL		 191582013-04-23 Xinliang David Li <davidxl@google.com>
19159
19160 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19161
92e776e9
RB		 191622013-04-23 Richard Biener <rguenther@suse.de>
19163
19164 PR middle-end/57036
19165 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19166 parameter, only add abnormal goto edges from the copied body
19167 if the call could perform abnormal gotos.
19168 (copy_cfg_body): Adjust.
19169
a15ee567
SN		 191702013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19171
19172 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19173
08c52234
AS		 191742013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19175
19176 * coretypes.h (gimple_stmt_iterator): Add struct to make
19177 compatible with C.
19178
999c1171
RB		 191792013-04-23 Richard Biener <rguenther@suse.de>
19180
19181 PR tree-optimization/57026
19182 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19183 from SSA names occuring in abnormal PHI nodes.
19184
53669259
AK		 191852013-04-22 Andi Kleen <ak@linux.intel.com>
19186
19187 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19188
1a0ad150
AK		 191892013-04-22 Andi Kleen <ak@linux.intel.com>
19190
19191 * lto/lto.c (print_lto_report_1): Declare early.
19192 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19193
057f8f20
AK		 191942013-04-22 Andi Kleen <ak@linux.intel.com>
19195
19196 * common.opt (-flto-report-wpa): Add.
19197 * doc/invoke.texi (-flto-report-wpa): Add.
19198 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19199 (lto_main): dito.
19200
473b1e05
XDL		 192012013-04-22 Xinliang David Li <davidxl@google.com>
19202
19203 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19204 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19205 * Makefile.in: New dependency
19206
cabf91cd 19207 David Daney <ddaney.cavm@gmail.com>
b1485a33 19208
cabf91cd
AM		 19209 * configure.ac (gcc_cv_as_micromips_support): Use the
19210 --fatal-warnings option.
19211 * configure: Regenerate.
b1485a33 19212
829d0168
MP		 192132013-04-22 Marek Polacek <polacek@redhat.com>
19214
19215 PR sanitizer/56990
19216 * tsan.c (instrument_expr): Don't instrument expression
19217 in case its size is zero.
19218
6d9b7208
UB		 192192013-04-22 Uros Bizjak <ubizjak@gmail.com>
19220
19221 PR target/57032
19222 Revert:
19223 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19224
19225 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19226
ea679d55
JG		 192272013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19228
19229 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19230 (gimple_stmt_iterator): New typedef.
19231 * gimple.h (gimple_stmt_iterator): Rename to...
19232 (gimple_stmt_iterator_d): ... This.
19233 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19234 trees be valid for GIMPLE and GENERIC.
19235 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19236 * gimple-fold.c (gimple_fold_call): Call target hook
19237 gimple_fold_builtin.
19238 * hooks.c (hook_bool_gsiptr_false): New.
19239 * hooks.h (hook_bool_gsiptr_false): New.
19240 * target.def (fold_stmt): New.
19241 * doc/tm.texi: Regenerate.
19242
88a581da
VM		 192432013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19244
19245 PR target/57018
19246 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19247 a set sp if no stack realignment.
19248
92be22dc
NC		 192492013-04-22 Nick Clifton <nickc@redhat.com>
19250
19251 * config.gcc (tilegx-linux): Extend extra_objs rather than
19252 overwriting it.
19253 (tilepro-linux): Likewise.
19254
0ddec79f
JG		 192552013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19256
19257 * config/aarch64/aarch64-builtins.c
19258 (CF): Remove.
19259 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19260 (VAR<1-12>): Add MAP parameter.
19261 (BUILTIN_*): Likewise.
19262 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19263 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19264 (aarch64_ushl_n<mode>): Likewise.
19265 (aarch64_sshr_n<mode>): Likewise.
19266 (aarch64_ushr_n<mode>): Likewise.
19267 (aarch64_<maxmin><mode>): Likewise.
19268 (aarch64_sqrt<mode>): Likewise.
19269 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19270 (vshr<q>_n_*): Likewise.
19271
0050faf8
JG		 192722013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19273
19274 * config/aarch64/aarch64-builtins.c
19275 (aarch64_simd_builtin_type_mode): Handle SF types.
19276 (sf_UP): Define.
19277 (BUILTIN_GPF): Define.
19278 (aarch64_init_simd_builtins): Handle SF types.
19279 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19280 (frecps): Likewise.
19281 (frecpx): Likewise.
19282 * config/aarch64/aarch64-simd.md
19283 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19284 (aarch64_frecpe<mode>): New.
19285 (aarch64_frecps<mode>): Likewise.
19286 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19287 (v8type): Add frecp<esx>.
19288 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19289 (aarch64_frecps<mode>): Likewise.
19290 * config/aarch64/iterators.md (FRECP): New.
19291 (frecp_suffix): Likewise.
19292 * config/aarch64/arm_neon.h
19293 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19294
0fad3dbc 192952013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 19296
19297 PR target/56995
19298 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19299 (REG_CLASS_NAMES): Idem.
19300 (REG_CLASS_CONTENTS): Idem.
19301 (REGCLASS_HAS_FP_REG): Idem.
19302 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19303 (sh_conditional_register_usage): Idem.
19304
3e8a33f9
JL		 193052013-04-21 Jeff Law <law@redhat.com>
19306
19307 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19308 (ssa_forward_propagate_and_combine): Use it.
19309
f38e1b0a
VM		 193102013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19311
19312 * lra.c: Update the flow chart diagram.
19313
682303da
VM		 193142013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19315
19316 PR rtl-optimization/56847
19317 * lra-constraints.c (process_alt_operands): Discourage alternative
19318 with non-matche doffsettable memory constraint fro memory with
19319 known offset.
19320
f6b64c35
RB		 193212013-04-19 Richard Biener <rguenther@suse.de>
19322
19323 PR tree-optimization/56982
19324 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19325 function.
19326 * gimplify.c (gimplify_call_expr): Notice special calls.
19327 (gimplify_modify_expr): Likewise.
19328 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19329 abnormal control flow receivers.
19330 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19331 in the same way as cfun->has_nonlocal_labels.
19332 (gimple_purge_dead_abnormal_call_edges): Likewise.
19333 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19334 receivers start a basic-block.
19335
01d8bf07
RB		 193362013-04-19 Richard Biener <rguenther@suse.de>
19337
19338 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19339 member ...
19340 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19341 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19342 (SLP_TREE_LOAD_PERMUTATION): Add.
19343 (vect_transform_slp_perm_load): Adjust prototype.
19344 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19345 (vect_free_slp_instance): Likewise.
19346 (vect_create_new_slp_node): Likewise.
19347 (vect_supported_slp_permutation_p): Remove.
19348 (vect_slp_rearrange_stmts): Adjust.
19349 (vect_supported_load_permutation_p): Likewise. Inline
19350 vect_supported_slp_permutation_p here.
19351 (vect_analyze_slp_instance): Compute load permutations per
19352 slp node instead of per instance.
19353 (vect_get_slp_defs): Adjust.
19354 (vect_transform_slp_perm_load): Likewise.
19355 (vect_schedule_slp_instance): Remove redundant code.
19356 (vect_schedule_slp): Remove hack for PR56270, add it ...
19357 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19358 CSE loads for SLP. Adjust.
19359
ede22fc3
GY		 193602013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19361
19362 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19363 spelling in two comments.
19364
67bc84fb
GY		 193652013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19366
19367 PR target/56797
19368 * config/arm/arm.c (load_multiple_sequence): Require SP
19369 as base register for loads if SP is in the register list.
19370
e248d83f
MJ		 193712013-04-19 Martin Jambor <mjambor@suse.cz>
19372
19373 PR tree-optimization/56718
19374 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19375 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19376 and made public. Adjusted all callers.
19377 (ipa_intraprocedural_devirtualization): New function.
19378 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19379 (ipa_intraprocedural_devirtualization): Likewise.
19380 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19381
4891e8f8
RB		 193822013-04-19 Richard Biener <rguenther@suse.de>
19383
19384 PR tree-optimization/57000
19385 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19386
dad89f7c
TG		 193872013-04-19 Terry Guo <terry.guo@arm.com>
19388
19389 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19390 Replace with ...
19391 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19392 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19393 (cortex_m4_fmacs): Use new reservations.
19394 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19395
72ea0d47
VM		 193962013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19397
f1e6512c 19398 PR rtl-optimization/56999
72ea0d47
VM		 19399 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19400 related code.
19401 (lra_coalesce): Remove split_origin_bitmap and related code.
19402 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19403 ranges if necessary.
19404
780a5b71
UB		 194052013-04-18 Uros Bizjak <ubizjak@gmail.com>
19406
19407 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19408 New array.
19409 (ix86_expand_call): Remove clobbered_registers array and use
19410 x86_64_ms_sysv_extra_clobbered_registers instead.
19411 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19412 Declare here.
19413 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19414 predicate.
19415 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19416 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19417 (*call_value_rex64_ms_sysv): Ditto.
19418
6f5a366a
CC		 194192013-04-18 Cary Coutant <ccoutant@google.com>
19420
19421 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19422 parent instead of die_mark.
19423
475b8f37
DN		 194242013-04-18 Diego Novillo <dnovillo@google.com>
19425
19426 * gimple.c (create_gimple_tmp): New.
19427 (get_expr_type): New.
19428 (build_assign): New.
19429 (build_type_cast): New.
19430 * gimple.h (enum ssa_mode): Define.
19431 (gimple_seq_set_location): New.
19432 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 19433 to use build_assign and build_type_cast.
475b8f37 19434
08940f33
RB		 194352013-04-18 Richard Biener <rguenther@suse.de>
19436
19437 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19438 handle negative step. Remove redundant checks.
19439 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19440 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19441 for negative step and grouped loads fail to vectorize.
19442
0e0f87d4
SB		 194432013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19444
19445 * emit-rtl.c (reset_insn_used_flags): New function.
19446 (reset_all_used_flags): Use it.
19447 (verify_insn_sharing): New function.
19448 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19449
4c445590
JJ		 194502013-04-18 Jakub Jelinek <jakub@redhat.com>
19451
19452 PR tree-optimization/56984
19453 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19454 and (x >> M) >= N don't register any assertion if N << M is the
19455 minimum value.
19456
6873ecab
SB		 194572013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19458
19459 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19460 do not delete_insn insns that have not yet been emitted, only
19461 unlink them with remove_insn.
19462 * df-scan.c (df_insn_delete): Revert r197492.
19463
3ccb989e
SB		 194642013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19465
19466 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19467 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19468
194692013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 19470
19471 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19472 define_insn_and_split.
19473 (and_scc,ior_scc,negscc): Likewise.
19474 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19475
3ccb989e 194762013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 19477
19478 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19479 can benefit from using a sequence of LDRD instructions in epilogue
19480 instead of a single LDM instruction.
19481
6d10a203
MLI		 194822013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19483
19484 PR 45688
19485 * doc/extend.texi: Fix typo.
19486
6983e6b5
RB		 194872013-04-17 Richard Biener <rguenther@suse.de>
19488
19489 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19490 (vect_build_slp_tree): ... here.
19491 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19492 match. Remove special-casing of mismatched complex loads.
19493 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19494 re-try the match with swapped commutative operands.
19495 (vect_supported_load_permutation_p): Remove special-casing of
19496 mismatched complex loads.
19497 (vect_analyze_slp_instance): Adjust.
19498
ef23e6a2
RB		 194992013-04-17 Richard Biener <rguenther@suse.de>
19500
19501 PR rtl-optimization/56921
19502 * cfgloop.h (struct loop): Add simple_loop_desc member.
19503 (struct niter_desc): Mark with GTY(()).
19504 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19505 * loop-iv.c (get_simple_loop_desc): Likewise.
19506 (free_simple_loop_desc): Likewise.
19507
19508 Revert
19509 2013-04-16 Richard Biener <rguenther@suse.de>
19510
19511 PR rtl-optimization/56921
19512 * loop-init.c (pass_rtl_move_loop_invariants): Add
19513 TODO_do_not_ggc_collect to todo_flags_finish.
19514 (pass_rtl_unswitch): Same.
19515 (pass_rtl_unroll_and_peel_loops): Same.
19516 (pass_rtl_doloop): Same.
19517
fc6f94f5
EB		 195182013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19519
19520 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19521 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19522 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19523 references.
19524 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19525 * tree-streamer.c (record_common_node): Adjust reference in comment.
19526
10a88311
TG		 195272013-04-17 Terry Guo <terry.guo@arm.com>
19528
19529 * config/arm/cortex-m4.md: Add a new bypass.
19530
6d9b7208 195312013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 19532
19533 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19534 New pattern.
19535 (*subs_<optab><mode>_multp2): New pattern.
19536 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19537 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19538
6d9b7208 195392013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 19540
19541 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19542 (*subs_mul_imm_<mode>): New pattern.
19543
18a6701e
DE		 195442013-04-16 David Edelsohn <dje.gcc@gmail.com>
19545
19546 PR target/56948
19547 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19548 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19549 (vsx_movti_32bit): Change j->wa to O->wa.
19550
07c37b2f
RB		 195512013-04-16 Richard Biener <rguenther@suse.de>
19552
19553 PR rtl-optimization/56921
19554 * loop-init.c (pass_rtl_move_loop_invariants): Add
19555 TODO_do_not_ggc_collect to todo_flags_finish.
19556 (pass_rtl_unswitch): Same.
19557 (pass_rtl_unroll_and_peel_loops): Same.
19558 (pass_rtl_doloop): Same.
19559
0e0f87d4 195602013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 19561
19562 * config/arm/arm.c (emit_multi_reg_push): New declaration
19563 for an existing function.
19564 (arm_emit_strd_push): New function.
19565 (arm_expand_prologue): Used here.
19566 (arm_emit_ldrd_pop): New function.
19567 (arm_expand_epilogue): Used here.
19568 (arm_get_frame_offsets): Update condition.
19569 (arm_emit_multi_reg_pop): Add a special case for load of a single
19570 register with writeback.
19571
5e8e2af4
UB		 195722013-04-16 Uros Bizjak <ubizjak@gmail.com>
19573
19574 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19575 description.
19576
9fd9ccf7
RB		 195772013-04-16 Richard Biener <rguenther@suse.de>
19578
19579 PR tree-optimization/56756
19580 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19581 (first_mem_ref_loc): New.
19582 (execute_sm): Place the load temporarily before a previous
19583 access instead of in the latch edge to ensure its SSA dependencies
19584 are defined at points dominating the load.
19585
96fba521
SB		 195862013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19587
4c8af858
SB		 19588 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19589 correct fix by moving header and footer insn to the footer of
19590 the merged basic block. Clear BB_END of the merged-away block.
19591
96fba521
SB		 19592 PR middle-end/43631
19593 * emit-rtl.c (make_note_raw): New function.
19594 (link_insn_into_chain): New static inline function.
19595 (add_insn): Use it.
19596 (add_insn_before, add_insn_after): Factor insn chain linking code...
19597 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19598 using link_insn_into_chain.
19599 (note_outside_basic_block_p): New helper function for emit_note_after
19600 and emit_note_before.
19601 (emit_note_after): Use nobb variant of add_insn_after if the note
19602 should not be contained in a basic block.
19603 (emit_note_before): Use nobb variant of add_insn_before if the note
19604 should not be contained in a basic block.
19605 (emit_note_copy): Use make_note_raw.
19606 (emit_note): Likewise.
19607 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19608 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19609 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19610 the moved barrier the tail of the basic block it follows.
19611 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19612
7b8265ba
JJ		 196132013-04-15 Jakub Jelinek <jakub@redhat.com>
19614
19615 PR tree-optimization/56962
19616 * gimple-ssa-strength-reduction.c (record_increment): Only set
19617 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19618 either rhs1 or rhs2 is equal to c->base_expr.
19619
5185d248
RB		 196202013-04-15 Richard Biener <rguenther@suse.de>
19621
19622 PR tree-optimization/56933
19623 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19624 member.
19625 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19626 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19627 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19628 dependence check ...
19629 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19630 ... here.
19631 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19632 GROUP_READ_WRITE_DEPENDENCE.
19633
a24243a0
AK		 196342013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19635
19636 * emit-rtl.c (reset_all_used_flags): New function.
19637 (verify_rtl_sharing): Call reset_all_used_flags before and after
19638 performing the checks.
19639
1c50eada
KT		 196402013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19641
19642 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19643 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19644 * config/arm/constraints.md (De): New constraint.
19645 * config/arm/neon.md (anddi3_neon): Delete.
19646 (neon_vand<mode>): Expand to standard anddi3 pattern.
19647 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19648 Move earlier in the file.
19649 (neon_inv_logic_op2): Likewise.
19650 (arm_anddi_operand_neon): New predicate.
19651
e927b6ad
RO		 196522013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19653
19654 * configure.ac (gcc_cv_ld_as_needed): Set
19655 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19656 Use -z ignore, -z record on *-*-solaris2*.
19657 (HAVE_LD_AS_NEEDED): Update comment.
19658 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19659 * configure: Regenerate.
19660 * config.in: Regenerate.
19661 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19662 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19663 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19664 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19665 equivalents. Fix markup.
19666 * doc/tm.texi: Regenerate.
19667
e0ea8797
AH		 196682013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19669
19670 * config/i386/i386.opt: New option mstack-protector-guard=.
19671 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19672 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19673 TARGET_SSP_TLS_GUARD.
19674 * config/i386/i386.c (ix86_option_override_internal): Set
19675 ix86_stack_protector_guard.
19676 * config/i386/i386.md (stack_protect_set): Enable for
19677 TARGET_SSP_TLS_GUARD only.
19678 (stack_protect_set_<mode>): Ditto.
19679 (stack_protect_test): Ditto.
19680 (stack_protect_test_<mode>): Ditto.
19681 * doc/invoke.texi (i386 Option): Document.
19682
811b72f9
EB		 196832013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19684
19685 PR target/56890
19686 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19687 (S_MODES): Set H_MODE bit.
19688 (SF_MODES): Set only S_MODE and SF_MODE bits.
19689 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19690 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19691 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19692 <MODE_FLOAT>: Likewise.
19693
5529fdd6
JY		 196942013-04-15 Joey Ye <joey.ye@arm.com>
19695
19696 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19697
517b1da2
JY		 196982013-04-15 Joey Ye <joey.ye@arm.com>
19699
19700 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19701 for real far jump.
19702 (thumb_far_jump_used_p): Count instruction size and set
19703 far_jump_used.
19704
01007ae0
EB		 197052013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19706
19707 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19708 * resource.c (mark_target_live_regs): Reformat conditional block.
19709
c46f6580
SB		 197102013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19711
19712 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19713 notes, they are emitted only just before final.
19714 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19715
1f397f45
SB		 197162013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19717
19718 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19719 * cfgrtl.c (delete_insn): Call it here instead.
19720 * lra-spills.c (lra_final_code_change): Use delete_insn.
19721 * haifa-sched.c (sched_remove_insn): Likewise.
19722 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19723 returning to the nop pool.
19724 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19725 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19726
58a51369
SB		 197272013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19728
19729 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19730 * doc/tm.texi: Regenerated.
19731
33159866
UB		 197322013-04-12 Uros Bizjak <ubizjak@gmail.com>
19733
19734 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19735 QImode checks.
19736
226e378f
SB		 197372013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19738
19739 * df-core.c (df_find_def): Compare register numbers.
19740 (df_find_use): Likewise.
19741
fafb9b18
VM		 197422013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19743
19744 PR target/56903
19745 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19746 lra_in_progress for return.
19747
9a946fd6
GY		 197482013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19749
19750 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19751 define_insn into define_insn_and_split and emit movsicc patterns.
19752
41b83758
GY		 197532013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19754
19755 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19756
d3afd9aa
RB		 197572013-04-12 Richard Biener <rguenther@suse.de>
19758
19759 * tree-pass.h (TODO_do_not_ggc_collect): New.
19760 * passes.c (execute_one_ipa_transform_pass): Honor
19761 TODO_do_not_ggc_collect.
19762 (execute_one_pass): Likewise.
19763
19764 Revert
19765 2013-04-10 Richard Biener <rguenther@suse.de>
19766
19767 * passes.c (init_optimization_passes): Remove reload pass.
19768 * ira.c (do_reload): Merge into ...
19769 (ira): ... this.
19770 (rest_of_handle_reload): Remove.
19771 (pass_reload): Likewise.
19772 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19773 of reload for vzeroupper pass placement.
19774
06f9b387
JJ		 197752013-04-12 Jakub Jelinek <jakub@redhat.com>
19776
19777 PR tree-optimization/56918
19778 PR tree-optimization/56920
19779 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19780 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19781 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19782 use wide_mul_with_sign method.
19783
953094d2
RB		 197842013-04-12 Richard Biener <rguenther@suse.de>
19785
19786 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19787 not be considered a gimple constant.
19788
bb506982
MG		 197892013-04-12 Marc Glisse <marc.glisse@inria.fr>
19790
19791 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19792 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19793 operations.
19794
4b84d650
JJ		 197952013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19796 Jakub Jelinek <jakub@redhat.com>
19797
19798 * opts.c: Include diagnostic-color.h.
19799 (common_handle_option): Handle OPT_fdiagnostics_color_.
19800 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19801 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19802 (diagnostic-color.o): New.
19803 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19804 (diagnostic_color_rule): New enum.
19805 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19806 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19807 the location string.
19808 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19809 either NULL, or color kind.
19810 * diagnostic-color.c: New file.
19811 * diagnostic-color.h: New file.
19812 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19813 arguments.
19814 * doc/invoke.texi (-fdiagnostics-color): Document.
19815 * pretty-print.h (pp_show_color): Define.
19816 (struct pretty_print_info): Add show_color field.
19817 * diagnostic.c: Include diagnostic-color.h.
19818 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19819 macros. Colorize error:, warning: etc. strings and also the location
19820 string.
19821 (diagnostic_show_locus): Colorize the caret line.
19822 * pretty-print.c: Include diagnostic-color.h.
19823 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19824 inside of %< %> quotes or quoted through q format modifier.
19825
067a1e71
AK		 198262013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19827
33159866 19828 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 19829
33159866 198302013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 19831
19832 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19833 code in CC_NZ mode.
19834 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19835 pattern.
19836
7b55f98f
MP		 198372013-04-11 Marek Polacek <polacek@redhat.com>
19838
19839 PR tree-optimization/48184
33159866 19840 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 19841
966b587e
EB		 198422013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19843
19844 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19845 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19846 (skip_simple_arithmetic): Tidy up.
19847 * tree.h (skip_simple_constant_arithmetic): Declare.
19848
33159866 198492013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 19850
19851 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19852
1d42c1ec
RB		 198532013-04-11 Richard Biener <rguenther@suse.de>
19854
19855 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19856 generate vector constants.
19857
4ba5ea11
RB		 198582013-04-11 Richard Biener <rguenther@suse.de>
19859
19860 PR tree-optimization/56878
19861 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19862 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19863 New function.
19864 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19865 Prefer to align the DR with the most invariant base address.
19866
f0defe58
SKS		 198672013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19868
19869 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19870 comment.
19871
d07458be
JG		 198722013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19873
19874 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19875 floating-point vector comparisons against 0.
19876
146b8692
JJ		 198772013-04-11 Jakub Jelinek <jakub@redhat.com>
19878
19879 PR tree-optimization/56899
19880 * fold-const.c (extract_muldiv_1): Apply distributive law
19881 only if TYPE_OVERFLOW_WRAPS (ctype).
19882
b8578ff7
BC		 198832013-04-11 Bin Cheng <bin.cheng@arm.com>
19884
19885 PR target/56124
19886 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19887 loading has side effect.
19888
0ea8a6f9
SB		 198892013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19890
19891 * config/sparc/sparc.c: Include tree-pass.h.
19892 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19893 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19894 head of file. Change return type. Split off gate function.
19895 (sparc_gate_work_around_errata): New function.
19896 (pass_work_around_errata): New pass definition.
19897 (insert_pass_work_around_errata) New pass insert definition to
19898 insert pass_work_around_errata just after delayed-branch scheduling.
19899 (sparc_option_override): Insert the pass.
19900 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19901
42e37616
DM		 199022013-04-10 David S. Miller <davem@davemloft.net>
19903
89deeb3b
DM		 19904 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19905 or -mcpu=hypersparc.
19906
42e37616
DM		 19907 * target.def (cstore_mode): New hook.
19908 * target.h: Include insn-codes.h
19909 * targhooks.c: Likewise.
19910 (default_cstore_mode): New function.
19911 * targhooks.h: Declare it.
19912 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19913 * doc/tm.texi: Rebuild.
19914 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19915 target hook, rather than inspecting the insn_data.
19916 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19917 (TARGET_CSTORE_MODE): Redefine.
19918 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19919 result patterns.
19920 * config/sparc/predicates.md (cstore_result_operand): New special
19921 predicate.
19922 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19923 Use it for operand 0.
19924 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19925 (*snesi_special): Likewise.
19926 (*snesi_zero): Likewise.
19927 (*seqsi_zero): Likewise.
19928 (*sltu_insn): Likewise.
19929 (*sgeu_insn): Likewise.
19930 (*seqdi_special): Make operand 0 and comparison operation be of
19931 DImode.
19932 (*snedi_special): Likewise.
19933 (*snedi_special_vis3): Likewise.
19934 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
19935 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
19936 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
19937 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
19938 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
19939 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
19940 (*sltu_extend_sp64): Likewise.
19941 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
19942 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
19943 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
19944 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
19945 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
19946
95ca411e
YZ		 199472013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
19948
19949 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
19950 (aarch64_start_file): Use the new function.
19951
6782438d 199522013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 19953 Jason Merrill <jason@redhat.com>
6782438d
SKS		 19954
19955 * common.opt: Add -gdwarf.
19956 * opts.c (common_handle_option): Handle it.
19957 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
19958
bb313b93
RB		 199592013-04-10 Richard Biener <rguenther@suse.de>
19960
19961 * passes.c (execute_todo): Do not call ggc_collect conditional here.
19962 (execute_one_ipa_transform_pass): But unconditionally here.
19963 (execute_one_pass): And here.
19964 (init_optimization_passes): Remove reload pass.
19965 * tree-pass.h (TODO_ggc_collect): Remove.
19966 (pass_reload): Likewise.
19967 * ira.c (do_reload): Merge into ...
19968 (ira): ... this.
19969 (rest_of_handle_reload): Remove.
19970 (pass_reload): Likewise.
19971 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19972 of reload for vzeroupper pass placement.
19973 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
19974 and todo_flags_finish of all passes.
19975
793d9a16
RB		 199762013-04-10 Richard Biener <rguenther@suse.de>
19977
19978 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
19979 first_const_oprnd field, rename first_def_type to first_op_type.
19980 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
19981 (vect_get_and_check_slp_defs): Always use the type of the
19982 operand. Allow mixed vect_external_def, vect_constant_def types.
19983 (vect_get_constant_vectors): Handle mixed vect_external_def,
19984 vect_constant_def types.
19985
12211b99 199862013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 19987
19988 PR tree-optimization/55524
19989 * tree-ssa-math-opts.c
19990 (convert_mult_to_fma): Don't use an fms construct
19991 when we don't have an fms operation, but fnma, and it looks
19992 likely that we'll be able to use the latter.
19993
12211b99 199942013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 19995
19996 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
19997 function.
19998 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
19999 inline fail caused by overwritable functions.
20000
34ab4a5b
CJW		 200012013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20002
20003 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20004 unnecessary bits in the constant power of two case.
20005
abf9bfbc
RB		 200062013-04-10 Richard Biener <rguenther@suse.de>
20007
20008 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20009 broken code swapping operands.
20010 (vect_build_slp_tree): Do not compute load permutations here.
20011 (vect_analyze_slp_instance): Compute load permutations here,
20012 after building the SLP tree.
20013
f408477e
CB		 200142013-04-09 Christian Bruel <christian.bruel@st.com>
20015
20016 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20017 of next/prev_real_insn.
20018
5ac42672
JH		 200192013-04-09 Jan Hubicka <jh@suse.cz>
20020
abf9bfbc
RB		 20021 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20022 Drop aliased parameter.
5ac42672
JH		 20023 (function_and_variable_visibility): Do not handle alias pairs.
20024 * cgraph.c (varpool_externally_visible_p): Update prototype.
20025 * varpool.c (varpool_add_new_variable): Update.
20026
5017f1d2
KT		 200272013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20028
20029 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20030
48eecbee
SB		 200312013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20032
a949cf1c
SB		 20033 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20034
48eecbee
SB		 20035 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20036
75ef8e3d
MP		 200372013-04-09 Marek Polacek <polacek@redhat.com>
20038
20039 PR tree-optimization/48762
33159866 20040 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 20041
23847df4
RB		 200422013-04-09 Richard Biener <rguenther@suse.de>
20043
20044 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20045 dealing with cost.
20046 (vect_build_slp_tree): Likewise.
20047 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20048 calculating the cost of a SLP instance.
20049 (vect_analyze_slp_instance): Use it from here, after building
20050 the SLP tree.
20051
ea3a0fde
JJ		 200522013-04-09 Jakub Jelinek <jakub@redhat.com>
20053
20054 PR middle-end/56883
20055 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20056 expand_omp_for_static_chunk): Use simple_p = true in
20057 force_gimple_operand_gsi calls when assigning to addressable decls.
20058
a32dfe9d
JL		 200592013-04-09 Jeff Law <law@redhat.com>
20060
20061 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20062 when the boolean was created by converting a wider object which
20063 had a boolean range.
20064
d755c7ef
RB		 200652013-04-09 Richard Biener <rguenther@suse.de>
20066
20067 * tree-vectorizer.h (slp_void_p): Remove.
20068 (slp_tree): Typedef before _slp_tree declaration.
20069 (struct _slp_tree): Use a vector of slp_tree as children.
20070 (vect_get_place_in_interleaving_chain): Remove.
20071 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20072 Move ...
20073 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20074 and make static.
20075 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20076 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20077 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20078 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20079 Use slp_node instead of slp_void_p and adjust.
20080
3d741091
RB		 200812013-04-09 Richard Biener <rguenther@suse.de>
20082
20083 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20084 work that is not necessary.
20085
39307ba7
JJ		 200862013-04-09 Jakub Jelinek <jakub@redhat.com>
20087
20088 PR tree-optimization/56854
20089 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20090 forward into clobber stmts if it would change MEM_REF lhs into
20091 non-MEM_REF.
20092
343881fd
MK		 200932013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20094
20095 * tree.c (type_hash_lookup, type_hash_add): Make static.
20096 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20097
3922658a
RB		 200982013-04-09 Richard Biener <rguenther@suse.de>
20099
20100 * tree.h (unsave_expr_now): Remove.
20101 * tree-inline.c (mark_local_for_remap_r): Remove.
20102 (unsave_expr_1): Likewise.
20103 (unsave_r): Likewise.
20104 (unsave_expr_now): Likewise.
20105 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20106 (propagate_tree_value): Likewise.
20107
9fb6b620
SB		 201082013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20109
20110 * doc/rtl.texi (sequence): Rewrite documentation to match the
20111 current use of SEQUENCE rtl objects.
20112 * rtl.def (SEQUENCE): Likewise.
20113
20114 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20115 Update documentation.
20116 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20117 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20118
20119 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20120
8ddb5a29
TJ		 201212013-04-08 Teresa Johnson <tejohnson@google.com>
20122
20123 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20124 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 20125 methods.
8ddb5a29
TJ		 20126 (estimate_edge_size_and_time): Add comment to suggest using rounding
20127 methods.
20128 (estimate_node_size_and_time): Ditto.
20129 (remap_edge_change_prob): Use helper rounding divide methods.
20130 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20131 (gimple_mod_pow2_value_transform): Ditto.
20132 (gimple_mod_subtract_transform): Ditto.
20133 (gimple_ic_transform): Ditto.
20134 (gimple_stringops_transform): Ditto.
20135 * stmt.c (conditional_probability): Ditto.
20136 (emit_case_dispatch_table): Ditto.
20137 * lto-cgraph.c (merge_profile_summaries): Ditto.
20138 * tree-optimize.c (execute_fixup_cfg): Ditto.
20139 * cfgcleanup.c (try_forward_edges): Ditto.
20140 * cfgloopmanip.c (scale_loop_profile): Ditto.
20141 (loopify): Ditto.
20142 (duplicate_loop_to_header_edge): Ditto.
20143 (lv_adjust_loop_entry_edge): Ditto.
20144 * tree-vect-loop.c (vect_transform_loop): Ditto.
20145 * profile.c (compute_branch_probabilities): Ditto.
20146 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20147 * lto-streamer-in.c (input_cfg): Ditto.
20148 * gimple-streamer-in.c (input_bb): Ditto.
20149 * ipa-cp.c (update_profiling_info): Ditto.
20150 (update_specialized_profile): Ditto.
20151 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20152 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 20153 rounding methods.
8ddb5a29
TJ		 20154 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20155 (compute_trg_info): Ditto.
20156 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20157 (purge_dead_edges): Ditto.
20158 * loop-unswitch.c (unswitch_loop): Ditto.
20159 * cgraphclones.c (cgraph_clone_edge): Ditto.
20160 (cgraph_clone_node): Ditto.
20161 * tree-inline.c (copy_bb): Ditto.
20162 (copy_edges_for_bb): Ditto.
20163 (initialize_cfun): Ditto.
20164 (copy_cfg_body): Ditto.
20165 (expand_call_inline): Ditto.
20166
661e6bd7
KT		 201672013-04-08 Kai Tietz <ktietz@redhat.com>
20168
20169 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20170 TARGET_CYGWIN64 by TARGET_64BIT.
20171
105766f3
JR		 201722013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20173
20174 * config/epiphany/epiphany.md (GPR_1): New constant.
20175 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20176 * config/epiphany/epiphany.c (gen_compare_reg):
20177 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20178 is already in place.
20179 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20180 Don't require being called during rtl expansion; If y operlaps r0,
20181 return 0.
20182 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20183 (epiphany_expand_epilogue): Likewise.
20184
8afab237
JR		 20185 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20186 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20187 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20188
fa7707d6
JR		 20189 * config/epiphany/constraints.md (CnL): New constraint.
20190 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20191 * config/epiphany/predicates.md (add_operand): Allow 1024.
20192
5138e18d
JR		 20193 * config/epiphany/epiphany.md (logical_op): New code iterator.
20194 (op_mnc): New code attribute.
20195 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20196 (mov_f+1, mov_f+2): New peephole2 patterns.
20197
2ccc703d
JR		 20198 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20199 (cstoresi4): Also allow re-use of zero result when doing a NE
20200 comparison to a non-zero operand.
aefb0819 20201 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20202
093ac0a5
JR		 20203 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20204 Use gen_addsi3_i / gen_subsi3_i.
20205
f223bb13
JJ		 202062013-04-08 Jakub Jelinek <jakub@redhat.com>
20207
20208 PR c++/34949
20209 PR c++/50243
20210 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20211 contain anything but clobbers, at most one __builtin_stack_restore,
20212 optionally debug stmts and final resx, and if it has at least one
20213 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20214 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20215 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20216 which isn't defaut definition, remove them.
20217 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20218 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20219 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20220 with MEM_REF LHS with SSA_NAME address.
20221
4481581f
JL		 202222013-04-08 Jeff Law <law@redhat.com>
20223
20224 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20225
451dabda
RB		 202262013-04-08 Richard Biener <rguenther@suse.de>
20227
20228 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20229 extra newline.
20230 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20231 determined vector type.
20232 (vect_analyze_data_refs): Likewise.
20233 (vect_get_new_vect_var): Adjust.
20234 (vect_create_destination_var): Preserve SSA name versions.
20235 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20236 not dump anything here.
20237
3b088b47
JR		 202382013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20239
20240 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20241 Add member lr_slot_known.
20242 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20243 if necessary.
20244 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20245 Remove code that sets lr_slot_offset according to what a previous
20246 version of epiphany_emit_save_restore used to do.
20247 (epiphany_emit_save_restore): When doing an lr save or restore,
20248 set/verify lr_slot_known and lr_slot_offset.
20249
d8484d41
XQ		 202502013-04-08 Xinyu Qi <xyqi@marvell.com>
20251
33159866
UB		 20252 PR target/54338
20253 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 20254 in ALL_REGS.
20255
481be1c4
RB		 202562013-04-08 Richard Biener <rguenther@suse.de>
20257
20258 * alias.c (find_base_term): Fix thinko in previous change.
20259
401f3a81
JJ		 202602013-04-08 Jakub Jelinek <jakub@redhat.com>
20261
20262 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20263 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20264 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20265 if possible to compute val.
20266 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20267 For QImode integers don't require anything about precision. Use
20268 const_with_all_bytes_same to find out if the constant doesn't have
20269 repeated bytes in it.
20270
03ed99a8
AK		 202712013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20272
20273 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20274 within mode size.
20275
781b2e62
MP		 202762013-04-08 Marek Polacek <polacek@redhat.com>
20277
20278 PR rtl-optimization/48182
20279 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20280 value to 1.
20281
27e430a2
JDA		 202822013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20283
20284 PR target/55487
20285 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20286 nuses, make sure we have a label.
20287
4902aa64
BS		 202882013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20289
20290 PR target/56843
20291 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20292 (rs6000_emit_swdiv_low_precision): Remove.
20293 (rs6000_emit_swdiv): Rewrite to handle between one and four
20294 iterations of Newton-Raphson generally; modify required number of
20295 iterations for some cases.
20296 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20297
7bca81dc
SB		 202982013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20299
20300 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20301 set-but-unused variable.
20302
20303 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20304 basic blocks of released function bodies garbage-collectable.
20305
20306 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20307 (struct rtl_opt_pass): Add TODO_df_finish.
20308
20309 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20310
4542a38a
GY		 203112013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20312
20313 * config/arm/constraints.md (q): New constraint.
20314 * config/arm/ldrdstrd.md: New file.
20315 * config/arm/arm.md (ldrdstrd.md) New include.
20316 (arm_movdi): Use "q" instead of "r" constraint
20317 for double-word memory access.
20318 (movdf_soft_insn): Likewise.
20319 * config/arm/vfp.md (movdi_vfp): Likewise.
20320 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20321 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 20322 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20323 (mem_ok_for_ldrd_strd): Likewise.
20324 (output_move_double): Update assertion.
20325
2385b218
GY		 203262013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20327
20328 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20329
75fe1cb5
GY		 203302013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20331
20332 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20333 define_insn_and_split.
20334 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20335
dd660e8e
GY		 203362013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20337
20338 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20339 define_insn_and_split.
33159866 20340 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 20341 (shiftsi3_compare): New pattern.
20342 (rrx): New pattern.
20343 * config/arm/unspecs.md (UNSPEC_RRX): New.
20344
045e472c
GY		 203452013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20346
20347 * config/arm/arm.md (negdi_extendsidi): New pattern.
20348 (negdi_zero_extendsidi): Likewise.
20349
3f3bf1a8
GY		 203502013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20351
20352 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20353 define_insn_and_split.
20354 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20355 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20356
b0b49556
GY		 203572013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20358
20359 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20360 define_insn_and_split.
20361 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20362 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20363
d633dd84
GY		 203642013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20365
20366 * config/arm/arm.md (subsi3_carryin): New pattern.
20367 (subsi3_carryin_const): Likewise.
20368 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20369 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20370
f4499066
GY		 203712013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20372
20373 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20374
ceef6fd9
GY		 203752013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20376
20377 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20378 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20379
ddbdd8a7
KT		 203802013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20381
20382 * config/arm/arm.c (arm_expand_builtin): Change fcode
20383 type to unsigned int.
20384
8456d78a
RR		 203852013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20386
20387 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20388
526c230f
ILT		 203892013-04-04 Ian Lance Taylor <iant@google.com>
20390
33159866
UB		 20391 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20392 language standard.
526c230f 20393
3cfbe04d
SB		 203942013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20395
20396 PR middle-end/56729
20397 * df-scan.c (df_insn_delete): Disable failing assert.
20398
dfa3f8d0
KT		 203992013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20400
20401 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20402 New function prototype.
20403 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20404 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20405 (arm_builtin_vectorized_function): New function.
20406
f7837758
KT		 204072013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20408
20409 * config/arm/arm_neon_builtins.def: New file.
20410 * config/arm/arm.c (neon_builtin_data): Move contents to
20411 arm_neon_builtins.def.
20412 (enum arm_builtins): Include neon builtin definitions.
20413 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 20414 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 20415
39c1b6db
MP		 204162013-04-04 Marek Polacek <polacek@redhat.com>
20417
20418 PR tree-optimization/48186
20419 * predict.c (maybe_hot_frequency_p): Return false if
20420 HOT_BB_FREQUENCY_FRACTION is 0.
20421 (cgraph_maybe_hot_edge_p): Likewise.
20422
314f64eb
RB		 204232013-04-04 Richard Biener <rguenther@suse.de>
20424
20425 PR tree-optimization/56826
20426 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20427 more accurately.
20428
51a905b2
RB		 204292013-04-04 Richard Biener <rguenther@suse.de>
20430
20431 PR tree-optimization/56213
20432 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 20433 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 20434
f1bf4f3a
RB		 204352013-04-04 Richard Biener <rguenther@suse.de>
20436
20437 PR tree-optimization/56837
20438 * tree-loop-distribution.c (classify_partition): For non-zero
20439 values require that the value has the same precision as its
20440 mode to be useful as memset value.
20441
0bca7ded
NC		 204422013-04-03 Nick Clifton <nickc@redhat.com>
20443
33159866 20444 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 20445 (fmssf4): Use fmsf.s on E3V5 architectures.
20446 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20447 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20448
b4019227
JL		 204492013-04-03 Jeff Law <law@redhat.com>
20450
20451 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20452 (lra-eliminations.o): Likewise.
20453
f57ddb5b
TJ		 204542013-04-03 Teresa Johnson <tejohnson@google.com>
20455
20456 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 20457 compute_working_sets here from profile.c.
f57ddb5b 20458 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 20459 (gcov_working_set_t): Moved typedef here from basic-block.h
20460 (compute_working_set): Declare.
f57ddb5b
TJ		 20461 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20462 (get_working_sets): Renamed from compute_working_set,
33159866 20463 replace most of body with call to new compute_working_sets.
f57ddb5b 20464 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 20465 to get_working_sets.
20466 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 20467 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 20468 to get_working_sets.
f57ddb5b
TJ		 20469 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20470 * gcov-dump.c (dump_working_sets): New function.
20471
12211b99 204722013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 20473
20474 * hwint.c (sext_hwi, zext_hwi): New functions.
20475 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20476 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20477 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20478 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20479 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20480 (sext_hwi, zext_hwi): New functions.
0bca7ded 20481
be672e08
JL		 204822013-04-03 Jeff Law <law@redhat.com>
20483
20484 PR tree-optimization/56799
33159866
UB		 20485 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20486 back test for widening conversion erroneously dropped in prior change.
be672e08 20487
9d821fa5
KT		 204882013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20489
20490 PR target/56809
20491 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20492 instead of next_real_insn.
20493
71c581e7
MP		 204942013-04-03 Marek Polacek <polacek@redhat.com>
20495
20496 PR sanitizer/55702
33159866 20497 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 20498
4a32ef80
KT		 204992013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20500
20501 PR target/56809
20502 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20503 next_real_insn.
20504 (thumb1_output_casesi): Likewise.
20505 (thumb2_output_casesi): Likewise.
20506
1b2253d4
RB		 205072013-04-03 Richard Biener <rguenther@suse.de>
20508
20509 PR tree-optimization/56817
20510 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20511 Split out ...
20512 (tree_unroll_loops_completely_1): ... new function to manually
20513 walk the loop tree, properly defering outer loops of unrolled
20514 loops to later iterations.
20515
38000232
MG		 205162013-04-03 Marc Glisse <marc.glisse@inria.fr>
20517
20518 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20519 (vectorizable_load): Likewise.
20520 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20521 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20522
3fa35298
MG		 205232013-04-03 Marc Glisse <marc.glisse@inria.fr>
20524
20525 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20526 BIT_FIELD_REF.
20527
b3d45ff0
UW		 205282013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20529
20530 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20531
ec9202a8
BC		 205322013-04-03 Bin Cheng <bin.cheng@arm.com>
20533
20534 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20535
6805bd36
MG		 205362013-04-03 Marc Glisse <marc.glisse@inria.fr>
20537
20538 PR tree-optimization/56790
33159866
UB		 20539 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20540 folding.
6805bd36 20541
da694a77
MG		 205422013-04-03 Marc Glisse <marc.glisse@inria.fr>
20543
20544 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20545 Handle VEC_MERGE.
20546 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20547 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20548 equal arguments.
20549
4e7d7b3d
JJ		 205502013-04-03 Jakub Jelinek <jakub@redhat.com>
20551
20552 PR c/19449
20553 * tree.h (force_folding_builtin_constant_p): New decl.
20554 * builtins.c (force_folding_builtin_constant_p): New variable.
20555 (fold_builtin_constant_p): Fold immediately also if
20556 force_folding_builtin_constant_p.
20557
e6c9d234
RB		 205582013-04-03 Richard Biener <rguenther@suse.de>
20559
20560 PR tree-optimization/56812
20561 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20562 DRs of the same interleaving chain are independent.
20563
984d07dd
JM		 205642013-04-02 Jason Merrill <jason@redhat.com>
20565
20566 * gdbinit.in (pbb): Use debug fn.
20567
622849c9
LC		 205682013-04-02 Lawrence Crowl <crowl@google.com>
20569
20570 * sese.h (struct ivtype_map_elt_s): Remove unused.
20571 (extern debug_ivtype_map): Remove unused.
20572 (extern eq_ivtype_map_elts): Remove unused.
20573 * sese.c (debug_ivtype_map): Removed unused.
20574 (debug_ivtype_map_1): Removed unused.
20575 (debug_ivtype_elt): Remove unused.
20576 (eq_ivtype_map_elts): Remove unused.
20577
20578
82c0e1a0
KT		 205792013-04-02 Kai Tietz <ktietz@redhat.com>
20580
20581 PR target/52790
20582 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20583 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 20584 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20585 function.
82c0e1a0
KT		 20586 (legitimize_pe_coff_symbol): Likewise.
20587 (is_imported_p): New helper-function.
20588 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20589 for Windows x64 targets.
20590 (ix86_expand_prologue): Optimize for pe-coff targets.
20591 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20592 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20593 medium/large code-model.
20594 (legitimize_pic_address): Likewise.
20595 (legitimize_tls_address): Likewise.
20596 (ix86_expand_call): Likewise.
20597 (x86_output_mi_thunk): Likewise.
20598 (get_dllimport_decl): Add new beimport argument.
20599 (construct_plt_address): Don't assert for x64 pe-coff targets.
20600 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20601 targets.
20602 (SYMBOL_FLAG_STUBVAR): New macro.
20603 (SYMBOL_REF_STUBVAR_P): Likewise.
20604 * config/i386/winnt.c (stub_list): New structure.
20605 (stub_head): New local variable.
20606 (i386_pe_record_stub): New function.
20607 (i386_pe_file_end): Emit refptr-stubs.
20608
5d751b0c
JJ		 206092013-04-02 Jakub Jelinek <jakub@redhat.com>
20610
09bb4c99
JJ		 20611 PR rtl-optimization/56745
20612 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20613 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20614
a1d8947a
JJ		 20615 PR c++/34949
20616 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20617 and both of them are MEM_REFs, just compare first argument for
20618 equality and attempt to deal even with differing offsets.
20619
5d751b0c
JJ		 20620 PR c++/34949
20621 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20622 of gimple_clobber_p to be MEM_REF.
20623 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20624 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20625 after gimplification.
20626 * asan.c (get_mem_ref_of_assignment): Don't instrument
20627 gimple_clobber_p stmts.
20628 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20629 gimple_clobber_p stmt if they have MEM_REF lhs and
20630 are dead because of another gimple_clobber_p stmt.
20631 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20632 gimple_clobber_p stmts like debug stmts.
20633 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20634 that refer to unused VAR_DECLs or uninitialized values.
20635 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20636 gimple_clobber_p stmts if they refer to removed parameters.
20637 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20638 formatting.
20639
e594716a
UB		 206402013-04-02 Uros Bizjak <ubizjak@gmail.com>
20641
20642 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20643 using SWI48 mode attribute.
20644
7a80735b
WM		 206452013-04-02 Wei Mi <wmi@google.com>
20646
20647 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20648 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20649 *<rotate_insn><mode>3_mask in i386.md.
20650
f423a9e4
AI		 206512013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20652
20653 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20654
90eb75f2
RB		 206552013-04-02 Richard Biener <rguenther@suse.de>
20656
20657 PR tree-optimization/56778
20658 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20659 Runtime alias tests are not supported for gather loads.
20660 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20661 stmts referenced from SSA operands before updating SSA form.
20662
d8c69a92
IC		 206632013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20664 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20665
20666 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 20667 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20668 * config/arm/cortex-a53.md: New file.
20669 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20670 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20671 * config/arm/arm.c (arm_issue_rate): Likewise.
20672 * config/arm/arm-tune.md: Regenerate
20673 * config/arm/arm-tables.opt: Regenerate.
20674 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 20675
239eb04c
ZC		 206762013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20677
20678 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20679 non-static link.
20680
c902d3c8
SN		 206812013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20682
d8c69a92
IC		 20683 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20684 scalar load/store operations using B/H registers.
c902d3c8
SN		 20685 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20686
051d0e2f
SN		 206872013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20688
20689 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20690 scalar move.
20691 * config/aarch64/aarch64.c
20692 (aarch64_simd_scalar_immediate_valid_for_move): New.
20693 * config/aarch64/aarch64-protos.h
20694 (aarch64_simd_scalar_immediate_valid_for_move): New.
20695 * config/aarch64/constraints.md (Dh, Dq): New.
20696 * config/aarch64/iterators.md (hq): New.
20697
0ee1e3d9
EB		 206982013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20699
20700 * reorg.c (get_branch_condition): Deal with conditional returns.
20701 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20702
136fb3f7
WM		 207032013-04-01 Wei Mi <wmi@google.com>
20704
20705 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20706 Truncate operand 2 using %b asm operand modifier.
20707 (*<shift_insn><mode>3_mask): Ditto.
20708 (*<rotate_insn><mode>3_mask): Ditto.
20709
6388c738
SB		 207102013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20711
20712 PR middle-end/56798
20713 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20714
f7a4d826
KK		 207152013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20716
20717 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20718 of next_real_insn.
20719 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20720
4b943a49
LC		 207212013-03-30 Lawrence Crowl <crowl@google.com>
20722
20723 * dse.c (clear_alias_sets): Remove never set.
20724 (disqualified_clear_alias_sets): Remove never set.
20725 (clear_alias_mode_pool): Remove never set.
20726 (dse_step0): Remove condition that is never true.
20727 (canon_address): Remove condition that is never true.
20728 (dse_step7): Remove condition that is never true.
20729 (rest_of_handle_dse): Remove condition that is never true.
20730 (rest_of_handle_dse::did_global): Remove never read from above.
20731 (dse_step2_spill): Remove never called from above.
20732 (dse_step5_spill): Remove never called from above.
20733
39718607
SB		 207342013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20735
da5c6bde
SB		 20736 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20737 JUMP_TABLE_DATA changes.
20738 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20739 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20740 (Insns) <jump_table_data>: New entry.
20741 * doc/tm.texi: Regenerate.
20742
39718607
SB		 20743 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20744
20745 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20746 for table jump at the end of a basic block using tablejump_p.
20747 * targhooks.c (default_invalid_within_doloop): Likewise.
20748 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20749 target hook implementation that is identical to the default hook.
20750 (rs6000_invalid_within_doloop): Remove.
20751
20752 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20753 unused variable from tablejump_p call.
20754
20755 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20756 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20757 (INSN_DELETED_P): Likewise.
20758 (emit_jump_table_data): New prototype.
20759 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20760 after 4th as unused.
20761 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20762 * sched-vis.c (print_insn): Likewise.
20763 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20764 insn for compatibility with back ends that use next_active_insn to
20765 identify jump table data.
20766 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20767 (remove_insn): Likewise.
20768 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20769 to be emitted.
20770 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20771 (emit_jump_table_data): New function.
20772
20773 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20774 basic block, a JUMP_TABLE_DATA never is.
20775 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20776 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20777 off from code handling real insns.
20778 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20779 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20780 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20781 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20782 is not a NONDEBUG_INSN_P.
20783 * ira-costs.c (scan_one_insn): Likewise.
20784 * jump.c (mark_all_labels): Likewise.
20785 (mark_jump_label_1): Likewise.
20786 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20787 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20788 a basic block.
20789 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20790 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20791 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20792 JUMP_TABLE_DATA_P insns.
20793 (calculate_elim_costs_all_insns): Likewise.
20794 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20795 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20796 (delete_output_reload): Code style fixups.
20797 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20798 insn flags on this non-insn.
20799 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20800 as scheduling barriers, for pre-change compatibility.
20801 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20802 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20803
20804 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20805 redundant JUMP_TABLE_DATA_P test.
20806 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20807 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20808 (frv_for_each_packet): Likewise.
20809 * config/i386/i386.c (min_insn_size): Likewise.
20810 (ix86_avoid_jump_mispredicts): Likewise.
20811 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20812 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20813 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20814 (mips16_insn_length): Robustify.
20815 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20816 (mips16_split_long_branches): Likewise.
20817 * config/pa/pa.c (pa_combine_instructions): Likewise.
20818 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20819 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20820 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20821 as contributing to pool range lengths.
20822 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20823 Remove redundant JUMP_TABLE_DATA_P test.
20824 (sh_loop_align): Likewise.
20825 (split_branches): Likewise.
20826 (sh_insn_length_adjustment): Likewise.
20827 * config/spu/spu.c (get_branch_target): Likewise.
20828
0208f7da
JH		 208292013-03-29 Jan Hubicka <jh@suse.cz>
20830
20831 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20832 gcov streaming; stream hot bb threshold to ltrans.
20833 * predict.c (get_hot_bb_threshold): Break out from ....
20834 (maybe_hot_count_p): ... here.
20835 (set_hot_bb_threshold): New function.
20836 * lto-section-in.c (lto_section_name): Add profile.
20837 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20838 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20839 and data-streamer.h
20840 (histogram_entry): New structure.
20841 (histogram, histogram_pool): New global vars.
20842 (histogram_hash): New structure.
20843 (histogram_hash::hash): New method.
20844 (histogram_hash::equal): Likewise.
20845 (account_time_size): New function.
20846 (cmp_counts): New function.
20847 (dump_histogram): New function.
20848 (ipa_profile_generate_summary): New function.
20849 (ipa_profile_write_summary): New function.
20850 (ipa_profile_read_summary): New function.
20851 (ipa_profile): Decide on threshold.
e594716a
UB		 20852 (pass_ipa_profile): Add ipa_profile_write_summary and
20853 ipa_profile_read_summary.
0208f7da
JH		 20854 * Makefile.in (ipa.o): Update dependencies.
20855 * lto-streamer.h (LTO_section_ipa_profile): New section.
20856
5a6ccc94
GDR		 208572013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20858
20859 * tree.h (VAR_P): New.
20860
39385fa6
PC		 208612013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20862
20863 PR lto/56777
20864 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20865
34f0d87a
SB		 208662013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20867
20868 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20869 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20870 (control_flow_insn_p): Likewise.
20871 * cfgrtl.c (duplicate_insn_chain): Likewise.
20872 * final.c (get_attr_length_1): Likewise.
20873 (shorten_branches): Likewise.
20874 (final_scan_insn): Likewise.
20875 * function.c (instantiate_virtual_regs): Likewise.
20876 * gcse.c (insert_insn_end_basic_block): Likewise.
20877 * ira-costs.c (scan_one_insn): Likewise.
20878 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20879 * lra.c (check_rtl): Likewise.
20880 * reload1.c (elimination_costs_in_insn): Likewise.
20881 * reorg.c (follow_jumps): Likewise.
20882
20883 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20884 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20885 (thumb_far_jump_used_p): Likewise.
20886 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20887 (workaround_speculation): Likewise.
20888 (add_sched_insns_for_speculation): Likewise.
20889 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20890 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20891 (frv_for_each_packet): Likewise.
20892 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20893 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20894 (final_emit_insn_group_barriers): Likewise.
20895 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20896 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20897 (mips16_insn_length): Likewise.
20898 * config/pa/pa.c (pa_reorg): Likewise.
20899 (pa_combine_instructions): Likewise.
20900 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20901 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20902 (sh_reorg): Likewise.
20903 (split_branches): Likewise.
20904 * config/spu/spu.c (get_branch_target): Likewise.
20905
20906 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20907 JUMP_TABLE_DATA_P.
20908
4ac761b0
KY		 209092013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20910
39385fa6 20911 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 20912 Fix declaration name.
20913
58a49189
LC		 209142013-03-28 Lawrence Crowl <crowl@google.com>
20915
20916 * graphds.h (struct graph.indicies): Remove unused.
20917 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20918 (SCOP_ORIGINAL_PDDRS): Remove unused.
20919 * sese.h (extern insert_loop_close_phis): Removed unused.
20920 (extern insert_guard_phis): Removed unused.
20921 (extern ivtype_map_elt_info): Removed unused.
20922 (new_ivtype_map_elt): Removed unused.
20923 * sese.c (ivtype_map_elt_info): Removed unused.
20924
7b3b6ae4
LC		 209252013-03-28 Lawrence Crowl <crowl@google.com>
20926
20927 * Makefile.in: Add several missing include dependences.
20928 (DUMPFILE_H): New.
20929 (test-dump.o): New. This object is not added to any executable,
20930 but is present for ad-hoc testing.
20931 * bitmap.c
20932 (debug (const bitmap_head_def &)): New.
20933 (debug (const bitmap_head_def *)): New.
20934 * bitmap.h
20935 (extern debug (const bitmap_head_def &)): New.
20936 (extern debug (const bitmap_head_def *)): New.
20937 * cfg.c
20938 (debug (edge_def &)): New.
20939 (debug (edge_def *)): New.
20940 * cfghooks.c
20941 (debug (basic_block_def &)): New.
20942 (debug (basic_block_def *)): New.
20943 * dumpfile.h
20944 (dump_node (const_tree, int, FILE *)): Correct source file.
20945 * dwarf2out.c
20946 (debug (die_struct &)): New.
20947 (debug (die_struct *)): New.
20948 * dwarf2out.h
20949 (extern debug (die_struct &)): New.
20950 (extern debug (die_struct *)): New.
20951 * gimple-pretty-print.c
20952 (debug (gimple_statement_d &)): New.
20953 (debug (gimple_statement_d *)): New.
20954 * gimple-pretty-print.h
20955 (extern debug (gimple_statement_d &)): New.
20956 (extern debug (gimple_statement_d *)): New.
20957 * ira-build.c
20958 (debug (ira_allocno_copy &)): New.
20959 (debug (ira_allocno_copy *)): New.
20960 (debug (ira_allocno &)): New.
20961 (debug (ira_allocno *)): New.
20962 * ira-int.h
20963 (extern debug (ira_allocno_copy &)): New.
20964 (extern debug (ira_allocno_copy *)): New.
20965 (extern debug (ira_allocno &)): New.
20966 (extern debug (ira_allocno *)): New.
20967 * ira-lives.c
20968 (debug (live_range &)): New.
20969 (debug (live_range *)): New.
20970 * lra-int.h
20971 (debug (lra_live_range &)): New.
20972 (debug (lra_live_range *)): New.
20973 * lra-lives.c
20974 (debug (lra_live_range &)): New.
20975 (debug (lra_live_range *)): New.
20976 * omega.c
20977 (debug (omega_pb_d &)): New.
20978 (debug (omega_pb_d *)): New.
20979 * omega.h
20980 (extern debug (omega_pb_d &)): New.
20981 (extern debug (omega_pb_d *)): New.
20982 * print-rtl.c
20983 (debug (const rtx_def &)): New.
20984 (debug (const rtx_def *)): New.
20985 * print-tree.c
20986 (debug_tree (tree): Move within file.
20987 (debug_raw (const tree_node &)): New.
20988 (debug_raw (const tree_node *)): New.
20989 (dump_tree_via_hooks (const tree_node *, int)): New.
20990 (debug (const tree_node &)): New.
20991 (debug (const tree_node *)): New.
20992 (debug_verbose (const tree_node &)): New.
20993 (debug_verbose (const tree_node *)): New.
20994 (debug_head (const tree_node &)): New.
20995 (debug_head (const tree_node *)): New.
20996 (debug_body (const tree_node &)): New.
20997 (debug_body (const tree_node *)): New.
20998 (debug_vec_tree (tree): Move and reimplement in terms of dump.
20999 (debug (vec<tree, va_gc> &)): New.
21000 (debug (vec<tree, va_gc> *)): New.
21001 * rtl.h
21002 (extern debug (const rtx_def &)): New.
21003 (extern debug (const rtx_def *)): New.
21004 * sbitmap.c
21005 (debug_raw (simple_bitmap_def &)): New.
21006 (debug_raw (simple_bitmap_def *)): New.
21007 (debug (simple_bitmap_def &)): New.
21008 (debug (simple_bitmap_def *)): New.
21009 * sbitmap.h
21010 (extern debug (simple_bitmap_def &)): New.
21011 (extern debug (simple_bitmap_def *)): New.
21012 (extern debug_raw (simple_bitmap_def &)): New.
21013 (extern debug_raw (simple_bitmap_def *)): New.
21014 * sel-sched-dump.c
21015 (debug (vinsn_def &)): New.
21016 (debug (vinsn_def *)): New.
21017 (debug_verbose (vinsn_def &)): New.
21018 (debug_verbose (vinsn_def *)): New.
21019 (debug (expr_def &)): New.
21020 (debug (expr_def *)): New.
21021 (debug_verbose (expr_def &)): New.
21022 (debug_verbose (expr_def *)): New.
21023 (debug (vec<rtx> &)): New.
21024 (debug (vec<rtx> *)): New.
21025 * sel-sched-dump.h
21026 (extern debug (vinsn_def &)): New.
21027 (extern debug (vinsn_def *)): New.
21028 (extern debug_verbose (vinsn_def &)): New.
21029 (extern debug_verbose (vinsn_def *)): New.
21030 (extern debug (expr_def &)): New.
21031 (extern debug (expr_def *)): New.
21032 (extern debug_verbose (expr_def &)): New.
21033 (extern debug_verbose (expr_def *)): New.
21034 (extern debug (vec<rtx> &)): New.
21035 (extern debug (vec<rtx> *)): New.
21036 * sel-sched-ir.h
21037 (_list_iter_cond_expr): Make inline instead of static.
21038 * sreal.c
21039 (debug (sreal &)): New.
21040 (debug (sreal *)): New.
21041 * sreal.h
21042 (extern debug (sreal &)): New.
21043 (extern debug (sreal *)): New.
21044 * tree.h
21045 (extern debug_raw (const tree_node &)): New.
21046 (extern debug_raw (const tree_node *)): New.
21047 (extern debug (const tree_node &)): New.
21048 (extern debug (const tree_node *)): New.
21049 (extern debug_verbose (const tree_node &)): New.
21050 (extern debug_verbose (const tree_node *)): New.
21051 (extern debug_head (const tree_node &)): New.
21052 (extern debug_head (const tree_node *)): New.
21053 (extern debug_body (const tree_node &)): New.
21054 (extern debug_body (const tree_node *)): New.
21055 (extern debug (vec<tree, va_gc> &)): New.
21056 (extern debug (vec<tree, va_gc> *)): New.
21057 * tree-cfg.c
21058 (debug (struct loop &)): New.
21059 (debug (struct loop *)): New.
21060 (debug_verbose (struct loop &)): New.
21061 (debug_verbose (struct loop *)): New.
21062 * tree-dump.c: Add header dependence.
21063 * tree-flow.h
21064 (extern debug (struct loop &)): New.
21065 (extern debug (struct loop *)): New.
21066 (extern debug_verbose (struct loop &)): New.
21067 (extern debug_verbose (struct loop *)): New.
21068 * tree-data-ref.c
21069 (debug (data_reference &)): New.
21070 (debug (data_reference *)): New.
21071 (debug (vec<data_reference_p> &)): New.
21072 (debug (vec<data_reference_p> *)): New.
21073 (debug (vec<ddr_p> &)): New.
21074 (debug (vec<ddr_p> *)): New.
21075 * tree-data-ref.h
21076 (extern debug (data_reference &)): New.
21077 (extern debug (data_reference *)): New.
21078 (extern debug (vec<data_reference_p> &)): New.
21079 (extern debug (vec<data_reference_p> *)): New.
21080 (extern debug (vec<ddr_p> &)): New.
21081 (extern debug (vec<ddr_p> *)): New.
21082 * tree-ssa-alias.c
21083 (debug (pt_solution &)): New.
21084 (debug (pt_solution *)): New.
21085 * tree-ssa-alias.h
21086 (extern debug (pt_solution &)): New.
21087 (extern debug (pt_solution *)): New.
21088 * tree-ssa-alias.c
21089 (debug (_var_map &)): New.
21090 (debug (_var_map *)): New.
21091 (debug (tree_live_info_d &)): New.
21092 (debug (tree_live_info_d *)): New.
21093 * tree-ssa-alias.h
21094 (extern debug (_var_map &)): New.
21095 (extern debug (_var_map *)): New.
21096 (extern debug (tree_live_info_d &)): New.
21097 (extern debug (tree_live_info_d *)): New.
21098
be77e1e5
JH		 210992013-03-28 Jan Hubicka <jh@suse.cz>
21100
21101 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21102
777e6976
IB		 211032013-03-28 Ian Bolton <ian.bolton@arm.com>
21104
21105 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21106 record only when desired or required.
21107
9cd347ae
UB		 211082013-03-28 Uros Bizjak <ubizjak@gmail.com>
21109
21110 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21111 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21112
33159866 211132013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 21114
21115 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21116 (*andsi3_compare0_uxtw): New pattern.
21117 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21118 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21119
89ab31c1
JH		 211202013-03-28 Jan Hubicka <jh@suse.cz>
21121
21122 * data-streamer-in.c (streamer_read_gcov_count): New function.
21123 * gimple-streamer-out.c: Include value-prof.h.
21124 (output_gimple_stmt): Output histogram.
21125 (output_bb): Use streamer_write_gcov_count.
21126 * value-prof.c: Include data-streamer.h
21127 (dump_histogram_value): Add HIST_TYPE_MAX.
21128 (stream_out_histogram_value): New function.
21129 (stream_in_histogram_value): New function.
21130 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21131 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21132 * data-streamer-out.c (streamer_write_gcov_count): New function.
21133 (streamer_write_gcov_count_stream): New function.
21134 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21135 (lto_output_node): Likewise.
21136 (input_node, input_edge): Likewise.
21137 * lto-streamer-out.c (output_cfg): Update streaming.
21138 * lto-streamer-in.c (input_cfg): Likewise.
21139 * data-streamer.h (streamer_write_gcov_count,
21140 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21141 * gimple-streamer-in.c: Include value-prof.h
21142 (input_gimple_stmt): Input histograms.
21143 (input_bb): Update profile streaming.
21144
e594716a 211452013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 21146
33159866 21147 * genmodes.c (emit_max_int): New function.
8697be17 21148 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 21149 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21150 Added doc.
8697be17 21151 * machmode.def: Fixed comment.
89ab31c1 21152
e594716a 211532013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 21154
21155 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 21156 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21157
74946978
MP		 211582013-03-28 Marek Polacek <polacek@redhat.com>
21159 Richard Biener <rguenther@suse.de>
21160
21161 PR tree-optimization/56695
21162 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21163 build signed result of a vector comparison.
21164 * tree-cfg.c (verify_gimple_comparison): Check that a result
21165 of a vector comparison has signed type.
21166
a64b9c26
RB		 211672013-03-28 Richard Biener <rguenther@suse.de>
21168
21169 PR tree-optimization/37021
21170 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21171 do not restrict gaps between groups.
21172 * tree-vect-stmts.c (vectorizable_load): Properly account for
21173 a gap between groups.
21174
a9dc2a2f
EB		 211752013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21176
21177 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21178 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21179 is not enabled.
21180
53cb97f9
GP		 211812013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21182
21183 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21184 * doc/extend.texi (Named Address Spaces): Ditto.
21185 (Variable Attributes): Ditto.
21186
b802ae5c
KT		 211872013-03-27 Kai Tietz <ktietz@redhat.com>
21188
eddae10a
KT		 21189 * config.build: Add support for cygwin x64 target.
21190 * config.gcc: Likewise.
21191 * config.host: Likewise.
21192 * configure.ac: Likewise
21193 * configure: Regenerated.
b802ae5c 21194
371e77e3 211952013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 21196
21197 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21198 * config/i386/t-cygwin-w64: New file.
21199 * config/i386/cygwin-w64.h: New file.
21200 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21201 and add support for x64-cygwin target.
21202 (CPP_SPEC): Likewise.
21203 (CXX_WRAP_SPEC_LIST): Undefine before define.
21204 (LIBGCJ_SONAME): Use 15 as version.
21205
f49b33cb
RB		 212062013-03-27 Richard Biener <rguenther@suse.de>
21207
21208 PR tree-optimization/56716
21209 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21210 dumping for ref nodes.
21211
b37a6ce5
MJ		 212122013-03-27 Martin Jambor <mjambor@suse.cz>
21213
21214 PR tree-optimization/55334
21215 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21216 restricted pointers to arrays.
21217
9469b9b2
GDR		 212182013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21219
21220 * Makefile.in (.SUFFIXES): Add .cc.
21221 (.c.o): Apply same recipe for implicit rule .cc.o.
21222
7d24f650
RB		 212232013-03-27 Richard Biener <rguenther@suse.de>
21224
21225 PR tree-optimization/37021
21226 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21227 REALPART/IMAGPART_EXPRs around the supported refs.
21228 * tree-ssa-structalias.c (find_func_aliases): Assume that
21229 floating-point values are not used to transfer pointers.
21230
2f251a05
AI		 212312013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21232
f43245d1
UB		 21233 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21234 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21235 * doc/tm.texi: Regenerate.
21236 * targhooks.h (default_has_ifunc_p): New.
21237 * targhooks.c (default_has_ifunc_p): Ditto.
21238 * config/linux-protos.h: New file.
21239 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21240 hook for linux which disables support of indirect functions in android.
21241 * config/linux-android.c: New file.
21242 * config/t-linux-android.c: Ditto.
21243 * config.gcc: Added new object file linux-android.o.
21244 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21245 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21246 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 21247 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21248 doesn't support indirect functions.
21249 * configure: Regenerate.
21250
78b4e425
BC		 212512013-03-27 Bin Cheng <bin.cheng@arm.com>
21252
21253 PR target/56102
21254 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21255 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21256 mult-word mode.
21257
a5ba7b92
AK		 212582013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21259
21260 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21261
be624986
TG		 212622013-03-27 Terry Guo <terry.guo@arm.com>
21263
21264 * config/arm/arm-cores.def: Added core cortex-r7.
21265 * config/arm/arm-tune.md: Regenerated.
21266 * config/arm/arm-tables.opt: Regenerated.
21267 * doc/invoke.texi: Added entry for core cortex-r7.
21268
0a514f47
WL		 212692013-03-27 Walter Lee <walt@tilera.com>
21270
21271 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21272 double-decrement of next_scratch_regno.
21273
9332b0d2
WL		 212742013-03-27 Walter Lee <walt@tilera.com>
21275
801d9b2a 21276 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 21277 input operands.
21278 (insn_v1mulus): Ditto.
21279 (insn_v2muls): Ditto.
21280
f54ea5dd
WL		 212812013-03-27 Walter Lee <walt@tilera.com>
21282
e594716a 21283 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 21284 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21285
bf60f4ca
WL		 212862013-03-27 Walter Lee <walt@tilera.com>
21287
e594716a 21288 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 21289 (*sibcall_value): Ditto.
21290
e3b51eeb
WL		 212912013-03-27 Walter Lee <walt@tilera.com>
21292
21293 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21294 (insn_mnz_v8qi): ... this ...
21295 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21296 vector equivalent.
21297 (insn_v<n>mnz): Replaced by ...
21298 (insn_v1mnz): ... this ...
21299 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21300 equivalent.
21301 (insn_mz_<mode>): Replaced by ...
21302 (insn_mz_v8qi): ... this ...
21303 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21304 vector equivalent.
21305 (insn_v<n>mz): Replaced by ...
21306 (insn_v1mz): ... this ...
21307 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21308 equivalent.
21309
4fc7b145
EB		 213102013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21311
21312 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21313
6f33817e
RM		 213142013-03-26 Roland McGrath <mcgrathr@google.com>
21315
21316 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21317 than fprintf with a non-constant, non-format string.
21318
93a4145b
UB		 213192013-03-26 Uros Bizjak <ubizjak@gmail.com>
21320
21321 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21322 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21323 operand 0 predicate.
21324 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21325 attribute. Use general_x64nomem_operand as operand 1 predicate.
21326 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21327 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21328 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21329 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21330 (mov<mode>_insv_1): Remove expander. Merge insn with
21331 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21332 Use general_x64nomem_operand as operand 1 predicate.
21333 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21334 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21335 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21336 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21337 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21338 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21339 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21340 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21341 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21342 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21343 (general_x64nomem_operand): Ditto.
21344
c6a9ed5a 213452013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 21346
21347 * config/rtems.opt: Add -pthread option.
21348
31b0a960
RB		 213492013-03-26 Richard Biener <rguenther@suse.de>
21350
93a4145b
UB		 21351 * alias.c (find_base_term): Avoid redundant and not used recursion.
21352 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 21353 (true_dependence_1): Compute and pass base terms to base_alias_check.
21354 (write_dependence_p): Likewise.
21355 (may_alias_p): Likewise.
21356
79517551
SN		 213572013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21358
21359 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21360 PC-relative load in SI modes and above only.
21361
a76213b9
XQ		 213622013-03-26 Xinyu Qi <xyqi@marvell.com>
21363
21364 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 21365 * config/arm/iwmmxt.md (WCGR0): Update.
21366 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21367
37ff9355
UB		 213682013-03-26 Uros Bizjak <ubizjak@gmail.com>
21369
21370 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21371 Use x64 and nox64 isa attributes.
21372
30b0317c
RB		 213732013-03-26 Richard Biener <rguenther@suse.de>
21374
21375 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21376 alignment computations and rely on get_object_alignment_1
21377 for the !TYPE_P case.
21378 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21379
cbcdb140
WL		 213802013-03-26 Walter Lee <walt@tilera.com>
21381
21382 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21383 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21384
dd884525 213852013-03-25 Jeff Law <law@redhat.com>
be672e08 21386
33159866
UB		 21387 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21388 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 21389 wrong version of prior patch.
21390
0051d3ec
WL		 213912013-03-25 Walter Lee <walt@tilera.com>
21392
21393 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21394 TILEGX_INSN_SHUFFLEBYTES1.
21395 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21396 shufflebytes1.
21397 (tilegx_builtins): Ditto.
21398 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21399
065a3605
WL		 214002013-03-25 Walter Lee <walt@tilera.com>
21401
21402 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 21403 (floatunssisf2): New pattern.
21404 (floatsidf2): New pattern.
21405 (floatunssidf2): New pattern.
065a3605 21406
5b2a3c82
WL		 214072013-03-25 Walter Lee <walt@tilera.com>
21408
21409 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21410 tests for constraint J, K, N, P.
21411
192ea533
WL		 214122013-03-25 Walter Lee <walt@tilera.com>
21413
21414 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21415 Use indirect/pcrel encoding.
21416 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21417 Ditto.
21418
855e0d0b
SE		 214192013-03-25 Steve Ellcey <sellcey@mips.com>
21420
21421 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21422 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21423 * config/mips/mips.c (mips_option_override): Set IMADD default.
21424 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21425 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21426 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21427 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21428 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21429
39e45653
JL		 214302013-03-25 Jeff Law <law@redhat.com>
21431
21432 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21433 slightly to avoid creating and folding useless trees. Simplify
21434 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21435
7482c470
UB		 214362013-03-25 Uros Bizjak <ubizjak@gmail.com>
21437
21438 * config/i386/i386.md (*zero_extendsidi2): Merge with
21439 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21440 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21441 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21442 targets. Clarify comment.
21443
4a53743e
MJ		 214442013-03-25 Martin Jambor <mjambor@suse.cz>
21445
21446 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21447 pass-through jump functions differently.
21448 (ipa_read_jump_function): Likewise. Also use setter functions to set
21449 up jump functions.
21450
162712de
MJ		 214512013-03-25 Martin Jambor <mjambor@suse.cz>
21452
21453 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21454 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21455 process it.
21456 (ipa_get_indirect_edge_target): New function.
21457 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21458 ipa_get_indirect_edge_target. Update all callers.
21459 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21460 ipa_get_indirect_edge_target_1 instead of calling
21461 ipa_get_indirect_edge_target.
21462 (create_specialized_node): Pass aggvlas to
21463 ipcp_discover_new_direct_edges.
21464
a5a4c20a
KT		 214652013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21466
21467 * config/arm/arm.md (f_sels, f_seld): New types.
21468 (*cmov<mode>): New pattern.
21469 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21470 predicate.
21471
f992b9fc
KT		 214722013-03-25 Kai Tietz <ktietz@redhat.com>
21473
21474 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21475 POSIX-printf for mingw-hosted builds.
21476
c16fd676
RB		 214772013-03-25 Richard Biener <rguenther@suse.de>
21478
21479 PR middle-end/56694
21480 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21481 must-not-throw stmt location.
21482
5ad29f12
KT		 214832013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21484
21485 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21486 Emit load-acquire versions when acq is true.
21487 (arm_emit_store_exclusive): Add rel parameter.
21488 Emit store-release versions when rel is true.
21489 (arm_split_compare_and_swap): Use acquire-release instructions
21490 instead.
21491 of barriers when appropriate.
21492 (arm_split_atomic_op): Likewise.
21493 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21494 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21495 (VUNSPEC_SLX): Likewise.
21496 (VUNSPEC_LDA): Likewise.
21497 (VUNSPEC_STL): Likewise.
21498 * config/arm/sync.md (atomic_load<mode>): New pattern.
21499 (atomic_store<mode>): Likewise.
21500 (arm_load_acquire_exclusive<mode>): Likewise.
21501 (arm_load_acquire_exclusivesi): Likewise.
21502 (arm_load_acquire_exclusivedi): Likewise.
21503 (arm_store_release_exclusive<mode>): Likewise.
21504
03a7dddb
CM		 215052013-03-25 Catherine Moore <clm@codesourcery.com>
21506
21507 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21508 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21509 * config/mip/predicates.md (lwsp_swsp_operand,
21510 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21511 sb16_operand, db4_operand, db7_operand, ib3_operand,
21512 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21513 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21514 andi16_operand): New predicates.
21515 * config/mips/mips.md (compression): New attribute.
21516 (enabled): New attribute.
21517 (length): Consider compression in computing length.
21518 (shift_compression): New code attribute.
21519 (*add<mode>3): New operands. Record compression.
21520 (sub<mode>3): Likewise.
21521 (one_cmpl<mode>2): Likewise.
21522 (*and<mode>3): Likewise.
21523 (*ior<mode>3): Likewise.
21524 (unnamed pattern for xor): Likewise.
21525 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21526 (*<optab><mode>3): Likewise.
21527 (*mov<mode>_internal: Likewise.
21528 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21529 (mips_unsigned_immediate_p): New.
21530 (umips_lwsp_swsp_address_p): New.
21531 (m16_based_address_p): New.
21532 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21533 (mips_unsigned_immediate_p): New prototype.
21534 (lwsp_swsp_address_p): New prototype.
21535 (m16_based_address_p): New prototype.
21536 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21537 (mips_signed_immediate_p): New function.
21538 (m16_based_address_p): New function.
21539 (lwsp_swsp_address_p): New function.
21540 (mips_print_operand_punctuation): Recognize short delay slot insns
21541 for microMIPS.add<mode>3"
21542
f35c297f 215432013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 21544
f35c297f
KT		 21545 PR target/56720
21546 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21547 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21548
051b9446
RB		 215492013-03-25 Richard Biener <rguenther@suse.de>
21550
21551 PR tree-optimization/56689
21552 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21553 any edge.
21554
374001cb
RB		 215552013-03-25 Richard Biener <rguenther@suse.de>
21556
21557 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21558 of bitmap.
21559 (memory_references): Likewise.
21560 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21561 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21562 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21563 (gather_mem_refs_in_loops): Fold into ...
21564 (analyze_memory_references): ... this. Move initialization
21565 to tree_ssa_lim_initialize.
21566 (fill_always_executed_in): Rename to ...
21567 (fill_always_executed_in_1): ... this.
21568 (fill_always_executed_in): Move contains_call computation to
21569 this new function from ...
21570 (tree_ssa_lim_initialize): ... here.
21571 (tree_ssa_lim): Call fill_always_executed_in.
21572
57895947
EB		 215732013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21574
21575 * postreload.c (reload_combine): Fix code detecting returns.
21576
1e1b18c1
EB		 215772013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21578
21579 * function.c (emit_use_return_register_into_block): On cc0 targets,
21580 do not emit the sequence between cc0 setter and user.
21581
9216baf1
KT		 215822013-03-25 Kai Tietz <ktietz@redhat.com>
21583
1e1b18c1
EB		 21584 * config/i386/predicates.md (local_symbolic_operand): Interpret
21585 dll-imported symbols as none-local.
9216baf1 21586
76421b44
RB		 215872013-03-25 Richard Biener <rguenther@suse.de>
21588
21589 * tree-ssa-loop-im.c (struct depend): Remove.
21590 (struct lim_aux_data): Make depends a vec of gimples.
21591 (free_lim_aux_data): Adjust.
21592 (add_dependency): Likewise.
21593 (set_level): Likewise.
21594
d154bfa2
RB		 215952013-03-25 Richard Biener <rguenther@suse.de>
21596
21597 PR middle-end/56434
21598 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21599 the pointer returned by calls with ECF_MALLOC set.
21600
2cd9804e
UB		 216012013-03-24 Uros Bizjak <ubizjak@gmail.com>
21602
acef5fe0
UB		 21603 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21604
216052013-03-24 Uros Bizjak <ubizjak@gmail.com>
21606
21607 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 21608 using MMXMODE mode iterator.
21609 (*move<mode>_internal): Merge with *movv2sf_internal and
21610 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21611
225ccc68
SB		 216122013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21613
85c0f02d
SB		 21614 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21615 (record_last_mem_set_info): Likewise.
21616
225ccc68
SB		 21617 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21618 of XNEWVEC followed by memset.
21619 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21620
b64925dc
SB		 216212013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21622
21623 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21624 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21625 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21626 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21627 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21628 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21629 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21630 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21631 BARRIER_P instead of GET_CODE.
21632
b9aaf52e
EB		 216332013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21634
21635 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21636 inaccuracy in the probing code.
21637
21638 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21639 (ctrapdi4): Likewise.
21640
66de4d7c
EB		 216412013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21642
21643 * calls.c (expand_call): Add missing guard to code handling return
21644 of non-BLKmode structures in MSB.
21645 * function.c (expand_function_end): Likewise.
21646
4f25c8fb
EB		 216472013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21648
21649 * combine.c (try_combine): Adjust comment. Do not add the set of
21650 insn #0 if the destination indirectly is set or dies in insn #2.
21651 Tidy up code to distribute a new note.
21652
23b7850d
UB		 216532013-03-22 Uros Bizjak <ubizjak@gmail.com>
21654
21655 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21656 also for alternatives 16 and 17.
21657
bed852cf
UB		 216582013-03-22 Uros Bizjak <ubizjak@gmail.com>
21659
21660 * config/i386/sse.md (*mov<mode>_internal): Merge with
21661 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21662 Emit insn template depending on type attribute. Use
21663 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21664 movd instead of movq mnemonic for interunit moves. Rewrite mode
21665 attribute calculation. Remove unit attribute calculation.
21666 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21667 Set prefix_data16 attribute for DImode ssemov types.
21668 Use Ym instead of y for SSE-MMX conversion alternatives.
21669 Reorder operand constraints.
21670
1e8a7937
SB		 216712013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21672
21673 * df.h (df_insn_delete): Adjust prototype.
21674 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21675 and let it decide whether mark the basic block dirty.
21676 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21677 * df-scan.c (df_insn_info_delete): New helper function, split
21678 off from df_insn_delete.
21679 (df_scan_free_bb_info): Use it.
21680 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21681 Likewise.
21682 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21683 that the insn is actually an insn and it has a non-NULL basic block.
21684 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21685
e14d094c
RB		 216862013-03-22 Richard Biener <rguenther@suse.de>
21687
21688 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21689 dep_ref members.
21690 (mem_ref_alloc): Do not allocate them.
21691 (refs_independent_p): Do not query or maintain a cache.
21692
e9cf7316
RB		 216932013-03-22 Richard Biener <rguenther@suse.de>
21694
21695 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21696 (gather_mem_refs_in_loops): Do not compute it.
21697 (analyze_memory_references): Do not allocate it.
21698 (tree_ssa_lim_finalize): Do not free it.
21699 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21700
f046e81b
RB		 217012013-03-22 Richard Biener <rguenther@suse.de>
21702
21703 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 21704
a05c0ddf
IB		 217052013-03-22 Ian Bolton <ian.bolton@arm.com>
21706
21707 * config/aarch64/aarch64.c (aarch64_print_operand): New
21708 format specifier for printing a constant in hex.
21709 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21710 format specifier for printing second operand.
21711
c00217fc
RB		 217122013-03-22 Richard Biener <rguenther@suse.de>
21713
21714 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21715 bitmaps.
21716 (gather_mem_refs_in_loops): Perform store accumulation here.
21717 (create_vop_ref_mapping_loop): Remove.
21718 (create_vop_ref_mapping): Likewise.
21719 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 21720 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 21721 (record_indep_loop): Remove.
21722 (record_dep_loop): New function.
21723 (ref_indep_loop_p_1): Adjust to only walk over references
21724 in the loop, not its subloops.
21725 (ref_indep_loop_p): Rename to ...
21726 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21727 maintaining a more fine-grained cache.
21728 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21729 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21730
15d19bf8
RB		 217312013-03-22 Richard Biener <rguenther@suse.de>
21732
21733 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21734 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21735 aggregate mem_ref_loc.
21736 (free_mem_ref_locs): Inline into ...
21737 (memref_free): ... this and adjust.
21738 (mem_ref_alloc): Adjust.
21739 (mem_ref_locs_alloc): Remove.
21740 (record_mem_ref_loc): Adjust.
21741 (get_all_locs_in_loop): Rewrite into ...
21742 (for_all_locs_in_loop): ... this iterator.
21743 (rewrite_mem_ref_loc): New functor.
21744 (rewrite_mem_refs): Use for_all_locs_in_loop.
21745 (sm_set_flag_if_changed): New functor.
21746 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21747 (ref_always_accessed): New functor.
21748 (ref_always_accessed_p): Use for_all_locs_in_loop.
21749
6f37411d
MG		 217502013-03-21 Marc Glisse <marc.glisse@inria.fr>
21751
21752 * tree-pass.h (PROP_gimple_lvec): New.
21753 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21754 (init_optimization_passes): Move pass_lower_vector.
21755 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21756 PROP_gimple_lvec.
21757 (pass_lower_vector): Provide PROP_gimple_lvec.
21758 (pass_lower_vector_ssa): Likewise.
21759 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21760
511ef689
MW		 217612013-03-21 Mark Wielaard <mjw@redhat.com>
21762
21763 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21764
cc1df30b
UB		 217652013-03-21 Uros Bizjak <ubizjak@gmail.com>
21766
21767 * config/i386/i386.md (*movdi_internal): Disparage slightly
21768 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21769 conversion alternatives.
21770
c845cfe1
JJ		 217712013-03-21 Jakub Jelinek <jakub@redhat.com>
21772
37e99116
JJ		 21773 PR middle-end/48087
21774 * diagnostic.def (DK_WERROR): New kind.
21775 * diagnostic.h (werrorcount): Define.
21776 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21777 promoted to DK_ERROR, increment DK_WERROR counter instead of
21778 DK_ERROR counter.
21779 * toplev.c (toplev_main): Call print_ignored_options even if
21780 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21781 even if just werrorcount is non-zero.
21782
c845cfe1
JJ		 21783 PR debug/55608
21784 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21785 on failure.
21786 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21787 (string_cst_pool_decl): New function.
21788 (optimize_one_addr_into_implicit_ptr): New function.
21789 (resolve_addr_in_expr): Optimize DWARF location expression
21790 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21791 which doesn't live in memory, but has DW_AT_location or
21792 DW_AT_const_value, or refers to a string literal, into
21793 DW_OP_GNU_implicit_pointer.
21794 (optimize_location_into_implicit_ptr): New function.
21795 (resolve_addr): If removing DW_AT_location of a variable because
21796 it was DW_OP_addr of address of the variable, but the variable doesn't
21797 live in memory, try to emit const value attribute for the initializer.
21798
08e0cda6
MG		 217992013-03-21 Marc Glisse <marc.glisse@inria.fr>
21800
21801 * tree.h (VECTOR_TYPE_P): New macro.
21802 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21803 TYPE_MODE): Use it.
21804 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21805 VEC_COND_EXPR cannot be lvalues.
21806 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21807
d08633b4
MG		 218082013-03-21 Marc Glisse <marc.glisse@inria.fr>
21809
21810 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21811 Restrict the transformation to equal modes.
21812
e6647190
RB		 218132013-03-21 Richard Biener <rguenther@suse.de>
21814
21815 PR tree-optimization/39326
21816 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21817 (MEM_ANALYZABLE): Adjust.
21818 (record_mem_ref_loc): Move bitmap ops ...
21819 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21820 unanalyzable refs, do not record locations for it.
21821 (analyze_memory_references): Allocate ref zero as shared
21822 unanalyzable ref.
21823 (refs_independent_p): Do not test for unanalyzed mems here.
21824 (ref_indep_loop_p_1): Special-case disambiguation against
21825 the unanalyzed ref.
cc1df30b 21826 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 21827
65074f54
CL		 218282013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21829
21830 * config/arm/arm-protos.h (tune_params): Add
21831 prefer_neon_for_64bits field.
21832 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21833 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21834 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21835 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21836 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21837 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21838 (arm_option_override): Handle -mneon-for-64bits new option.
21839 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21840 (prefer_neon_for_64bits): Declare new variable.
21841 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 21842 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 21843 (arch_enabled): Handle new arch types. Remove support for onlya8
21844 and nota8.
21845 (one_cmpldi2): Use new arch names.
bc5faa5b 21846 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 21847 * config/arm/arm.opt (mneon-for-64bits): Add option.
21848 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21849 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21850 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21851 of onlya8.
21852 * doc/invoke.texi (-mneon-for-64bits): Document.
21853
5a2d2a79
RB		 218542013-03-21 Richard Biener <rguenther@suse.de>
21855
21856 PR tree-optimization/39326
21857 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21858 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 21859 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 21860
5abe1e05
RB		 218612013-03-21 Richard Biener <rguenther@suse.de>
21862
21863 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21864 (vect_insert_into_interleaving_chain): Likewise.
21865 (vect_drs_dependent_in_basic_block): Inline ...
21866 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21867 split out from ...
21868 (vect_analyze_data_ref_dependence): ... here. Simplify.
21869 (vect_check_interleaving): Simplify.
21870 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21871 (vect_slp_analyze_data_ref_dependences): ... this new function.
21872 (dr_group_sort_cmp): New function.
21873 (vect_analyze_data_ref_accesses): Compute data-reference groups
21874 here instead of in vect_analyze_data_ref_dependence. Use
21875 a more efficient algorithm.
21876 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21877 vect_slp_analyze_data_ref_dependences. Call
21878 vect_analyze_data_ref_accesses earlier.
21879 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21880 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21881 (vect_slp_analyze_data_ref_dependences): New prototype.
21882
cad1735b
RB		 218832013-03-21 Richard Biener <rguenther@suse.de>
21884
21885 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21886 ref is stored in the loop.
21887 (find_refs_for_sm): Walk only over all stores.
21888 (store_motion_loop): Allocate from lim_bitmap_obstack.
21889 (store_motion): Likewise.
21890
141310ef
RB		 218912013-03-21 Richard Biener <rguenther@suse.de>
21892
21893 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21894 Update virtual SSA form.
21895
5022315a
RO		 218962013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21897
21898 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21899 * configure: Regenerate.
21900 * config.in: Regenerate.
21901 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21902 if !HAVE_LD_EH_FRAME_CIEV3.
21903
50d4421c
RB		 219042013-03-21 Richard Biener <rguenther@suse.de>
21905
21906 * tree-cfg.c (verify_expr_no_block): New function.
21907 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21908 nor DECL_VALUE_EXPR have locations with associated blocks.
21909 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21910 (clear_unused_block_pointer): Remove code dealing with
21911 blocks in DECL_DEBUG_EXPR locations.
21912
839b422f
RB		 219132013-03-21 Richard Biener <rguenther@suse.de>
21914
21915 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21916 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21917 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21918 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21919 instead of DECL_DEBUG_EXPR_IS_FROM.
21920 * gimplify.c (gimplify_modify_expr): Likewise.
21921 * tree-cfg.c (verify_expr_location_1): Likewise.
21922 * tree-complex.c (create_one_component_var): Likewise.
21923 * tree-sra.c (create_access_replacement): Likewise.
21924 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21925 (clear_unused_block_pointer): Likewise.
21926 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21927 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
21928 * var-tracking.c (var_debug_decl): Likewise.
21929 (track_expr_p): Likewise.
21930 * tree-inline.c (add_local_variables): Likewise. Set
21931 DECL_HAS_DEBUG_EXPR_P after copying it.
21932 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
21933 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
21934
fe04878d
UB		 219352013-03-21 Uros Bizjak <ubizjak@gmail.com>
21936
21937 PR bootstrap/56656
21938 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
21939 * configure: Regenerate.
21940 * config.in: Regenerate.
21941 * config/i386/i386.md (*movdf_internal): Use
21942 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21943 movd instead of movq mnemonic for interunit moves.
21944 (*movdi_internal): Ditto.
21945
bd059b26 219462013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 21947
21948 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
21949 (abd<mode>_3): New pattern.
21950 (aba<mode>_3): New pattern.
21951 (fabd<mode>_3): New pattern.
21952
bd059b26 219532013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 21954
21955 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
21956 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
21957 occurrence of REGISTER_PREFIX as its empty string.
21958
4f1f78b9
JL		 219592013-03-20 Jeff Law <law@redhat.com>
21960
21961 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
21962 addititional equivalences for equality comparisons between an SSA_NAME
21963 and a constant where the SSA_NAME was set from a widening conversion.
21964
327a1118
WL		 219652013-03-20 Walter Lee <walt@tilera.com>
21966
21967 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
21968
813e0036
UB		 219692013-03-20 Uros Bizjak <ubizjak@gmail.com>
21970
21971 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
21972 depending on type attribute.
21973 (*movti_internal): Ditto.
21974 (*movtf_internal): Ditto.
21975 (*movxf_internal): Ditto.
21976 (*movdf_internal): Ditto.
21977 (*movsf_internal): Ditto.
21978
7cf34aae
UB		 219792013-03-20 Uros Bizjak <ubizjak@gmail.com>
21980
21981 * config/i386/i386.md (*movti_internal): Set prefix attribute to
21982 maybe_vex for sselog1 and ssemov types.
21983 (*movdi_internal): Reorder operand constraints.
21984 (*movsi_internal): Ditto. Set prefix attribute to
21985 maybe_vex for sselog1 and ssemov types.
21986 (*movtf_internal): Set prefix attribute to maybe_vex
21987 for sselog1 and ssemov types.
21988 (*movdf_internal): Ditto. Set prefix_data16 attribute for
21989 DImode ssemov types. Reorder operand constraints.
21990 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
21991 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
21992 attribute for SImode ssemov types. Reorder operand constraints.
21993
19321415
MJ		 219942013-03-20 Martin Jambor <mjambor@suse.cz>
21995
21996 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
21997 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
21998
12211b99 219992013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 22000
22001 * config/rs6000/predicates.md (indexed_address, update_address_mem
22002 update_indexed_address_mem): New predicates.
22003 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22004 attribute for load/store instructions.
22005 * config/rs6000/dfp.md (movsd_store): Likewise.
22006 (movsd_load): Likewise.
22007 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22008 (unnamed HI->DI extend define_insn): Likewise.
22009 (unnamed SI->DI extend define_insn): Likewise.
22010 (unnamed QI->SI extend define_insn): Likewise.
22011 (unnamed QI->HI extend define_insn): Likewise.
22012 (unnamed HI->SI extend define_insn): Likewise.
22013 (unnamed HI->SI extend define_insn): Likewise.
22014 (extendsfdf2_fpr): Likewise.
22015 (movsi_internal1): Likewise.
22016 (movsi_internal1_single): Likewise.
22017 (movhi_internal): Likewise.
22018 (movqi_internal): Likewise.
22019 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22020 attribute for load/store instructions.
22021 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22022 instructions.
22023 (mov<mode>_softfloat): Likewise.
22024 (mov<mode>_hardfloat32): Likewise.
22025 (mov<mode>_hardfloat64): Likewise.
22026 (mov<mode>_softfloat64): Likewise.
22027 (movdi_internal32): Likewise.
22028 (movdi_internal64): Likewise.
22029 (probe_stack_<mode>): Likewise.
22030
220312013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22032
22033 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22034 floating point, and decimal floating point to reload iterator.
22035
22036 * config/rs6000/constraints.md (wl constraint): New constraints to
22037 return FLOAT_REGS if certain options are used to reduce the number
22038 of separate patterns that exist in the file.
22039 (wx constraint): Likewise.
22040 (wz constraint): Likewise.
22041
22042 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22043 -mdebug=reg, print wg, wl, wx, and wz constraints.
22044 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22045 Initialize the reload functions for 64-bit binary/decimal floating
22046 point types.
22047 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22048 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22049 create the buffer on the stack to overcome not having a 32-bit
22050 load and store.
22051 (rs6000_emit_move): Likewise.
22052 (rs6000_secondary_memory_needed_rtx): Likewise.
22053 (rs6000_alloc_sdmode_stack_slot): Likewise.
22054 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22055 via xxlxor, just like DFmode 0.0.
22056
22057 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22058 define as 1 if we are running on a power7 or newer.
22059 (enum r6000_reg_class_enum): Add new constraints.
22060
22061 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22062 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22063 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 22064 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 22065 (movsd splitter): Likewise.
22066 (movsd_hardfloat): Likewise.
22067 (movsd_softfloat): Likewise.
22068
22069 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22070 binary and decimal floating point moves.
22071 (fmove_ok): New attributes to combine binary and decimal floating
22072 point moves, and to combine power6x (mfpgpr) moves along normal
22073 floating moves.
22074 (real_value_to_target): Likewise.
22075 (f32_lr): Likewise.
22076 (f32_lm): Likewise.
22077 (f32_li): Likewise.
22078 (f32_sr): Likewise.
22079 (f32_sm): Likewise.
22080 (f32_si): Likewise.
22081 (movsf): Combine binary and decimal floating point moves. Combine
22082 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 22083 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 22084 (mov<mode> for SFmode/SDmode); Likewise.
22085 (SFmode/SDmode splitters): Likewise.
22086 (movsf_hardfloat): Likewise.
22087 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22088 (movsf_softfloat): Likewise.
22089 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22090
22091 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22092 wx and wz constraints.
22093
22094 * config/rs6000/constraints.md (wg constraint): New constraint to
22095 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22096
22097 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22098 constraint.
22099
22100 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22101 -mdebug=reg, print wg, wl, wx, and wz constraints.
22102 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22103 Initialize the reload functions for 64-bit binary/decimal floating
22104 point types.
22105 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22106 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22107 create the buffer on the stack to overcome not having a 32-bit
22108 load and store.
22109 (rs6000_emit_move): Likewise.
22110 (rs6000_secondary_memory_needed_rtx): Likewise.
22111 (rs6000_alloc_sdmode_stack_slot): Likewise.
22112 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22113 via xxlxor, just like DFmode 0.0.
22114
c6d5ff83
MM		 22115 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22116 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22117 with other moves by using conditional constraits (wg). Use LFIWZX
22118 and STFIWX for loading SDmode on power7.
22119 (movdd splitters): Likewise.
22120 (movdd_hardfloat32): Likewise.
22121 (movdd_softfloat32): Likewise.
22122 (movdd_hardfloat64_mfpgpr): Likewise.
22123 (movdd_hardfloat64): Likewise.
22124 (movdd_softfloat64): Likewise.
22125
22126 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22127 64-bit binary and decimal floating point moves.
22128 (FMOVE64X): Likewise.
22129 (movdf): Combine 64-bit binary and decimal floating point moves.
22130 Combine power6x (mfpgpr) moves with other moves by using
22131 conditional constraits (wg).
22132 (mov<mode> for DFmode/DDmode): Likewise.
22133 (DFmode/DDmode splitters): Likewise.
22134 (movdf_hardfloat32): Likewise.
22135 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22136 (movdf_softfloat32): Likewise.
22137 (movdf_hardfloat64_mfpgpr): Likewise.
22138 (movdf_hardfloat64): Likewise.
22139 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22140 (movdf_softfloat64): Likewise.
22141 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22142 (reload_<mode>_load): Move to later in the file so they aren't in
22143 the middle of the floating point move insns.
22144 (reload_<mode>_store): Likewise.
22145
22146 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22147 constraint.
22148
22149 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22150 constraint if -mdebug=reg.
bd059b26
UB		 22151 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22152 Enable using dd reload support if needed.
c6d5ff83
MM		 22153
22154 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22155 binary and decimal floating point moves in rs6000.md.
22156 (movtd_internal): Likewise.
22157
22158 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22159 decimal floating point moves.
22160 (movtf): Likewise.
22161 (movtf_internal): Likewise.
22162 (mov<mode>_internal, TDmode/TFmode): Likewise.
22163 (movtf_softfloat): Likewise.
22164 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22165
22166 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22167 movdi_internal64, using wg constraint for move direct operations.
22168 (movdi_internal64): Likewise.
22169
22170 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22171 MODES_TIEABLE_P for selected modes. Print the numerical value of
22172 the various virtual registers. Use GPR/FPR first/last values,
22173 instead of hard coding the register numbers. Print which modes
22174 have reload functions registered.
bd059b26
UB		 22175 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22176 settings before/after setting cpu, target and subtarget settings.
22177 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22178 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 22179 (rs6000_secondary_reload_fail): Likewise.
22180 (rs6000_secondary_reload_inner): Likewise.
22181
22182 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22183 macros for first/last GPR and FPR registers.
22184 (LAST_GPR_REGNO): Likewise.
22185 (FIRST_FPR_REGNO): Likewise.
22186 (LAST_FPR_REGNO): Likewise.
22187
22188 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22189 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22190 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22191 (vcond<mode><mode>): Likewise.
22192 (vcondu<mode><mode>): Likewise.
22193 (vector_gtu<mode>): Likewise.
22194 (vector_gte<mode>): Likewise.
22195 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22196 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 22197 (ior<mode>3): Likewise.
22198 (and<mode>3): Likewise.
22199 (one_cmpl<mode>2): Likewise.
22200 (nor<mode>3): Likewise.
22201 (andc<mode>3): Likewise.
22202
22203 * config/rs6000/constraints.md (wt constraint): New constraint
22204 that returns VSX_REGS if TImode is allowed in VSX registers.
22205
22206 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22207 constant under VSX.
22208
22209 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22210 similar to TImode, but it is restricted to being in the GPRs.
22211
22212 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22213 TImode to occupy a single VSX register.
22214
22215 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22216 -mvsx-timode for power7/power8.
22217 (power7 cpu): Likewise.
22218 (power8 cpu): Likewise.
22219
22220 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22221 sure that TFmode/TDmode take up two registers if they are ever
22222 allowed in the upper VSX registers.
22223 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22224 registers.
22225 (rs6000_init_hard_regno_mode_ok): Likewise.
22226 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22227 constraint. Print if LRA is turned on.
22228 (rs6000_option_override_internal): Give an error if -mvsx-timode
22229 and VSX is not enabled.
22230 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22231 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22232 to reg+offset addressing. Use PTImode when checking offset
22233 addresses for validity.
22234 (reg_offset_addressing_ok_p): Likewise.
22235 (rs6000_legitimate_offset_address_p): Likewise.
22236 (rs6000_legitimize_address): Likewise.
22237 (rs6000_legitimize_reload_address): Likewise.
22238 (rs6000_legitimate_address_p): Likewise.
22239 (rs6000_eliminate_indexed_memrefs): Likewise.
22240 (rs6000_emit_move): Likewise.
22241 (rs6000_secondary_reload): Likewise.
22242 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22243 reloads to fpr registers to continue to use reg+offset addressing,
22244 but 64-bit reloads to altivec registers need reg+reg addressing.
22245 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22246 it. Treat LO_SUM like a PLUS operation.
22247 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22248 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 22249 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22250 registers to share a register with a smaller sized type, since VSX
22251 puts scalars in the upper 64-bits.
22252 (print_operand): Add support for PTImode.
22253 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22254 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22255 registers, but don't have arithmetic support.
22256 (rs6000_memory_move_cost): Add test for VSX.
22257 (rs6000_opt_masks): Add -mvsx-timode.
22258
22259 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22260 for TImode.
22261 (VSs): Likewise.
22262 (VSr): Use wt constraint for TImode.
22263 (VSv): Drop TImode support.
22264 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22265 (vsx_movti_64bit): Likewise.
22266 (vsx_movti_32bit): Likewise.
22267 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22268 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22269 one '?' on the appropriate output constraint. Do not allow TImode
22270 logical operations on 32-bit systems.
22271 (vsx_ior<mode>3): Likewise.
22272 (vsx_xor<mode>3): Likewise.
22273 (vsx_one_cmpl<mode>2): Likewise.
22274 (vsx_nor<mode>3): Likewise.
22275 (vsx_andc<mode>3): Likewise.
22276 (vsx_concat_<mode>): Likewise.
22277 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22278
22279 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22280 OPTION_MASK_VSX_TIMODE.
22281 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22282 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22283
22284 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22285 (TI2 iterator): New iterator for TImode, PTImode.
22286 (wd mode attribute): Add values for vector types.
bd059b26
UB		 22287 (movti_string): Replace TI move operations with operations for TImode
22288 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 22289 (mov<mode>_string, TImode/PTImode): Likewise.
22290 (movti_ppc64): Likewise.
22291 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22292 (TI mode splitters): Likewise.
22293
22294 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22295 constraint.
22296
1fc5eced
MG		 222972013-03-20 Marc Glisse <marc.glisse@inria.fr>
22298
22299 PR tree-optimization/56355
22300 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22301 Also handle integers with undefined overflow.
22302
22c4c869
CM		 223032013-03-20 Catherine Moore <clm@codesourcery.com>
22304 Maciej W. Rozycki <macro@codesourcery.com>
22305 Tom de Vries <tom@codesourcery.com>
12211b99 22306 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 22307 Iain Sandoe <iain@codesourcery.com>
22308 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22309 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 22310
22311 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22312 Document new function attributes.
22c4c869
CM		 22313 * doc/invoke.texi (minterlink-compressed, mmicromips,
22314 m14k, m14ke, m14kec): Document new options.
22315 (minterlink-mips16): Update documentation.
22316 * doc/md.texi (ZC, ZD): Document new constraints.
22317 * configure.ac (gcc_cv_as_micromips): Check if linker
22318 supports the .set micromips directive.
22319 * configure: Regenerate.
22320 * config.in: Regenerate.
22321 * config/mips/mips-tables.opt: Regenerate.
22322 * config/mips/micromips.md: New file.
22323 * constraints.md (ZC, ZD): New constraints.
22324 * config/mips/predicates.md (movep_src_register): New predicate.
22325 (movep_src_operand): New predicate.
22326 (non_volatile_mem_operand): New predicate.
22327 * config/mips/mips.md (multimem): New type.
22328 (length): Differentiate between 17-bit and 18-bit branch offsets.
22329 (MOVEP1, MOVEP2): New mode iterator.
33159866 22330 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 22331 (mov_<load>r): Likewise.
22332 (mov_<store>l): Likewise.
22333 (mov_<store>r): Likewise.
22334 (*branch_equality<mode>_inverted): Add microMIPS support.
22335 (*branch_equality<mode>): Likewise.
22336 (*jump_absolute): Likewise.
22337 (indirect_jump_<mode>): Likewise.
22338 (tablejump_<mode>): Likewise.
22339 (<optab>_internal): Likewise.
22340 (sibcall_internal): Likewise.
22341 (sibcall_value_internal): Likewise.
22342 (prefetch): Use constraint ZD.
22343 * config/mips/mips.opt (minterlink-compressed): New option.
22344 (minterlink-mips16): Now an alias for minterlink-compressed.
22345 (mmicromips): New option.
22346 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22347 (compare_and_swap_12): Likewise.
22348 (sync_add<mode>): Likewise.
22349 (sync_<optab>_12): Likewise.
22350 (sync_old_<optab>_12): Likewise.
22351 (sync_new_<optab>_12): Likewise.
22352 (sync_nand_12): Likewise.
22353 (sync_old_nand_12): Likewise.
22354 (sync_new_nand_12): Likewise.
22355 (sync_sub<mode>): Likewise.
22356 (sync_old_add<mode>): Likewise.
22357 (sync_old_sub<mode>): Likewise.
22358 (sync_new_add<mode>): Likewise.
22359 (sync_new_sub<mode>): Likewise.
22360 (sync_<optab><mode>): Likewise.
22361 (sync_old_<optab><mode>): Likewise.
22362 (sync_new_<optab><mode>): Likewise.
22363 (sync_nand<mode>): Likewise.
22364 (sync_old_nand<mode>): Likewise.
22365 (sync_new_nand<mode>): Likewise.
22366 (sync_lock_test_and_set<mode>): Likewise.
22367 (test_and_set_12): Likewise.
22368 (atomic_compare_and_swap<mode>): Likewise.
22369 (atomic_exchange<mode>_llsc): Likewise.
22370 (atomic_fetch_add<mode>_llsc): Likewise.
22371 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22372 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22373 (umips_save_restore_pattern_p): Likewise.
22374 (umips_load_store_pair_p): Likewise.
22375 (umips_output_load_store_pair): Likewise.
22376 (umips_movep_target_p): Likewise.
22377 (umips_12bit_offset_address_p): Likewise.
22378 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22379 (mips_base_mips16): Rename this...
22380 (mips_base_compression_flags): ...to this. Update all uses.
22381 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22382 (mips_mips16_decl_p): Delete.
22383 (mips_nomips16_decl_p): Delete.
33159866 22384 (mips_get_compress_on_flags): New function.
22c4c869
CM		 22385 (mips_get_compress_off_flags): New function.
22386 (mips_get_compress_mode): New function.
22387 (mips_get_compress_on_name): New function.
22388 (mips_get_compress_off_name): New function.
22389 (mips_insert_attributes): Support multiple compression types.
22390 (mips_merge_decl_attributes): Likewise.
22391 (umips_12bit_offset_address_p): New function.
22392 (mips_start_function_definition): Emit .set micromips directive.
22393 (mips_call_may_need_jalx_p): New function.
22394 (mips_function_ok_for_sibcall): Add microMIPS support.
22395 (mips_print_operand_punctuation): Support short delay slots and
22396 compact jumps.
22397 (umips_swm_mask, umips_swm_encoding): New.
22398 (umips_build_save_restore): New function.
22399 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22400 (was_mips16_p): Remove.
22401 (old_compression_mode): New.
22402 (mips_set_compression_mode): New function.
22403 (mips_set_current_function): Add microMIPS support.
22404 (mips_option_override): Likewise.
22405 (umips_save_restore_pattern_p): New function.
22406 (umips_output_save_restore): New function.
22407 (umips_load_store_pair_p_1): New function.
22408 (umips_load_store_pair_p): New function.
22409 (umips_output_load_store_pair_1): New function.
22410 (umips_output_load_store_pair): New function.
22411 (umips_movep_target_p) New function.
22412 (mips_prepare_pch_save): Add microMIPS support.
22413 * config/mips/mips.h (TARGET_COMPRESSION): New.
22414 (TARGET_CPU_CPP_BUILTINS): Update macro
22415 to use new compression flags and to support microMIPS.
22416 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22417 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22418 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22419 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22420 (ASM_SPEC): Support mmicromips and mno-micromips.
22421 (M16STORE_REG_P): New macro.
22422 (MIPS_CALL): Support TARGET_MICROMIPS.
22423 (MICROMIPS_J): New macro.
22424 (mips_base_mips16): Rename this...
22425 (mips_base_compression_flags): ...to this.
22426 (UMIPS_12BIT_OFFSET_P): New macro.
22427 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22428 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 224292013-03-20 Richard Biener <rguenther@suse.de>
22430
22431 PR tree-optimization/56661
22432 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22433 the result does not have to be distinct.
22434
54714c68
RB		 224352013-03-20 Richard Biener <rguenther@suse.de>
22436
22437 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22438 remap_gimple_op_r.
22439
cca1130d 224402013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 22441 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 22442
22443 PR rtl-optimization/56605
22444 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22445
2a930377
UB		 224462013-03-20 Uros Bizjak <ubizjak@gmail.com>
22447
22448 PR bootstrap/56656
22449 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22450 that require movd instead of movq.
22451
d6d305fe
RB		 224522013-03-20 Richard Biener <rguenther@suse.de>
22453
22454 * tree-ssa-structalias.c (struct variable_info): Add pointer
22455 to the first field of an aggregate with sub-vars. Make
22456 this and the pointer to the next subfield its ID.
22457 (vi_next): New function.
22458 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22459 storedanything_id, integer_id): Increment by one.
22460 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22461 get_call_clobber_vi): Adjust.
22462 (solution_set_expand): Simplify and speedup.
22463 (solution_set_add): Inline into ...
22464 (set_union_with_increment): ... this. Adjust accordingly.
22465 (do_sd_constraint): Likewise.
22466 (do_ds_constraint): Likewise.
22467 (do_complex_constraint): Simplify.
22468 (build_pred_graph): Adjust.
22469 (solve_graph): Likewise. Simplify and speedup.
22470 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22471 get_constraint_for_component_ref, get_constraint_for_1,
22472 first_vi_for_offset, first_or_preceding_vi_for_offset,
22473 create_function_info_for, create_variable_info_for_1,
22474 create_variable_info_for, intra_create_variable_infos): Adjust.
22475 (init_base_vars): Push NULL for ID zero.
22476 (compute_points_to_sets): Adjust.
22477
a271b42d
RB		 224782013-03-20 Richard Biener <rguenther@suse.de>
22479
22480 * cfgloop.c (verify_loop_structure): Streamline and avoid
22481 ICEing on corrupt loop tree.
22482 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22483 loop tree.
22484
ebd65954
RB		 224852013-03-20 Richard Biener <rguenther@suse.de>
22486
22487 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22488 check whether an SSA update is needed.
22489
4547b7ee
RS		 224902013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22491
22492 * config/mips/constraints.md (T): Rename to...
22493 (Yf): ...this.
22494 (U): Rename to...
22495 (Yd): ...this.
22496 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22497 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22498
0b8cdc58
IB		 224992013-03-19 Ian Bolton <ian.bolton@arm.com>
22500
22501 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22502 (*subsi3_carryin_uxtw): Likewise.
22503
5977a10d
IB		 225042013-03-19 Ian Bolton <ian.bolton@arm.com>
22505
22506 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22507 (*rorsi3_insn_uxtw): Likewise.
22508
bd83ff2c
IB		 225092013-03-19 Ian Bolton <ian.bolton@arm.com>
22510
22511 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22512 (*extrsi5_insn_uxtw): Likewise.
22513
ba65123c
RB		 225142013-03-19 Richard Biener <rguenther@suse.de>
22515
22516 PR tree-optimization/56273
22517 * passes.c (init_optimization_passes): Move second VRP after DOM.
22518
2eac0476
UB		 225192013-03-19 Uros Bizjak <ubizjak@gmail.com>
22520
22521 * config/i386/i386.md (*movti_internal): Merge from
22522 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22523 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22524 nox64 isa attributes.
22525
6e55eda7
RB		 225262013-03-18 Richard Biener <rguenther@suse.de>
22527
22528 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22529 (unite): Likewise.
22530 (merge_node_constraints): Likewise.
22531 (build_succ_graph): Likewise.
22532 (valid_graph_edge): Inline into single caller.
22533 (unify_nodes): Likewise. Use bitmap_set_bit return value
22534 and cache varinfo.
22535 (scc_visit): Fix formatting and variable use.
22536 (do_sd_constraint): Use gcc_checking_assert.
22537 (do_ds_constraint): Likewise.
22538 (do_complex_constraint): Likewise.
22539 (condense_visit): Likewise. Cleanup.
22540 (dump_pred_graph): New function.
22541 (perform_var_substitution): Dump the pred-graph before
22542 variable substitution.
22543 (find_equivalent_node): Use gcc_checking_assert.
22544 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22545
4bdd44c4
RB		 225462013-03-18 Richard Biener <rguenther@suse.de>
22547
22548 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22549 Remove cond_expr_stmt_list argument and do not gimplify the
22550 built expression.
22551 (vect_loop_versioning): Adjust.
22552 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22553 Cleanup to use less temporaries.
22554 (vect_create_data_ref_ptr): Cleanup.
22555
38c56a5b
JJ		 225562013-03-18 Jakub Jelinek <jakub@redhat.com>
22557
22558 PR tree-optimization/56635
22559 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22560 require types_compatible_p types.
22561
20b2e6a0
NC		 225622013-03-18 Nick Clifton <nickc@redhat.com>
22563
a6178a25
NC		 22564 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22565 spurious backslash.
22566
20b2e6a0
NC		 22567 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22568 Add missing line to comment describing function.
22569
92608d0e
RB		 225702013-03-18 Richard Biener <rguenther@suse.de>
22571
22572 PR tree-optimization/56210
22573 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22574 Handle string / character search functions.
22575 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22576
31348d52
RB		 225772013-03-18 Richard Biener <rguenther@suse.de>
22578
22579 PR middle-end/56483
22580 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22581 and implement properly.
22582 * gimple.h (gimple_cond_single_var_p): Remove.
22583
fcac74a1
RB		 225842013-03-18 Richard Biener <rguenther@suse.de>
22585
22586 * tree-data-ref.h (find_data_references_in_loop): Declare.
22587 * tree-data-ref.c (get_references_in_stmt): Use a stack
22588 vector pre-allocated in the callers.
22589 (find_data_references_in_stmt): Adjust.
22590 (graphite_find_data_references_in_stmt): Likewise.
22591 (create_rdg_vertices): Likewise.
22592 (find_data_references_in_loop): Export.
22593 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22594 Compute dependences here...
22595 (vect_analyze_data_refs): ...not here. When we encounter
22596 a non-vectorizable data reference in basic-block vectorization
22597 truncate the data reference vector. Do not bother to
22598 fixup data-dependence information for gather loads.
22599 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22600 of data references, as reported.
22601
0d5a1b56
RB		 226022013-03-18 Richard Biener <rguenther@suse.de>
22603
22604 PR tree-optimization/3713
22605 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22606 has_constants and expr.
22607 (stmt_has_constants): Properly valueize SSA names when deciding
22608 whether the stmt has constants.
22609
789c34e3
RB		 226102013-03-18 Richard Biener <rguenther@suse.de>
22611
22612 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22613 whole function when there is nothing to do.
22614 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22615 * tree-vectorizer.c (vectorize_loops): Update virtual and
22616 loop-closed SSA once.
22617 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22618
076b4605
RB		 226192013-03-18 Richard Biener <rguenther@suse.de>
22620
22621 PR middle-end/56113
22622 * domwalk.c (bb_postorder): New global static.
22623 (cmp_bb_postorder): New function.
22624 (walk_dominator_tree): Replace scheme imposing an order for
22625 visiting dominator sons by one sorting them at the time they
22626 are pushed on the stack.
22627
bdb01696
RB		 226282013-03-18 Richard Biener <rguenther@suse.de>
22629
22630 PR tree-optimization/39326
22631 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22632 (struct mem_ref): Replace mem member with ao_ref typed member.
22633 (MEM_ANALYZABLE): Adjust.
22634 (memref_eq): Likewise.
22635 (mem_ref_alloc): Likewise.
22636 (gather_mem_refs_stmt): Likewise.
22637 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22638 (execute_sm_if_changed_flag_set): Adjust.
22639 (execute_sm): Likewise.
22640 (ref_always_accessed_p): Likewise.
22641 (refs_independent_p): Likewise.
22642 (can_sm_ref_p): Likewise.
22643
12d2dc5e
JJ		 226442013-03-18 Jakub Jelinek <jakub@redhat.com>
22645
22646 PR c/56566
22647 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22648 return 1 even for !unsignedp.
22649
286fb677
UB		 226502013-03-17 Uros Bizjak <ubizjak@gmail.com>
22651
22652 * config/i386/i386.md (isa): Add x64 and nox64.
22653 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22654 (*pushtf): Enable *roF alternative for x64 isa only.
22655 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22656 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22657 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22658 (*movtf_internal): Merge from *movtf_internal_rex64 and
22659 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22660 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22661 nox64 isa attributes.
22662 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22663 nox64 isa attributes.
22664 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22665
88b97037
UB		 226662013-03-17 Uros Bizjak <ubizjak@gmail.com>
22667
22668 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22669
9f4f1735
JJ		 226702013-03-17 Jakub Jelinek <jakub@redhat.com>
22671
22672 PR target/56640
22673 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22674 class names. Remove trailing comma after "ALL_REGS".
22675
48f4a6fa
JH		 226762013-03-16 Jan Hubicka <jh@suse.cz>
22677
22678 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22679 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22680 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22681 of cgraph_get_create_node.
22682 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22683
98e81123
JM		 226842013-03-16 Jason Merrill <jason@redhat.com>
22685
22686 PR debug/49090
22687 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22688 with DW_AT_default_value.
22689
8a298c04
JJ		 226902013-03-16 Jakub Jelinek <jakub@redhat.com>
22691
22692 * BASE-VER: Set to 4.9.0.
22693
4323afa0
AK		 226942013-03-14 Andi Kleen <ak@linux.intel.com>
22695
22696 PR target/56619
22697 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22698 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22699 Document _x* TSX intrinsics.
22700
b3c0d469
JJ		 227012013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22702 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 22703
22704 * configure.ac: Add MicroBlaze TLS support detection.
22705 * configure: Regenerate.
88b97037
UB		 22706 * config/microblaze/microblaze-protos.h
22707 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22708 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22709 * config/microblaze/microblaze.c (microblaze_address_type): Add
22710 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 22711 (microblaze_address_info): Add tls_reloc.
22712 (TARGET_HAVE_TLS): Define.
22713 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22714 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 22715 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22716 load_tls_operand, microblaze_call_tls_get_addr,
22717 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 22718 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22719 (get_base_reg): Use microblaze_tls_symbol_p.
22720 (microblaze_classify_address): Handle TLS.
88b97037
UB		 22721 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22722 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 22723 (microblaze_legitimize_address): Handle TLS.
22724 (microblaze_address_insns): Handle ADDRESS_TLS.
22725 (pic_address_needs_scratch): Handle TLS.
22726 (print_operand_address): Handle TLS.
22727 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22728 (microblaze_expand_move): Handle TLS.
88b97037
UB		 22729 (microblaze_legitimate_constant_p): Check
22730 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 22731 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22732 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22733 (PIC_OFFSET_TABLE_REGNUM): Set.
22734 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22735 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22736 (addsi3, movsi_internal2, movdf_internal): Update constraints
22737 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 22738 (move_operand): Redefine as move_src_operand,
22739 check microblaze_tls_referenced_p.
8cc9a5a5 22740
d803a491
IB		 227412013-03-14 Ian Bolton <ian.bolton@arm.com>
22742
22743 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22744 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22745
a8504f22
IB		 227462013-03-14 Ian Bolton <ian.bolton@arm.com>
22747
22748 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22749 CC mode for AND.
22750
df2dfaea
JJ		 227512013-03-14 Jakub Jelinek <jakub@redhat.com>
22752
fbd28bc3
JJ		 22753 PR tree-optimization/53265
22754 * common.opt (Waggressive-loop-optimizations): New option.
22755 * tree-ssa-loop-niter.c: Include tree-pass.h.
22756 (do_warn_aggressive_loop_optimizations): New function.
22757 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22758 if number_of_latch_executions returned constant.
22759 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22760 early. If number_of_latch_executions returned constant, set
22761 nb_iterations_upper_bound back to it.
22762 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22763 field.
22764 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22765 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22766
df2dfaea
JJ		 22767 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22768 (MULTILIB_OSDIRNAMES): Set.
22769 * genmultilib: If defaultosdirname doesn't start with :: , set
22770 defaultosdirname2 instead, clear it and emit two . multilib_raw
22771 entries instead of just one.
22772
ee0d2b68
KK		 227732013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22774
22775 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22776 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22777 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22778 (SUBTARGET_OVERRIDE_OPTIONS): New.
22779
decc676e
OE		 227802013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22781
22782 PR target/49880
22783 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22784 (musermode): Convert to Var(TARGET_USERMODE).
22785 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22786 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22787 * config/sh/sh.c (sh_option_override): Use
22788 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22789 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22790 condition.
22791 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22792 TARGET_SH4.
22793 (udivsi3_i4_single, divsi3_i4_single): Use
22794 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22795
f4b719c7
DK		 227962013-03-13 Dave Korn <dave.korn.cygwin@....>
22797
22798 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22799 default setting.
22800
c40eced0
RB		 228012013-03-13 Richard Biener <rguenther@suse.de>
22802
22803 PR tree-optimization/56608
22804 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22805 calls when vectorizing basic-blocks.
22806
1bfa5973
JJ		 228072013-03-13 Jakub Jelinek <jakub@redhat.com>
22808
22809 PR plugins/45078
22810 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22811 tm_file.
22812
6fcf5434
JJ		 228132013-03-12 Jakub Jelinek <jakub@redhat.com>
22814
22815 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22816
4f38fa8c
JH		 228172013-03-11 Jan Hubicka <jh@suse.cz>
22818
22819 PR lto/56557
22820 * lto-streamer-out.c (output_symbol_p): Skip references from
22821 constructors of external variables.
22822
c5c5ba89
JH		 228232013-03-11 Jan Hubicka <jh@suse.cz>
22824
22825 PR middle-end/56571
22826 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22827 from pseudos.
22828 * emit-rtl.c (verify_rtx_sharing): Likewise.
22829 (copy_insn_1): Likewise.
22830 * rtl.c (copy_rtx): Likewise.
22831
c2a939b1
GJL		 228322013-03-11 Georg-Johann Lay <avr@gjlay.de>
22833
22834 PR target/56591
22835 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22836 output_operand_lossage message.
88b97037 22837
3c3279fb
RE		 228382013-03-11 Richard Earnshaw <rearnsha@arm.com>
22839
22840 PR target/56470
22841 * arm.c (shift_op): Validate RTL pattern on the fly.
22842 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22843 the RTL.
22844
aef5ef9d
JDA		 228452013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22846
22847 PR target/56347
22848 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22849 new call patterns that clobber %fr12.
22850 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22851 split and postreload patterns.
22852 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22853 registers %fr12 and %fr12R as call used.
22854
456610d3
SB		 228552013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22856
22857 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22858 (canon_address, record_store, replace_read, check_mem_read_rtx,
22859 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22860 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22861 rest_of_handle_dse): Likewise.
22862
4b1baac8
RS		 228632013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22864
22865 PR middle-end/56524
22866 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22867 Add base_optabs.
22868 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22869 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22870 (save_optabs_if_changed): Replace with...
22871 (init_tree_optimization_optabs): ...this.
22872 * optabs.c (save_optabs_if_changed): Rename to...
22873 (init_tree_optimization_optabs): ...this. Take the optimization node
22874 as argument. Do nothing if the base optabs are already correct.
22875 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22876 to recompute optabs.
22877 * function.h (function): Remove optabs field.
22878 * function.c (invoke_set_current_function_hook): Call
22879 init_tree_optimization_optabs. Use the result to initialize
22880 this_fn_optabs.
22881
b7a78683
AH		 228822013-02-27 Aldy Hernandez <aldyh@redhat.com>
22883
22884 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22885 if GTMA_HAS_NO_INSTRUMENTATION.
22886 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22887 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22888 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22889 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22890 GTMA_HAS_NO_INSTRUMENTATION.
22891
6384c29b
JJ		 228922013-03-08 Jakub Jelinek <jakub@redhat.com>
22893
22894 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22895 libasan_preinit.o.
22896
ecd4f20a
MP		 228972013-03-08 Marek Polacek <polacek@redhat.com>
22898 Jakub Jelinek <jakub@redhat.com>
22899
22900 PR tree-optimization/56478
22901 * predict.c (is_comparison_with_loop_invariant_p): Change the
22902 type of loop_step to tree.
22903 (predict_loops): Adjust.
22904 (predict_iv_comparison): Perform the computations on double_ints.
22905
64366d35
RB		 229062013-03-08 Richard Biener <rguenther@suse.de>
22907
22908 PR tree-optimization/56570
22909 * tree-cfg.c (verify_expr_location_1): Verify locations for
22910 DECL_DEBUG_EXPR.
22911 * tree-sra.c (create_access_replacement): Strip locations
22912 from DECL_DEBUG_EXPRs.
22913
a9d5a059
RB		 229142013-03-08 Richard Biener <rguenther@suse.de>
22915
22916 * tree-inline.c (expand_call_inline): Do not associate
22917 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22918 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22919
b7aa4e9a
RB		 229202013-03-08 Richard Biener <rguenther@suse.de>
22921
22922 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22923 or block changes with -Og. Fix for location / block encoding
22924 changes and PHI arguments with locations.
22925
c4c2f9fa
SB		 229262013-03-07 Steven Bosscher <steven@gcc.gnu.org>
22927
22928 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
22929 for all counters.
22930 (struct output_info): Likewise.
22931 (register_overhead): Remove bad gcc_assert.
22932 (bitmap_find_bit): If there is only a single bitmap element, do not
22933 count a miss as a search.
22934 (print_statistics): Update for counter type changes.
22935 (dump_bitmap_statistics): Likewise. Print headers such that they
22936 are properly lined up with the printed counters.
22937
5bf6606a
JJ		 229382013-03-07 Jakub Jelinek <jakub@redhat.com>
22939
22940 PR tree-optimization/56559
22941 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
22942 check that it has only a single use.
22943
2c653d46
RB		 229442013-03-07 Richard Biener <rguenther@suse.de>
22945
22946 * doc/invoke.texi (fwhole-program): Discourage use in combination
22947 with -flto.
22948
a72d8780
JJ		 229492013-03-06 Jakub Jelinek <jakub@redhat.com>
22950
01a454df
JJ		 22951 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
22952
c1781047
JJ		 22953 PR tree-optimization/56539
22954 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
22955 instead of GSI_CONTINUE_LINKING as last argument to
22956 force_gimple_operand_gsi. Adjust function comment.
22957
9772c47a
JJ		 22958 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
22959 aarch64-cores.def.
22960
a72d8780
JJ		 22961 PR middle-end/56548
22962 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
22963 promoted mode, convert the result back to the original mode.
22964
fa5556de
RB		 229652013-03-06 Richard Biener <rguenther@suse.de>
22966
22967 PR middle-end/56294
22968 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
22969 (insert_updated_phi_nodes_compare_uids): New function.
22970 (update_ssa): Sort symbols_to_rename after UID before
22971 traversing it to insert PHI nodes.
22972
010403d1
RB		 229732013-03-06 Richard Biener <rguenther@suse.de>
22974
22975 PR middle-end/50494
22976 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
22977 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
22978
22979 Revert
22980 2013-02-13 Richard Biener <rguenther@suse.de>
22981
22982 PR lto/50494
22983 * varasm.c (output_constant_def_1): Get the decl representing
22984 the constant as argument.
22985 (output_constant_def): Wrap output_constant_def_1.
22986 (make_decl_rtl): Use output_constant_def_1 with the decl
22987 representing the constant.
22988 (build_constant_desc): Optionally re-use a decl already
22989 representing the constant.
22990 (tree_output_constant_def): Adjust.
22991
3c27ce4c
JY		 229922013-03-06 Joey Ye <joey.ye@arm.com>
22993
22994 PR lto/50293
22995 * gcc.c (convert_white_space): New function.
22996 (main): Handles white space in function name.
22997
8f6d1c86
OE		 229982013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
22999
23000 PR target/56529
23001 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23002 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23003 to SH_DIV_CALL_TABLE for TARGET_SH2.
23004 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23005 list.
23006 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23007 call-table options.
23008
798a209f
SA		 230092013-03-05 Sterling Augustine <saugustine@google.com>
23010 Cary Coutant <ccoutant@google.com>
23011
23012 PR debug/55364
23013 * dwarf2out.c (resolve_addr): Don't call
23014 remove_loc_list_addr_table_entries a second time for the same
23015 expression.
23016
6cfa417f 230172013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 23018
6cfa417f
JJ		 23019 PR debug/56510
23020 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23021 (avoid_complex_debug_insns): New function.
23022 (expand_debug_locations): Call it.
23023
23024 PR rtl-optimization/56484
23025 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23026 lifetimes of hard registers on small register class machines.
23027
230282013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23029
23030 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 23031 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 23032 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 23033 fast_interrupt.
23034 (microblaze_fast_interrupt_function_p): New function.
23035 (microblaze_is_interrupt_handler): Rename to
23036 microblaze_is_interrupt_variant and add fast_interrupt check.
23037 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23038 (save_restore_insns): Likewise.
23039 (compute_frame_size): Likewise.
23040 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23041 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 23042 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23043 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 23044 microblaze_is_interrupt_variant.
23045
731aefac
KT		 230462013-03-05 Kai Tietz <ktietz@redhat.com>
23047
23048 * sdbout.c (sdbout_one_type): Switch to current function's section
23049 supporting cold/hot.
23050
a72d8780 230512013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 23052
23053 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23054 -mxl-reorder.
23055
334e71e8
JJ		 230562013-03-05 Jakub Jelinek <jakub@redhat.com>
23057
0b50e654
JJ		 23058 PR middle-end/56461
23059 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23060 if VALGRIND_GET_VBITS is defined, temporarily make object
23061 memory all defined, and restore previous valgrind addressability
23062 and definability afterwards. Free this_object at the end.
23063
4ccf8f43
JJ		 23064 PR middle-end/56461
23065 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23066 right before calling lra_create_live_ranges, also call it
23067 when clearing live_p. Only call lra_clear_live_ranges
23068 at the end if live_p.
23069
334e71e8
JJ		 23070 PR middle-end/56461
23071 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23072
9ca966ca
RB		 230732013-03-05 Richard Biener <rguenther@suse.de>
23074
23075 PR tree-optimization/56521
23076 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23077 value-id.
23078
d273b176
SB		 230792013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23080
23081 PR c++/55135
23082 * except.h (remove_unreachable_eh_regions): New prototype.
23083 * except.c (remove_eh_handler_splicer): New function, split out
23084 of remove_eh_handler.
23085 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23086 warning about running it on many EH regions one at a time.
23087 (remove_unreachable_eh_regions_worker): New function, walk the
23088 EH tree in depth-first order and remove non-marked regions.
23089 (remove_unreachable_eh_regions): New function.
23090 * tree-eh.c (mark_reachable_handlers): New function, split out
23091 from remove_unreachable_handlers.
23092 (remove_unreachable_handlers): Use mark_reachable_handlers and
23093 remove_unreachable_eh_regions.
23094 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23095 and remove_unreachable_eh_regions.
23096
5e7f1aef
RB		 230972013-03-05 Richard Biener <rguenther@suse.de>
23098
23099 PR middle-end/56525
23100 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23101 not freeing them until the end.
23102
f276b762
AK		 231032013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23104
23105 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23106
12a43ab8
RB		 231072013-03-05 Richard Biener <rguenther@suse.de>
23108
23109 PR tree-optimization/56270
23110 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23111 of loads after scheduling an SLP instance.
23112
93675444
JJ		 231132013-03-05 Jakub Jelinek <jakub@redhat.com>
23114
db4138e3
JJ		 23115 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23116 tic6x.exp.
23117 (check_gcc_parallelize): Run guality.exp as a separate job from
23118 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23119 struct-layout-1.exp with stackalign.exp.
23120
dd3d1ec0
JJ		 23121 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23122
b8d381a3
JJ		 23123 PR middle-end/56461
23124 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23125 load_index sbitmap even if some bit in it isn't set.
23126
b4f9786b
JJ		 23127 PR middle-end/56461
23128 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23129 (discover_iteration_bound_by_body_walk): Change queues to
23130 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23131 spelling in comment. Call safe_push on queues[bound_index] directly.
23132 Release queues[queue_index] in every iteration unconditionally.
23133 Release bounds vector.
23134
93675444
JJ		 23135 PR middle-end/56461
23136 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23137 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23138 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23139 inner_phis vector.
23140
3e492e9c
RB		 231412013-03-05 Richard Biener <rguenther@suse.de>
23142
23143 PR lto/56515
23144 * tree-inline.c (remap_blocks_to_null): New function.
23145 (expand_call_inline): When expanding a call stmt without
23146 an associated block inline remap all callee blocks to NULL.
23147
a5d1569a
JJ		 231482013-03-05 Jakub Jelinek <jakub@redhat.com>
23149
86efb5cd
JJ		 23150 PR rtl-optimization/56494
23151 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23152 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23153 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23154
85f5dbea
JJ		 23155 PR middle-end/56461
23156 * sel-sched-ir.c (free_sched_pools): Release
23157 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23158 if succs_info_pool.max_top isn't -1.
23159
a5d1569a
JJ		 23160 PR bootstrap/56509
23161 * opts.c (opts_obstack, opts_concat): Moved to...
23162 * opts-common.c (opts_obstack, opts_concat): ... here.
23163
4432aa6c
JJ		 231642013-03-04 Jakub Jelinek <jakub@redhat.com>
23165
23166 PR middle-end/56461
23167 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23168
f3a1fb91
MJ		 231692013-03-04 Martin Jambor <mjambor@suse.cz>
23170
23171 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23172 all appropriate places.
23173
22110e6c
EB		 231742013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23175
23176 PR tree-optimization/56424
23177 * ipa-split.c (split_function): Do not set the RSO flag if result is
23178 not by reference and its type is a register type.
23179
a72d8780 231802013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23181
88b97037 23182 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23183 (microblaze_legitimate_pic_operand): Likewise
88b97037 23184 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 23185 new function microblaze_legitimate_pic_operand
23186 * config/microblaze/microblaze-protos.h
23187 (microblaze_legitimate_pic_operand): Declare.
23188
a72d8780 231892013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23190
a72d8780 23191 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23192 New predicate for supported rtx code types.
a72d8780 23193 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 23194 call_insn_simple_operand predicate.
23195
541d9ac8
JJ		 231962013-03-04 Jakub Jelinek <jakub@redhat.com>
23197
5eb010bc
JJ		 23198 PR middle-end/56461
23199 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23200 partitions.ordered_remove.
23201
30862efc
JJ		 23202 PR middle-end/56461
23203 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23204 vec_oprnds0.create (1) for modifier == NONE.
23205
8930f723
JJ		 23206 PR middle-end/56461
23207 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23208 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23209 vec_oprnds1 right before pushing anything to it for
23210 scalar_shift_arg.
23211
541d9ac8
JJ		 23212 PR middle-end/56461
23213 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23214 set nbbs to 0 instead of having separate code path.
23215 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23216 instead of false as last argument if returning NULL.
23217
7aa7f2e3
SL		 232182013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23219
23220 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23221 the attribute is now called "target" instead of "option".
23222 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23223 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23224 attribute/pragma name for TARGET_OPTION_VALID_P and
23225 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23226 * doc/tm.texi: Regenerated.
23227
8930f723 232282013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23229
8930f723 23230 * config/microblaze/microblaze.c:
cb8a1637 23231 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 23232 * config/microblaze/microblaze.h: Add -mxl-reorder to
23233 DRIVER_SELF_SPECS.
23234 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23235 instructions emitted if TARGET_REORDER.
88b97037
UB		 23236 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23237 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23238 separately.
cb8a1637 23239
4dc1d68c
XDL		 232402013-03-01 Xinliang David Li <davidxl@google.com>
23241
23242 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23243 walk length.
23244
689eaba3
JJ		 232452013-03-01 Jakub Jelinek <jakub@redhat.com>
23246
9d676bf6
JJ		 23247 PR middle-end/56461
23248 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23249 vector even when returning true. Fix up function comment formatting.
23250
caff7edf
JJ		 23251 PR middle-end/56461
23252 * ira-build.c (ira_loop_nodes_count): New variable.
23253 (create_loop_tree_nodes): Initialize it.
23254 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23255
b6b9227d
JJ		 23256 PR middle-end/56461
23257 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23258 method on dr_chain and result_chain.
23259 * tree-vect-stmts.c (vectorizable_store): Only call
23260 result_chain.create if j == 0.
23261
689eaba3
JJ		 23262 PR middle-end/56461
23263 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23264 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23265 before overwriting it.
23266
40bf31ed
TB		 232672013-03-01 Tobias Burnus <burnus@net-b.de>
23268
23269 * doc/extended.texi (C Extensions): Change order in @menu
23270 to match @node.
23271 (Other MIPS Built-in Functions): Move last MIPS entry before
23272 "picoChip Built-in Functions".
23273 (SH Built-in Functions): Move after RX Built-in Functions.
23274 * doc/gcc.texi (Introduction): Change order in @menu
23275 to match @node.
23276 * doc/md.texi (Constraints): Ditto.
23277 * gty.texi (Type Information): Ditto.
23278 (User-provided marking routines for template types): Make
23279 subsection.
23280 * doc/invoke.texi (AArch64 Options): Move before
23281 "Adapteva Epiphany Options".
23282
e664c61c
KS		 232832013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23284 Jakub Jelinek <jakub@redhat.com>
23285
23286 PR sanitizer/56454
23287 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23288 no_address_safety_analysis attribute.
23289 * doc/extend.texi (no_address_safety_attribute): Rename to
23290 no_sanitize_address attribute, mention no_address_safety_analysis
23291 attribute as deprecated alias.
23292
37b5ec8f
JJ		 232932013-02-28 Jakub Jelinek <jakub@redhat.com>
23294
23295 PR middle-end/56461
23296 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23297 type to vec<vec<tree> > *.
23298 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23299 to be vec<tree> instead of vec<tree> *, set vec_defs
23300 to vNULL and call vec_defs.create (number_of_vects), adjust other
23301 uses of vec_defs.
23302 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23303 vectorizable_condition): Adjust vect_get_slp_defs callers.
23304
ba96cdfb
JG		 233052013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23306
23307 * config/aarch64/aarch64.c
23308 (aarch64_float_const_representable): Remove unused variable.
23309
6f549691
JG		 233102013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23311
23312 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23313
af55e82d
JG		 233142013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23315
23316 * config/aarch64/aarch64-builtins.c
23317 (aarch64_init_simd_builtins): Make static.
23318
1df3f464
JG		 233192013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23320
23321 * config/aarch64/aarch64.c
23322 (aarch64_simd_make_constant): Make static.
23323
f8f42513
MJ		 233242013-02-28 Martin Jambor <mjambor@suse.cz>
23325
23326 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23327 with no initialization to the RHS of debug statements.
23328
b48b3fc4
MJ		 233292013-02-28 Martin Jambor <mjambor@suse.cz>
23330
23331 PR tree-optimization/56294
23332 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23333 Adjust dumping.
23334 (get_access_replacement): Do not call create_access_replacement.
23335 Assert a replacement exists.
23336 (get_repl_default_def_ssa_name): Create the replacement declaration
23337 itself.
23338
c3ae224c
RR		 233392013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23340
23341 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23342 final_end_function.
23343
45fa0eef
MP		 233442013-02-28 Marek Polacek <polacek@redhat.com>
23345
23346 PR rtl-optimization/56466
23347 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23348 if we're changing a loop.
23349 (peel_loops_completely): Likewise.
23350
502c067d
PC		 233512013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23352
23353 PR c++/55813
23354 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23355
ff24afc8
GJL		 233562013-02-28 Georg-Johann Lay <avr@gjlay.de>
23357
23358 PR target/56445
23359 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23360 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23361 INTX_FTYPE_FX, FX_FTYPE_INTX.
23362 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23363
c7b1fc1b
GJL		 233642013-02-28 Georg-Johann Lay <avr@gjlay.de>
23365
23366 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23367 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23368 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23369 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23370 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23371 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23372 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23373 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23374 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23375 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23376 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23377 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23378 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23379 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23380 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23381 (avrxmega6): Increase max flash segments from 5 to 6.
23382 * config/avr/t-multilib: Regenerate.
23383 * config/avr/avr-tables.opt: Regenerate.
23384 * doc/avr-mmcu.texi: Regenerate.
23385
4a0e3cfe
GJL		 233862013-02-28 Georg-Johann Lay <avr@gjlay.de>
23387
23388 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23389 (avr_device_to_arch): Rename to avr_device_to_ld.
23390 (avr_device_to_as): New prototype.
23391 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23392 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23393 * config/avr/driver-avr.c (avr_device_to_as): New.
23394 (avr_device_to_arch): Rename to avr_device_to_ld.
23395
97785e52
JJ		 233962013-02-27 Jakub Jelinek <jakub@redhat.com>
23397
3f292312
JJ		 23398 PR middle-end/56461
23399 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23400 method on dr_chain and result_chain.
23401
a344216b
JJ		 23402 PR middle-end/56461
23403 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23404 pointer_set_destroy on not_executed_last_iteration.
23405
f121ad02 23406 PR middle-end/56461
88b97037 23407 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 23408
307f83a3
JJ		 23409 PR middle-end/56461
23410 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23411 FOR_EACH_DEFINED_FUNCTION when freeing state.
23412
e19624ee
JJ		 23413 PR middle-end/56461
23414 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23415 pool_free.
23416 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23417 overwriting it.
23418
90e709fd
JJ		 23419 PR middle-end/56461
23420 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23421 known_aggs[i].items and release known_aggs vector.
23422
97785e52
JJ		 23423 PR middle-end/56461
23424 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23425
234262013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 23427
97785e52
JJ		 23428 * config/microblaze/microblaze.c (microblaze_emit_compare):
23429 Use xor for EQ/NE comparisions.
23430 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23431 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 23432
6fa5e0ed
JJ		 234332013-02-27 Jakub Jelinek <jakub@redhat.com>
23434
23435 PR middle-end/56461
23436 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23437 vector.
23438 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23439 vec_safe_push, always update *slot.
23440 (redirect_edge_var_map_clear): Use vec_free.
23441 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23442 (free_var_map_entry): Use vec_free.
23443 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23444 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23445
436a956a
AB		 234462013-02-27 Andrey Belevantsev <abel@ispras.ru>
23447
23448 PR middle-end/45472
436a956a
AB		 23449 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23450 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 23451 Reorder tests for speculativeness in the logical and operator.
23452
f5c2caca
JJ		 234532013-02-27 Jakub Jelinek <jakub@redhat.com>
23454
0fcb564b
JJ		 23455 * incpath.c (add_standard_paths): Use reconcat instead of concat
23456 where appropriate and avoid leaking memory.
23457
dc357798
JJ		 23458 * opts.h: Include obstack.h.
23459 (opts_concat): New prototype.
23460 (opts_obstack): New declaration.
23461 * opts.c (opts_concat): New function.
23462 (opts_obstack): New variable.
23463 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23464 (finish_options): Use opts_concat instead of concat
23465 and XOBNEWVEC instead of XNEWVEC.
23466 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23467 generate_option): Likewise.
23468 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23469 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23470
f5c2caca
JJ		 23471 PR target/56455
23472 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23473 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23474
d0163673
JJ		 234752013-02-26 Jakub Jelinek <jakub@redhat.com>
23476
23477 PR middle-end/56461
23478 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23479
b5ad2b8e
JR		 234802013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23481
23482 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23483 (arm_block_move_unaligned_straight): Likewise.
23484 (arm_adjust_block_mem): Likewise.
23485
9b639e2c
JR		 234862013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23487
23488 PR target/48901
23489 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23490 temp, cond and label.
23491 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23492
e13a0ccb
JR		 23493 PR target/52500
23494 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23495 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23496
fbe4f171
JR		 23497 PR target/52501
23498 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23499 of prologue/epilogue functions.
23500
ae006543
JR		 23501 PR target/52550
23502 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23503 Remove unused variable cfa_offset.
23504 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23505
c81369fa
JR		 23506 PR target/54639
23507 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23508 type promotion to unsigned.
23509
f8a8fea7
JR		 23510 PR target/54640
23511 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23512 for HOST_WIDE_INT of 32 bit / same size as int.
23513 (arm_block_move_unaligned_straight): Likewise.
23514 (arm_adjust_block_mem): Likewise.
23515
f8be5169
JR		 23516 PR target/54662
23517 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23518 ALL_CFLAGS.
23519
f1ad3354
MP		 235202013-02-26 Marek Polacek <polacek@redhat.com>
23521
23522 PR tree-optimization/56426
88b97037 23523 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 23524
a6af0f42
RB		 235252013-02-26 Richard Biener <rguenther@suse.de>
23526
23527 PR target/56444
23528 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23529 unused variable loops.
23530
aca43c6c
JJ		 235312013-02-26 Jakub Jelinek <jakub@redhat.com>
23532
cecbe5d9
JJ		 23533 PR tree-optimization/56448
23534 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23535 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23536 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23537 later operands of the references, or even first operand for
23538 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23539
aca43c6c
JJ		 23540 PR tree-optimization/56443
23541 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23542 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23543 to type_for_mode langhook.
23544
53e2e141
MT		 235452013-02-25 Matt Turner <mattst88@gmail.com>
23546
23547 * doc/invoke.texi: Document r4700.
23548
259ee451
RB		 235492013-02-25 Richard Biener <rguenther@suse.de>
23550
23551 PR tree-optimization/56175
23552 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23553 split out from ...
23554 (simplify_bitwise_binary): ... here. Also guard the conversion
23555 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23556
6d65e8f1
CM		 235572013-02-25 Catherine Moore <clm@codesourcery.com>
23558
23559 Revert:
23560 2013-02-24 Catherine Moore <clm@codesourcery.com>
23561 Maciej W. Rozycki <macro@codesourcery.com>
23562 Tom de Vries <tom@codesourcery.com>
a72d8780 23563 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 23564 Iain Sandoe <iain@codesourcery.com>
23565 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23566 Chao-ying Fu <fu@mips.com>
88b97037 23567
6d65e8f1 23568 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23569 Document new function attributes.
6d65e8f1
CM		 23570 * doc/invoke.texi (minterlink-compressed, mmicromips,
23571 m14k, m14ke, m14kec): Document new options.
23572 (minterlink-mips16): Update documentation.
23573 * doc/md.texi (ZC, ZD): Document new constraints.
23574 * configure.ac (gcc_cv_as_micromips): Check if linker
23575 supports the .set micromips directive.
23576 * configure: Regenerate.
23577 * config.in: Regenerate.
23578 * config/mips/mips-tables.opt: Regenerate.
23579 * config/mips/micromips.md: New file.
23580 * constraints.md (ZC, AD): New constraints.
23581 * config/mips/predicates.md (movep_src_register): New predicate.
23582 (movep_src_operand): New predicate.
23583 (non_volatile_mem_operand): New predicate.
23584 * config/mips/mips.md (multimem): New type.
23585 (length): Differentiate between 17-bit and 18-bit branch offsets.
23586 (MOVEP1, MOVEP2): New mode iterator.
23587 (mov_<load>l): Use ZC constraint.
23588 (mov_<load>r): Likewise.
23589 (mov_<store>l): Likewise.
23590 (mov_<store>r): Likewise.
23591 (*branch_equality<mode>_inverted): Add microMIPS support.
23592 (*branch_equality<mode>): Likewise.
23593 (*jump_absolute): Likewise.
23594 (indirect_jump_<mode>): Likewise.
23595 (tablejump_<mode>): Likewise.
23596 (<optab>_internal): Likewise.
23597 (sibcall_internal): Likewise.
23598 (sibcall_value_internal): Likewise.
23599 (prefetch): Use constraint ZD.
23600 * config/mips/mips.opt (minterlink-compressed): New option.
23601 (minterlink-mips16): Now an alias for minterlink-compressed.
23602 (mmicromips): New option.
23603 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23604 (compare_and_swap_12): Likewise.
23605 (sync_add<mode>): Likewise.
23606 (sync_<optab>_12): Likewise.
23607 (sync_old_<optab>_12): Likewise.
23608 (sync_new_<optab>_12): Likewise.
23609 (sync_nand_12): Likewise.
23610 (sync_old_nand_12): Likewise.
23611 (sync_new_nand_12): Likewise.
23612 (sync_sub<mode>): Likewise.
23613 (sync_old_add<mode>): Likewise.
23614 (sync_old_sub<mode>): Likewise.
23615 (sync_new_add<mode>): Likewise.
23616 (sync_new_sub<mode>): Likewise.
23617 (sync_<optab><mode>): Likewise.
23618 (sync_old_<optab><mode>): Likewise.
23619 (sync_new_<optab><mode>): Likewise.
23620 (sync_nand<mode>): Likewise.
23621 (sync_old_nand<mode>): Likewise.
23622 (sync_new_nand<mode>): Likewise.
23623 (sync_lock_test_and_set<mode>): Likewise.
23624 (test_and_set_12): Likewise.
23625 (atomic_compare_and_swap<mode>): Likewise.
23626 (atomic_exchange<mode>_llsc): Likewise.
23627 (atomic_fetch_add<mode>_llsc): Likewise.
23628 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23629 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23630 (umips_save_restore_pattern_p): Likewise.
23631 (umips_load_store_pair_p): Likewise.
23632 (umips_output_load_store_pair): Likewise.
23633 (umips_movep_target_p): Likewise.
23634 (umips_12bit_offset_address_p): Likewise.
23635 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23636 (mips_base_mips16): Rename this...
23637 (mips_base_compression_flags): ...to this. Update all uses.
23638 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23639 (mips_mips16_decl_p): Delete.
23640 (mips_nomips16_decl_p): Delete.
23641 (mips_get_compress_on_flags): New function.
23642 (mips_get_compress_off_flags): New function.
23643 (mips_get_compress_mode): New function.
23644 (mips_get_compress_on_name): New function.
23645 (mips_get_compress_off_name): New function.
23646 (mips_insert_attributes): Support multiple compression types.
23647 (mips_merge_decl_attributes): Likewise.
23648 (umips_12bit_offset_address_p): New function.
23649 (mips_start_function_definition): Emit .set micromips directive.
23650 (mips_call_may_need_jalx_p): New function.
23651 (mips_function_ok_for_sibcall): Add microMIPS support.
23652 (mips_print_operand_punctuation): Support short delay slots and
23653 compact jumps.
23654 (umips_swm_mask, umips_swm_encoding): New.
23655 (umips_build_save_restore): New function.
23656 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23657 (was_mips16_p): Remove.
23658 (old_compression_mode): New.
23659 (mips_set_compression_mode): New function.
23660 (mips_set_current_function): Add microMIPS support.
23661 (mips_option_override): Likewise.
23662 (umips_save_restore_pattern_p): New function.
23663 (umips_output_save_restore): New function.
23664 (umips_load_store_pair_p_1): New function.
23665 (umips_load_store_pair_p): New function.
23666 (umips_output_load_store_pair_1): New function.
23667 (umips_output_load_store_pair): New function.
23668 (umips_movep_target_p) New function.
23669 (mips_prepare_pch_save): Add microMIPS support.
23670 * config/mips/mips.h (TARGET_COMPRESSION): New.
23671 (TARGET_CPU_CPP_BUILTINS): Update macro
23672 to use new compression flags and to support microMIPS.
23673 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23674 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23675 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23676 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23677 (ASM_SPEC): Support mmicromips and mno-micromips.
23678 (M16STORE_REG_P): New macro.
23679 (MIPS_CALL): Support TARGET_MICROMIPS.
23680 (MICROMIPS_J): New macro.
23681 (mips_base_mips16): Rename this...
23682 (mips_base_compression_flags): ...to this.
23683 (UMIPS_12BIT_OFFSET_P): New macro.
23684 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23685 (MULTILIB_DIRNAMES): Likewise.
23686
5e5df392
TV		 236872013-02-25 Tom de Vries <tom@codesourcery.com>
23688
23689 PR rtl-optimization/56131
23690 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23691 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23692 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23693
3fdb53c1
TB		 236942013-02-25 Tobias Burnus <burnus@net-b.de>
23695
23696 * doc/invoke.texi (-fsanitize=): Move from optimization
23697 to debugging options.
23698
ed358aea
AB		 236992013-02-25 Andrey Belevantsev <abel@ispras.ru>
23700
23701 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23702
f45e9053
AB		 237032013-02-25 Andrey Belevantsev <abel@ispras.ru>
23704 Alexander Monakov <amonakov@ispras.ru>
23705
23706 PR middle-end/56077
23707 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23708 flush pending lists also on non-jumps. Adjust comment.
23709
6941b508
CM		 237102013-02-24 Catherine Moore <clm@codesourcery.com>
23711 Maciej W. Rozycki <macro@codesourcery.com>
23712 Tom de Vries <tom@codesourcery.com>
a72d8780 23713 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 23714 Iain Sandoe <iain@codesourcery.com>
23715 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23716 Chao-ying Fu <fu@mips.com>
6941b508
CM		 23717
23718 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23719 Document new function attributes.
6941b508
CM		 23720 * doc/invoke.texi (minterlink-compressed, mmicromips,
23721 m14k, m14ke, m14kec): Document new options.
23722 (minterlink-mips16): Update documentation.
23723 * doc/md.texi (ZC, ZD): Document new constraints.
23724 * configure.ac (gcc_cv_as_micromips): Check if linker
23725 supports the .set micromips directive.
23726 * configure: Regenerate.
23727 * config.in: Regenerate.
23728 * config/mips/mips-tables.opt: Regenerate.
23729 * config/mips/micromips.md: New file.
23730 * constraints.md (ZC, AD): New constraints.
23731 * config/mips/predicates.md (movep_src_register): New predicate.
23732 (movep_src_operand): New predicate.
23733 (non_volatile_mem_operand): New predicate.
23734 * config/mips/mips.md (multimem): New type.
23735 (length): Differentiate between 17-bit and 18-bit branch offsets.
23736 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 23737 (mov_<load>l): Use ZC constraint.
6941b508
CM		 23738 (mov_<load>r): Likewise.
23739 (mov_<store>l): Likewise.
23740 (mov_<store>r): Likewise.
23741 (*branch_equality<mode>_inverted): Add microMIPS support.
23742 (*branch_equality<mode>): Likewise.
23743 (*jump_absolute): Likewise.
23744 (indirect_jump_<mode>): Likewise.
23745 (tablejump_<mode>): Likewise.
23746 (<optab>_internal): Likewise.
23747 (sibcall_internal): Likewise.
23748 (sibcall_value_internal): Likewise.
23749 (prefetch): Use constraint ZD.
23750 * config/mips/mips.opt (minterlink-compressed): New option.
23751 (minterlink-mips16): Now an alias for minterlink-compressed.
23752 (mmicromips): New option.
23753 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23754 (compare_and_swap_12): Likewise.
23755 (sync_add<mode>): Likewise.
23756 (sync_<optab>_12): Likewise.
23757 (sync_old_<optab>_12): Likewise.
23758 (sync_new_<optab>_12): Likewise.
23759 (sync_nand_12): Likewise.
23760 (sync_old_nand_12): Likewise.
23761 (sync_new_nand_12): Likewise.
23762 (sync_sub<mode>): Likewise.
23763 (sync_old_add<mode>): Likewise.
23764 (sync_old_sub<mode>): Likewise.
23765 (sync_new_add<mode>): Likewise.
23766 (sync_new_sub<mode>): Likewise.
23767 (sync_<optab><mode>): Likewise.
23768 (sync_old_<optab><mode>): Likewise.
23769 (sync_new_<optab><mode>): Likewise.
23770 (sync_nand<mode>): Likewise.
23771 (sync_old_nand<mode>): Likewise.
23772 (sync_new_nand<mode>): Likewise.
23773 (sync_lock_test_and_set<mode>): Likewise.
23774 (test_and_set_12): Likewise.
23775 (atomic_compare_and_swap<mode>): Likewise.
23776 (atomic_exchange<mode>_llsc): Likewise.
23777 (atomic_fetch_add<mode>_llsc): Likewise.
23778 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23779 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23780 (umips_save_restore_pattern_p): Likewise.
23781 (umips_load_store_pair_p): Likewise.
23782 (umips_output_load_store_pair): Likewise.
23783 (umips_movep_target_p): Likewise.
23784 (umips_12bit_offset_address_p): Likewise.
23785 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23786 (mips_base_mips16): Rename this...
23787 (mips_base_compression_flags): ...to this. Update all uses.
23788 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23789 (mips_mips16_decl_p): Delete.
23790 (mips_nomips16_decl_p): Delete.
3fdb53c1 23791 (mips_get_compress_on_flags): New function.
6941b508
CM		 23792 (mips_get_compress_off_flags): New function.
23793 (mips_get_compress_mode): New function.
23794 (mips_get_compress_on_name): New function.
23795 (mips_get_compress_off_name): New function.
23796 (mips_insert_attributes): Support multiple compression types.
23797 (mips_merge_decl_attributes): Likewise.
23798 (umips_12bit_offset_address_p): New function.
23799 (mips_start_function_definition): Emit .set micromips directive.
23800 (mips_call_may_need_jalx_p): New function.
23801 (mips_function_ok_for_sibcall): Add microMIPS support.
23802 (mips_print_operand_punctuation): Support short delay slots and
23803 compact jumps.
23804 (umips_swm_mask, umips_swm_encoding): New.
23805 (umips_build_save_restore): New function.
23806 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23807 (was_mips16_p): Remove.
23808 (old_compression_mode): New.
23809 (mips_set_compression_mode): New function.
23810 (mips_set_current_function): Add microMIPS support.
23811 (mips_option_override): Likewise.
23812 (umips_save_restore_pattern_p): New function.
23813 (umips_output_save_restore): New function.
23814 (umips_load_store_pair_p_1): New function.
23815 (umips_load_store_pair_p): New function.
23816 (umips_output_load_store_pair_1): New function.
23817 (umips_output_load_store_pair): New function.
23818 (umips_movep_target_p) New function.
23819 (mips_prepare_pch_save): Add microMIPS support.
23820 * config/mips/mips.h (TARGET_COMPRESSION): New.
23821 (TARGET_CPU_CPP_BUILTINS): Update macro
23822 to use new compression flags and to support microMIPS.
23823 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23824 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23825 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23826 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23827 (ASM_SPEC): Support mmicromips and mno-micromips.
23828 (M16STORE_REG_P): New macro.
23829 (MIPS_CALL): Support TARGET_MICROMIPS.
23830 (MICROMIPS_J): New macro.
23831 (mips_base_mips16): Rename this...
23832 (mips_base_compression_flags): ...to this.
23833 (UMIPS_12BIT_OFFSET_P): New macro.
23834 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23835 (MULTILIB_DIRNAMES): Likewise.
23836
cdebbc6d
JJ		 238372013-02-24 Jakub Jelinek <jakub@redhat.com>
23838
23839 PR target/52555
23840 * target-globals.c (save_target_globals): For init_reg_sets and
23841 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23842
18c63565
JG		 238432013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23844
23845 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23846 * config/aarch64/t-aarch64
23847 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23848
2194f7a2
VM		 238492013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23850
23851 PR inline-asm/56148
23852 * lra-constraints.c (process_alt_operands): Reload operand
23853 conflicting with earlier clobber only if no more other conflicting
23854 operands.
23855
7d613735
JJ		 238562013-02-22 Jakub Jelinek <jakub@redhat.com>
23857
23858 PR sanitizer/56393
23859 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23860 if not linking a shared library.
23861
ac8d93a7
SL		 238622013-02-22 Seth LaForge <sethml@google.com>
23863
23864 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23865
e0237780
GY		 238662013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23867
3fdb53c1
TB		 23868 * config/arm/arm.md (split for extendsidi): Update condition.
23869 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23870 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23871 (qhs_zextenddi_cstr): Likewise.
e0237780 23872
d7fde18c
JJ		 238732013-02-21 Jakub Jelinek <jakub@redhat.com>
23874
be63b77d
JJ		 23875 PR middle-end/56420
23876 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23877 avoid signed wrapping.
23878 (expand_mult): Handle properly multiplication by
23879 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23880 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23881 in the compiler if coeff is HOST_WIDE_INT_MIN.
23882 (expand_divmod): Don't make ext_op1 static, change it's type to
23883 uhwi. Avoid undefined behavior in -INTVAL (op1).
23884
d7fde18c
JJ		 23885 PR rtl-optimization/50339
23886 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23887 field.
23888 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23889 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23890 into splitting_ashiftrt field.
23891 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23892 ASHIFTRT.
23893 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23894 choices.
23895
6aad4455
AH		 238962013-02-20 Aldy Hernandez <aldyh@redhat.com>
23897
23898 PR middle-end/56108
23899 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23900 are sure to go irrevocable.
23901
38fe784d
HPN		 239022013-02-21 Hans-Peter Nilsson <hp@axis.com>
23903
23904 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23905 scalars are valid operands.
23906
0fd44da3
MJ		 239072013-02-21 Martin Jambor <mjambor@suse.cz>
23908
23909 PR tree-optimization/56310
23910 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23911 only matching indices and non-negative final offsets.
23912 (intersect_aggregates_with_edge): Pass src_idx to
23913 agg_replacements_to_vector. Pass src_idx insstead of index to
23914 intersect_with_agg_replacements.
23915
7a92038b
MJ		 239162013-02-21 Martin Jambor <mjambor@suse.cz>
23917
23918 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23919 instead of hard-wired defaults.
23920
c0da9c37
MR		 239212013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23922
23923 * doc/invoke.texi (MIPS Options): Update documentation of the
23924 floating-point multiply-accumulate instruction restrictions.
23925
d247ea0c 239262013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 23927
23928 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
23929 asan_shadow_offset on x86_64 linux.
23930
22deefcb
RB		 239312013-02-21 Richard Biener <rguenther@suse.de>
23932
23933 PR tree-optimization/56415
23934 Revert
23935 2013-02-11 Richard Biener <rguenther@suse.de>
23936
23937 PR tree-optimization/56273
23938 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23939 first VRP run.
23940
7bcc6e75
JJ		 239412013-02-21 Jakub Jelinek <jakub@redhat.com>
23942
a0ad148f
JJ		 23943 PR bootstrap/56258
23944 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
23945 instead of @itemx.
23946
7bcc6e75
JJ		 23947 PR inline-asm/56405
23948 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
23949 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
23950
a0a7b611
JH		 239512013-02-20 Jan Hubicka <jh@suse.cz>
23952
23953 PR tree-optimization/56265
88b97037
UB		 23954 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
23955 when target is referenced for first time.
a0a7b611 23956
c0e50f72
RB		 239572013-02-20 Richard Biener <rguenther@suse.de>
23958
23959 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
23960 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
23961 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
23962 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
23963 not return anything.
23964 (rename_ssa_copies): Do not remove unused locals.
23965 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 23966 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 23967 * passes.c (execute_function_todo): Do not schedule unused locals
23968 removal if cleanup_tree_cfg did something.
23969 * tree-ssa-live.c (remove_unused_locals): Dump statistics
23970 about the number of removed locals.
23971
a52ca739
RB		 239722013-02-20 Richard Biener <rguenther@suse.de>
23973
23974 PR tree-optimization/56398
88b97037 23975 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 23976
ec9c9d1b
MJ		 239772013-02-20 Martin Jambor <mjambor@suse.cz>
23978
23979 PR tree-optimization/55334
23980 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
23981 restricted pointers to arrays.
23982
e91c8ed6 239832013-02-20 Richard Biener <rguenther@suse.de>
88b97037 23984 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 23985
23986 PR tree-optimization/56396
23987 * tree-ssa-ccp.c (n_const_val): New static variable.
23988 (get_value): Return NULL for SSA names we don't have a lattice
23989 entry for.
23990 (ccp_initialize): Initialize n_const_val.
23991 * tree-ssa-copy.c (n_copy_of): New static variable.
23992 (init_copy_prop): Initialize n_copy_of.
23993 (get_value): Return NULL_TREE for SSA names we don't have a
23994 lattice entry for.
23995
3d916479
MJ		 239962013-02-20 Martin Jambor <mjambor@suse.cz>
23997
23998 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
23999
71a86758
RB		 240002013-02-20 Richard Biener <rguenther@suse.de>
24001
24002 * genpreds.c (write_lookup_constraint): Do not compare first
24003 letter of the constraint again.
24004
79836a12
RB		 240052013-02-20 Richard Biener <rguenther@suse.de>
24006
24007 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24008 and ceil_log2.
24009 (get_use_iv_cost): Terminate hashtable walk when coming across
24010 an empty entry.
24011
bbe4fb2c
IZ		 240122013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24013
24014 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24015 reassociation for avx2 targets.
24016
a72d8780 240172012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 24018
c77f83d5
EI		 24019 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24020 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 24021 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 24022 version and TARGET_PATTERN_COMPARE check
24023 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 24024
a72d8780 240252012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 24026
a72d8780 24027 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 24028 function before branching.
24029
242387fa
AB		 240302012-02-19 Andrey Belevantsev <abel@ispras.ru>
24031
24032 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24033 DUMP_INSN_RTX_UID.
24034 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24035
b4979ab9
AB		 240362012-02-19 Andrey Belevantsev <abel@ispras.ru>
24037
24038 PR middle-end/55889
b4979ab9
AB		 24039 * sel-sched.c: Include ira.h.
24040 (implicit_clobber_conflict_p): New function.
24041 (moveup_expr): Use it.
88b97037 24042 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 24043
24d63016
RB		 240442013-02-19 Richard Biener <rguenther@suse.de>
24045
24046 PR tree-optimization/56384
24047 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24048 (vn_hash_type): Split out from ...
24049 (vn_hash_constant_with_type): ... here.
24050 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24051 (vn_phi_eq): Compare types from vn_phi_s structure.
24052 (vn_phi_lookup): Populate vn_phi_s type.
24053 (vn_phi_insert): Likewise.
24054
a475fd3d
JJ		 240552013-02-19 Jakub Jelinek <jakub@redhat.com>
24056
47cc28f5
JJ		 24057 PR tree-optimization/56350
24058 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24059 if haven't found reduction or nested cycle operand, rather than
24060 asserting we must find it.
24061
a475fd3d
JJ		 24062 PR tree-optimization/56381
24063 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24064 to fold_build3.
24065
240662013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 24067 Jakub Jelinek <jakub@redhat.com>
24068
24069 PR target/52555
24070 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24071 (swap_optab_enable): Same.
24072 (init_all_optabs): Use argument instead of global.
88b97037 24073 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 24074 * expr.h (init_all_optabs): Add argument to prototype.
24075 (TREE_OPTIMIZATION_OPTABS): New.
24076 (save_optabs_if_changed): Protoize.
24077 * optabs.h: Declare this_fn_optabs.
24078 * optabs.c (save_optabs_if_changed): New.
24079 Declare this_fn_optabs.
24080 (init_optabs): Add argument to init_all_optabs() call.
24081 * function.c (invoke_set_current_function_hook): Handle per
24082 function optabs.
24083 * function.h (struct function): New field optabs.
24084 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24085 optimization_current_node has changed.
24086 * target-globals.h (save_target_globals_default_opts): Protoize.
24087 * target-globals.c (save_target_globals_default_opts): New.
24088
3f587ca3
JDA		 240892013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24090
24091 PR target/56347
24092 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24093 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 24094
24095 PR target/56214
24096 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 24097 and HImode, require all displacements to be an integer multiple of
24098 their mode size.
ceaca33e
JDA		 24099 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24100 only allow QImode and HImode when reload is in progress and strict is
24101 true. Likewise for symbolic addresses. Use base14_operand to check
24102 displacements in REG+BASE addresses.
24103
fe0b4796
RB		 241042013-02-18 Richard Biener <rguenther@suse.de>
24105
24106 PR tree-optimization/56366
24107 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24108 handle sign-conversion of outer-loop initial induction value.
24109
6aaf596b
RB		 241102013-02-18 Richard Biener <rguenther@suse.de>
24111
73db8ff1 24112 PR middle-end/56349
6aaf596b
RB		 24113 * cfghooks.c (merge_blocks): If we merge a latch into another
24114 block adjust references to it.
24115 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24116 (verify_loop_structure): Verify that a recorded latch is in fact
24117 a latch.
24118
5e97dfb6
RB		 241192013-02-18 Richard Biener <rguenther@suse.de>
24120
24121 PR tree-optimization/56321
24122 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24123 order SSA name release and virtual operand unlinking.
24124
825527e8
EI		 241252013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24126
24127 * config/microblaze/microblaze.md (save_stack_block): Define.
24128 (restore_stack_block): Likewise.
24129
debd11d9
EI		 241302013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24131
24132 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24133 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24134 * config/microblaze/microblaze.c (microblaze_option_override):
24135 Bail out early for PIC modes when target does not support PIC.
24136
8ec77be0
EI		 241372013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24138
88b97037 24139 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 24140 Replace with a microblaze version.
24141 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 24142 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 24143 microblaze.
24144
b41288b3
JJ		 241452013-02-16 Jakub Jelinek <jakub@redhat.com>
24146 Dodji Seketeli <dodji@redhat.com>
24147
24148 PR asan/56330
88b97037 24149 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 24150 (instrument_mem_region_access): Do not forget to always put
24151 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24152 0) statement, even for cases where either 'base' or 'base + len'
24153 are not instrumented -- because they have been previously
24154 instrumented. Simplify the logic by putting all the statements
24155 instrument 'base + len' inside a sequence, and then insert that
24156 sequence right before the current insertion point. Then, to
24157 instrument 'base + len', just get an iterator on that statement.
24158 And do not forget to update the pointer to iterator the function
24159 received as argument.
24160
47918951
VM		 241612013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24162
24163 PR rtl-optimization/56348
24164 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24165
3fb7c699
SB		 241662013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24167
24168 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24169 (clean_graph_dump_file): Pass base to start_graph_dump.
24170
576fe41a
RH		 241712013-02-14 Richard Henderson <rth@redhat.com>
24172
24173 PR target/55941
24174 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24175
bb3accfa
SB		 241762013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24177
24178 * collect2-aix.h: Define F_LOADONLY.
24179
03143140
RB		 241802013-02-14 Richard Biener <rguenther@suse.de>
24181
24182 PR lto/50494
24183 * varasm.c (output_constant_def_1): Get the decl representing
24184 the constant as argument.
24185 (output_constant_def): Wrap output_constant_def_1.
24186 (make_decl_rtl): Use output_constant_def_1 with the decl
24187 representing the constant.
24188 (build_constant_desc): Optionally re-use a decl already
24189 representing the constant.
24190 (tree_output_constant_def): Adjust.
24191
8fb06726
DS		 241922013-02-14 Dodji Seketeli <dodji@redhat.com>
24193
24194 Fix an asan crash
24195 * asan.c (instrument_builtin_call): Really put the length of the
24196 second source argument into src1_len.
24197
c4bfe8bf
JJ		 241982013-02-13 Jakub Jelinek <jakub@redhat.com>
24199
24200 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24201 argument. If it is false, don't create edge from then_bb to
24202 fallthru_bb.
24203 (insert_if_then_before_iter): Pass true to it.
24204 (build_check_stmt): Pass false to it.
24205 (transform_statements): Flush hash table only on extended basic
24206 block boundaries, rather than at the beginning of every bb.
24207 Don't flush hash table on nonfreeing_call_p calls.
24208 * tree-flow.h (nonfreeing_call_p): New prototype.
24209 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24210
7afe2801
DM		 242112013-02-13 David S. Miller <davem@davemloft.net>
24212
24213 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24214
70cc3288
VM		 242152013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24216
24217 PR target/56184
24218 * ira.c (max_regno_before_ira): Move from ...
24219 (ira): ... here.
24220 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24221 vec_safe_length.
24222
6422242b
JJ		 242232013-02-13 Jakub Jelinek <jakub@redhat.com>
24224
24225 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24226
976a81ee
RB		 242272013-02-13 Richard Biener <rguenther@suse.de>
24228
24229 PR lto/56295
24230 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24231 globals in MEM_REFs.
24232
c1874a87
RB		 242332013-02-13 Richard Biener <rguenther@suse.de>
24234
24235 * loop-init.c (loop_optimizer_init): Clear loop state when
24236 re-initializing preserved loops.
24237 * loop-unswitch.c (unswitch_single_loop): Return whether
24238 we unswitched the loop. Do not verify loop state here.
88b97037 24239 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24240
b4ab7d34
KS		 242412013-02-13 Kostya Serebryany <kcc@google.com>
24242
88b97037
UB		 24243 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24244 on x86_64 linux.
b4ab7d34
KS		 24245 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24246
bdcbe80c
DS		 242472013-02-12 Dodji Seketeli <dodji@redhat.com>
24248
24249 Avoid instrumenting duplicated memory access in the same basic block
24250 * Makefile.in (asan.o): Add new dependency on hash-table.h
24251 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24252 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24253 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24254 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24255 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24256 (get_mem_ref_of_assignment): New functions.
24257 (get_mem_refs_of_builtin_call): Extract from
24258 instrument_builtin_call and tweak a little bit to make it fit with
24259 the new signature.
24260 (instrument_builtin_call): Use the new
24261 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24262 of is_gimple_builtin_call.
24263 (instrument_derefs, instrument_mem_region_access): Insert the
24264 instrumented memory reference into the hash table.
24265 (maybe_instrument_assignment): Renamed instrument_assignment into
24266 this, and change it to advance the iterator when instrumentation
24267 actually happened and return true in that case. This makes it
24268 homogeneous with maybe_instrument_assignment, and thus give a
24269 chance to callers to be more 'regular'.
24270 (transform_statements): Clear the memory reference hash table
24271 whenever we enter a new BB, when we cross a function call, or when
24272 we are done transforming statements. Use
24273 maybe_instrument_assignment instead of instrumentation. No more
24274 need to special case maybe_instrument_assignment and advance the
24275 iterator after calling it; it's now handled just like
24276 maybe_instrument_call. Update comment.
24277
4861a1f7
RB		 242782013-02-13 Richard Biener <rguenther@suse.de>
24279
24280 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24281 Fix loop discovery code.
24282
1c86bd80
VM		 242832013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24284
24285 PR inline-asm/56148
24286 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24287 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 24288 the operand is not reloaded. Prefer to reload conflicting operand
24289 if earlyclobber and matching operands are the same.
24290
7cbda518
RB		 242912013-02-12 Richard Biener <rguenther@suse.de>
24292
24293 PR lto/56297
24294 * lto-streamer-out.c (write_symbol): Do not output symbols
24295 for hard register variables.
24296
e68a4ef6
GJL		 242972013-02-12 Georg-Johann Lay <avr@gjlay.de>
24298
24299 PR target/54222
24300 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24301 (umulsidi3_insn, mulsidi3_insn): New insns.
24302
a72d8780 243032013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 24304
24305 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24306 (struct tune_params): Add vec_costs field.
24307 * config/arm/arm.c (arm_builtin_vectorization_cost)
24308 (arm_add_stmt_cost): New functions.
24309 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24310 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24311 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24312 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24313 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24314 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24315 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24316
43320568
RB		 243172013-02-12 Richard Biener <rguenther@suse.de>
24318
24319 PR lto/56295
24320 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24321 decls again if possible.
24322
a011aa39
RB		 243232013-02-12 Richard Biener <rguenther@suse.de>
24324
24325 PR middle-end/56288
24326 * tree-ssa.c (verify_ssa_name): Fix check, move
24327 SSA_NAME_IN_FREE_LIST check up.
24328
6da26889
JJ		 243292013-02-12 Jakub Jelinek <jakub@redhat.com>
24330 Steven Bosscher <steven@gcc.gnu.org>
24331
24332 PR rtl-optimization/56151
24333 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24334 equal to op0 or op1, and last_insn pattern is CODE operation
24335 with MEM dest and one of the operands matches that MEM.
24336
f80e0faf
ST		 243372013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24338
24339 * doc/extend.texi: Document Function Multiversioning and "default"
24340 parameter string to target attribute.
24341 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24342 target attribute parameter is "default".
24343 (ix86_compare_version_priority): Remove checks for target attribute.
24344 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24345 Remove check for target attribute equal to NULL. Add assert.
24346 (ix86_generate_version_dispatcher_body): Change error to sorry.
24347
6c59ffd1
IS		 243482013-02-11 Iain Sandoe <iain@codesourcery.com>
24349 Jack Howarth <howarth@bromo.med.uc.edu>
24350 Patrick Marlier <patrick.marlier@gmail.com>
24351
24352 PR libitm/55693
24353 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24354 define ENDFILE_SPEC as TM_DESTRUCTOR.
24355 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24356
a72d8780 243572013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 24358 Jack Howarth <howarth@bromo.med.uc.edu>
24359 Jakub Jelinek <jakub@redhat.com>
24360
24361 PR sanitizer/55617
24362 * config/darwin.c (cdtor_record): Rename ctor_record.
24363 (sort_cdtor_records): Rename sort_ctor_records.
24364 (finalize_dtors): New routine to sort destructors by
24365 priority before use in assemble_integer.
24366 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24367
b63fe007
UB		 243682013-02-11 Uros Bizjak <ubizjak@gmail.com>
24369
24370 PR rtl-optimization/56275
24371 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24372 offset is non-negative and less than cmode size before
24373 calling simplify_subreg.
24374
8e89b5b5
RB		 243752013-02-11 Richard Biener <rguenther@suse.de>
24376
24377 PR tree-optimization/56264
24378 * cfgloop.h (fix_loop_structure): Adjust prototype.
24379 * loop-init.c (fix_loop_structure): Return the number of
24380 newly discovered loops.
24381 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24382 are discovered, do a full loop-closed SSA rewrite.
24383
b4a4b56d
RB		 243842013-02-11 Richard Biener <rguenther@suse.de>
24385
24386 PR tree-optimization/56273
24387 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24388 first VRP run.
24389 (check_array_ref): Fix missing newline in dumps.
24390 (search_for_addr_array): Likewise.
24391
0c885229
DE		 243922013-02-09 David Edelsohn <dje.gcc@gmail.com>
24393
24394 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24395
59ac9a55
JJ		 243962013-02-09 Jakub Jelinek <jakub@redhat.com>
24397
24398 PR target/56256
24399 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24400
25bb0bb5
VM		 244012013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24402
24403 PR rtl-optimization/56246
0c885229 24404 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 24405 reload pseudo.
24406 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24407 constraints are satisfied.
24408
a698cc03
JL		 244092013-02-08 Jeff Law <law@redhat.com>
24410
24411 PR debug/53948
24412 * emit-rtl.c (reg_is_parm_p): New function.
24413 * regs.h (reg_is_parm_p): New prototype.
24414 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24415 callee-clobbered registers.
24416
e1122ddd
MM		 244172013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24418
24419 PR target/56043
24420 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24421 If there is no implicit builtin declaration, just return NULL.
24422
19c5f6e6
UB		 244232013-02-08 Uros Bizjak <ubizjak@gmail.com>
24424
24425 * config/i386/sse.md (FMAMODEM): New mode iterator.
24426 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24427 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24428
2480f2ca 244292013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 24430
2480f2ca
UB		 24431 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24432 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24433 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24434
244352013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24436
24437 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24438 (microblaze*-*-elf): Likewise.
24439 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24440 LINK_SPEC.
24441 * config/microblaze/microblaze-c.c: Add builtin defines for
24442 _LITTLE_ENDIAN and _BIG_ENDIAN.
24443 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24444 add to TARGET_DEFAULT flags.
76ef61fb 24445 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 24446 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24447 * config/microblaze/microblaze.md: Update extendsidi2 and
24448 movdi_internal instructions to use low-order / high-order reg
24449 print_operands.
24450 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24451 options and inversemask / mask of LITTLE_ENDIAN.
24452 * config/microblaze/t-microblaze: Expand multilib options to
24453 include mlittle-endian (le) and update exceptions patterns.
24454
600a5961
JJ		 244552013-02-08 Jakub Jelinek <jakub@redhat.com>
24456
5df81313
JJ		 24457 PR rtl-optimization/56195
24458 * lra-constraints.c (get_reload_reg): Don't reuse regs
24459 if they have smaller mode than requested, if they have
24460 wider mode than requested, try to return a SUBREG.
24461
600a5961
JJ		 24462 PR tree-optimization/56250
24463 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24464 if type is unsigned and code isn't MULT_EXPR.
24465
ff544649
GJL		 244662013-02-08 Georg-Johann Lay <avr@gjlay.de>
24467
24468 PR tree-optimization/56064
24469 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24470 bits according to mode.
24471 * fixed-value.h (fixed_from_double_int)
24472 (const_fixed_from_double_int): Adjust comments.
24473
e45cde98
RB		 244742013-02-08 Richard Biener <rguenther@suse.de>
24475
24476 PR lto/56231
24477 * lto-streamer.h (struct data_in): Remove current_file, current_line
24478 and current_col members.
24479 * lto-streamer-out.c (lto_output_location): Stream changed bits
24480 en-block for efficiency.
24481 * lto-streamer-in.c (clear_line_info): Remove.
24482 (lto_input_location): Cache current file, line and column
24483 globally via local statics. Read changed bits en-block.
24484 (input_function): Do not call clear_line_info.
24485 (lto_read_body): Likewise.
24486 (lto_input_toplevel_asms): Likewise.
24487
c1ca73d8
MM		 244882013-02-08 Michael Matz <matz@suse.de>
24489
24490 PR tree-optimization/52448
24491 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24492 (nt_call_phase): New static.
24493 (add_or_mark_expr): Only mark accesses with newer phase than any
24494 call seen.
24495 (nonfreeing_call_p): New.
24496 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24497 (nt_fini_block): Keep blocks marked as visited.
24498 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24499
57548aa2
RB		 245002013-02-08 Richard Biener <rguenther@suse.de>
24501
24502 * ira.c (ira): Free broken dominator information.
24503
8e10366f
UB		 245042013-02-08 Uros Bizjak <ubizjak@gmail.com>
24505
24506 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24507
f64fb0fa
MP		 245082013-02-08 Marek Polacek <polacek@redhat.com>
24509
8e10366f 24510 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 24511
0375167b
RB		 245122013-02-08 Richard Biener <rguenther@suse.de>
24513
24514 PR middle-end/56181
24515 * cfgloop.h (flow_loops_find): Adjust.
24516 (bb_loop_header_p): Declare.
24517 * cfgloop.c (bb_loop_header_p): New function split out from ...
24518 (flow_loops_find): ... here. Adjust function signature,
24519 support incremental loop structure update.
24520 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24521 * cfgloopmanip.c (fix_loop_structure): Move ...
24522 * loop-init.c (fix_loop_structure): ... here.
24523 (apply_loop_flags): Split out from ...
24524 (loop_optimizer_init): ... here.
24525 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24526 in incremental mode, only remove dead loops here.
24527
85d768f3
GJL		 245282013-02-08 Georg-Johann Lay <avr@gjlay.de>
24529
24530 PR target/54222
24531 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24532 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24533 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24534 (*round<mode>3.libgcc): New insns for fixed-modes.
24535 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24536 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24537 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24538 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24539 implementations. Define to __builtin_avr_absFX,
24540 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24541 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24542 __builtin_avr_countlsFX, respectively.
24543 * config/avr/avr-c.c (target.h): Include it.
24544 (enum avr_builtin_id): New enum.
24545 (avr_resolve_overloaded_builtin): New static function.
24546 (avr_register_target_pragmas): Use it to set
24547 targetm.resolve_overloaded_builtin.
24548 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24549 tree nodes used by DEF_BUILTIN.
24550 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24551 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24552 <AVR_BUILTIN_xxBITS>: Same.
24553
661bc682
RB		 245542013-02-08 Richard Biener <rguenther@suse.de>
24555
24556 * cfgloop.c (verify_loop_structure): Properly handle
24557 a loop exiting to another loop header.
24558 * ira-int.h (ira_loops): Remove.
24559 * ira.c (ira_loops): Remove.
24560 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24561 (do_reload): Use loop_optimizer_finalize.
24562 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24563 number_of_loops to access the loop tree.
24564 (more_one_region_p): Likewise.
24565 (finish_loop_tree_nodes): Likewise.
24566 (rebuild_regno_allocno_maps): Likewise.
24567 (mark_loops_for_removal): Likewise.
24568 (mark_all_loops_for_removal): Likewise.
24569 (remove_unnecessary_regions): Likewise.
24570 (ira_build): Likewise.
24571 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24572
0d5049b2
RB		 245732013-02-08 Richard Biener <rguenther@suse.de>
24574
24575 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24576 * ipa-pure-const.c (analyze_function): Avoid calling
24577 mark_irreducible_loops twice.
8e10366f 24578 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 24579
499f32e8
DM		 245802013-02-07 David S. Miller <davem@davemloft.net>
24581
24582 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24583 on 'reg'.
24584 * var-tracking.c (vt_add_function_parameter): Test the presence of
24585 HAVE_window_save properly and do not remap argument registers when
24586 we have a leaf function.
24587
6edc3e32
UB		 245882013-02-07 Uros Bizjak <ubizjak@gmail.com>
24589
24590 PR bootstrap/56227
24591 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24592 instead of "ll".
24593 * config/i386/i386.c (ix86_print_operand): Ditto.
24594
5306401f
VM		 245952013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24596
6edc3e32 24597 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 24598
027ece11
VM		 245992013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24600
24601 PR rtl-optimization/56225
24602 * lra-constraints.c (process_alt_operands): Check that reload hard
24603 reg can hold value for strict_low_part.
24604
f980dfdb
JJ		 246052013-02-07 Jakub Jelinek <jakub@redhat.com>
24606
24607 PR debug/56154
24608 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24609 dwarf2out_end_function.
24610 (in_first_function_p, maybe_at_text_label_p,
24611 first_loclabel_num_not_at_text_label): New variables.
24612 (dwarf2out_var_location): In the first function find out
6edc3e32 24613 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 24614 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24615 functions.
24616
73dd3123
EB		 246172013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24618
24619 PR rtl-optimization/56178
24620 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24621 SUBREG of a register. Tidy up related block of code.
24622 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24623 note if the source is a register or a SUBREG of a register.
24624
e3936f47
JJ		 246252013-02-07 Jakub Jelinek <jakub@redhat.com>
24626
24627 PR target/56228
24628 * config/rs6000/rs6000.md (ptrm): New mode attr.
24629 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24630 call_value_indirect_aix<pttrsize>,
24631 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24632 m in constraints.
24633
d96d674b
MH		 246342013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24635
24636 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24637 if -bnortl. Convert to strcmp and strncmp.
24638
921f2dee
AM		 246392013-02-07 Alan Modra <amodra@gmail.com>
24640
24641 PR target/54009
24642 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24643 addresses won't wrap when offsetting.
24644 (rs6000_secondary_reload): Provide secondary reloads needed for
24645 wrapping LO_SUM addresses.
24646
d09c7dba
TS		 246472013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24648
24649 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24650 MACH, just __MACH__.
24651
a44bbd48
RB		 246522013-02-06 Richard Biener <rguenther@suse.de>
24653
24654 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24655 instead of calling fix_loop_structure.
24656
6d840d99
JJ		 246572013-02-06 Jakub Jelinek <jakub@redhat.com>
24658
24659 PR middle-end/56217
24660 * omp-low.c (use_pointer_for_field): Return false if
24661 lower_send_shared_vars doesn't generate any copy-out code.
24662
0f33baa9
TV		 246632013-02-06 Tom de Vries <tom@codesourcery.com>
24664
24665 PR rtl-optimization/56131
24666 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24667 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24668 of the label is NULL. Add comment.
24669
11f1e3ab
JJ		 246702013-02-05 Jakub Jelinek <jakub@redhat.com>
24671
31502f9f
JJ		 24672 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24673
a74db9bd
JJ		 24674 PR sanitizer/55374
24675 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24676 (STATIC_LIBTSAN_LIBS): Likewise.
24677 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24678 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24679 is defined, don't add anything else beyond that.
24680 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24681 (LINK_COMMAND_SPEC): Use them.
24682
11f1e3ab
JJ		 24683 PR tree-optimization/56205
24684 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24685 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24686 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24687
14ac6aa2
RB		 246882013-02-05 Richard Biener <rguenther@suse.de>
24689
24690 PR tree-optimization/53342
24691 PR tree-optimization/53185
24692 * tree-vectorizer.h (vect_check_strided_load): Remove.
24693 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24694 not disallow peeling for vectorized strided loads.
24695 (vect_check_strided_load): Make static and simplify.
24696 (vect_analyze_data_refs): Adjust.
24697 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24698 correctly when vectorizing strided loads.
24699
6f22445a
RB		 247002013-02-05 Richard Biener <rguenther@suse.de>
24701
24702 * doc/install.texi: Refer to ISL, not PPL.
24703
39f9719e
JH		 247042013-02-05 Jan Hubicka <jh@suse.cz>
24705
ec4224ac
JH		 24706 PR tree-optimization/55789
24707 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24708
247092013-02-05 Jan Hubicka <jh@suse.cz>
24710
24711 PR tree-optimization/55789
39f9719e
JH		 24712 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24713 the dead call anyway.
24714
956d3b33
EB		 247152013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24716
24717 PR sanitizer/55374
24718 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24719
247202013-02-04 Alexander Potapenko <glider@google.com>
24721 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 24722 Jakub Jelinek <jakub@redhat.com>
24723
24724 PR sanitizer/55617
24725 * config/darwin.c (sort_ctor_records): Stabilized qsort
24726 on constructor priority by using original position.
24727 (finalize_ctors): New routine to sort constructors by
24728 priority before use in assemble_integer.
24729 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24730
7ac3af38
JJ		 247312013-02-04 Jakub Jelinek <jakub@redhat.com>
24732
24733 PR libstdc++/54314
24734 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24735 about visibility on artificial decls.
24736 * config/sol2.c (solaris_assemble_visibility): Likewise.
24737
152689dc
KT		 247382013-02-04 Kai Tietz <ktietz@redhat.com>
24739
24740 PR target/56186
24741 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24742 argument and improve checking of return-argument types for 16-byte
24743 modes.
24744 (ix86_function_value_1): Add additional valtype argument on call
24745 of function_value_64.
24746 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24747 handling infunction_value_64 function.
24748
2b5987b5
MGD		 247492013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24750
7ac3af38 24751 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 24752
14597080
RB		 247532013-02-04 Richard Biener <rguenther@suse.de>
24754
24755 PR tree-optimization/56188
24756 * tree-ssa-structalias.c (label_visit): Consider case with
24757 initially non-empty points-to set.
24758 (perform_var_substitution): Dump node mapping and clean up.
24759
ed73881e
RG		 247602013-02-04 Richard Guenther <rguenther@suse.de>
24761
24762 PR lto/56168
24763 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24764 node prevail as last resort.
24765 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 24766 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 24767
139a0707
RB		 247682013-02-04 Richard Biener <rguenther@suse.de>
24769
24770 PR tree-optimization/56113
24771 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24772 Merge into ...
24773 (equiv_class_lookup_or_add): ... this.
24774 (label_visit): Adjust and fix error in previous patch.
24775 (perform_var_substitution): Adjust.
24776
50fe8924
OE		 247772013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24778
24779 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24780 * config/sh/sh4-300.md: Likewise.
24781 * config/sh/sh4a.md: Likewise.
24782 * config/sh/constraints.md: Likewise.
24783 * config/sh/sh.md: Likewise.
24784 * config/sh/netbsd-elf.h: Likewise.
24785 * config/sh/predicates.md: Likewise.
24786 * config/sh/sh-protos.h: Likewise.
24787 * config/sh/ushmedia.h: Likewise.
24788 * config/sh/linux.h: Likewise.
24789 * config/sh/sh.c: Likewise.
24790 * config/sh/superh.h: Likewise.
24791 * config/sh/elf.h: Likewise.
24792 * config/sh/sh4.md: Likewise.
24793 * config/sh/sh.h: Likewise.
24794
1a04ac2b
JDA		 247952013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24796
24797 * config/pa/constraints.md: Adjust unused letters. Change "T"
24798 constraint to match_test floating_point_store_memory_operand().
24799 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24800 (base14_operand): New.
24801 (floating_point_store_memory_operand): New.
24802 (integer_store_memory_operand): Revise to use base14_operand and
24803 reg_plus_base_memory_operand.
24804 (move_dest_operand): Allow symbolic_memory_operands.
24805 (symbolic_memory_operand): Check for LO_SOM.
24806 (symbolic_operand): Change default case to break.
24807 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24808 CONST_DOUBLE values to be reloaded by putting them into memory when
24809 the destination is a floating point register.
24810 (movdf): Remove code to handle CONST_DOUBLE.
24811 (movsf): Likewise.
24812 (reload_indf_r1): New.
24813 (reload_insf_r1): New.
24814 Consistently use "Q" and "T" constraints with integer and floating
24815 point move instructions, respectively.
24816 (movdi): Remove FAIL.
24817 Change predicate for source operand unamed DImode move from
24818 general_operand to move_src_operand.
24819 (umulsidi3): Change predicate for destination operand to
24820 register_operand.
24821 Likewise for similar unamed patterns.
24822 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24823 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24824 (hppa_legitimize_address): Simplify mask calculation.
24825 (pa_emit_move_sequence): Revised handling of secondary reloads from
24826 REG+D addresses for floating point loads and stores. Directly handle
24827 loading CONST0_RTX (mode) to a floating point register.
24828 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24829 to floating point registers. Don't restrict secondary reloads to
24830 floating point registers to integer modes. Revise some comments and
24831 cleanup some code.
24832 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24833 (pa_legitimate_address_p): New.
24834 (pa_legitimize_reload_address): New.
24835 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24836 (STRICT_REG_OK_FOR_BASE_P): New.
24837 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24838 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24839
3bf65591
DE		 248402013-02-03 David Edelsohn <dje.gcc@gmail.com>
24841 Andrew Dixie <andrewd@gentrack.com>
24842
24843 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24844 flag set.
24845
c0a8a3e6
RS		 248462013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24847
24848 * expmed.c (extract_bit_field_1): Pass the full width of the
24849 structure to get_best_reg_extraction_insn.
24850
99113dff
DE		 248512013-02-01 David Edelsohn <dje.gcc@gmail.com>
24852
24853 PR target/54601
24854 * configure.ac (use_cxa_atexit): Add AIX.
24855 * configure: Regenerate.
24856
24857 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24858
ddd84654
JJ		 248592013-02-01 Jakub Jelinek <jakub@redhat.com>
24860
24861 PR debug/54793
24862 * final.c (need_profile_function): New variable.
24863 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24864 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24865 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24866 notes, targetm.asm_out.function_prologue doesn't emit anything,
24867 HAVE_prologue and profiler should be emitted before prologue,
24868 set need_profile_function instead of emitting it.
24869 (final_scan_insn): If need_profile_function, emit
24870 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24871 NOTE_INSN_FUNCTION_BEG note.
24872
248732013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 24874
24875 * config/rs6000/rs6000.md (smulditi3): New.
24876 (umulditi3): New.
24877
ff2a9d88
RH		 24878 * config/alpha/alpha.md (umulditi3): New.
24879
14d52b90
DE		 248802013-02-01 David Edelsohn <dje.gcc@gmail.com>
24881
24882 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24883 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24884
8c7ca45c
RB		 248852013-02-01 Richard Biener <rguenther@suse.de>
24886
24887 PR tree-optimization/56113
24888 * tree-ssa-structalias.c (label_visit): Reduce work for
24889 single-predecessor nodes.
24890
9f419393
EB		 248912013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24892
24893 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24894 range isn't testing for zero.
24895
c5f4be84
SB		 248962013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24897
24898 PR middle-end/56113
24899 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24900
dbdbd982
NC		 249012013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24902 Nick Clifton <nickc@redhat.com>
24903
24904 * config/v850/constraints.md (Q): Define as a memory constraint.
24905 * config/v850/predicates.md (label_ref_operand): New predicate.
24906 (e3v5_shift_operand): New predicate.
24907 (ior_operator): New predicate.
24908 * config/v850/t-v850: Add e3v5 multilib.
24909 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24910 (v850_gen_movdi): Prototype.
24911 * config/v850/v850.c: Add support for e3v5 architecture.
24912 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24913 TARGET_V850E_UP.
24914 (construct_save_jarl): Add e3v5 long JARL support.
24915 (v850_adjust_insn_length): New function. Adjust length of call
24916 insns when using e3v5 instructions.
24917 (v850_gen_movdi): New function: Generate instructions to move a
24918 DImode value.
24919 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24920 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24921 (TARGET_USE_FPU): Enable for e3v5.
24922 (CONST_OK_FOR_W): New macro.
24923 (ADJUST_INSN_LENGTH): Define.
24924 * config/v850/v850.md (UNSPEC_LOOP): Define.
24925 (attr cpu): Add v850e3v5.
24926 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24927 (movdi): New pattern.
24928 (movdi_internal): New pattern.
24929 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
24930 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
24931 (cstoresf4): Likewise.
24932 (cstoredf4): Likewise.
24933 (insv): New pattern.
24934 (rotlso3_a): New pattern.
24935 (rotlsi3_b): New pattern
24936 (rotlsi3_v850e3v5): New pattern.
24937 (doloop_begin): New pattern.
24938 (fix_loop_counter): New pattern.
24939 (doloop_end): New pattern.
24940 (branch_normal): Add e3v5 long branch support.
24941 (branch_invert): Likewise.
24942 (branch_z_normal): Likewise.
24943 (branch_z_invert): Likewise.
24944 (branch_nz_normal): Likewise.
24945 (branch_nz_invert): Likewise.
24946 (call_internal_short): Add e3v5 register-indirect JARL support.
24947 (call_internal_long): Likewise.
24948 (call_value_internal_short): Likewise.
24949 (call_value_internal_long): Likewise.
24950 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
24951 (mloop): New option.
24952 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 24953 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 24954
73861a41
PK		 249552013-01-31 Paul Koning <ni1d@arrl.net>
24956
24957 PR debug/55059
24958 PR debug/54508
24959 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
24960 children if parent is a class.
24961 (prune_unused_types_prune): Don't add DW_AT_declaration.
24962
e44978dc
RB		 249632013-01-31 Richard Biener <rguenther@suse.de>
24964
24965 PR tree-optimization/56157
24966 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
24967 match up operand with SLP child.
24968
90b10dec
JM		 249692013-01-31 Jason Merrill <jason@redhat.com>
24970
04d2dadd 24971 PR debug/54410
90b10dec
JM		 24972 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
24973 parameters the first time.
24974 (gen_scheduled_generic_parms_dies): Check completeness here.
24975
6e616110
RB		 249762013-01-31 Richard Biener <rguenther@suse.de>
24977
24978 PR middle-end/53073
24979 * common.opt (faggressive-loop-optimizations): New flag,
24980 enabled by default.
24981 * doc/invoke.texi (faggressive-loop-optimizations): Document.
24982 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
24983 infer_loop_bounds_from_undefined by it.
24984
636f59cf
RB		 249852013-01-31 Richard Biener <rguenther@suse.de>
24986
24987 PR tree-optimization/56150
24988 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
24989 visit virtual operands.
24990 (find_uses_to_rename_bb): Likewise.
24991
249922013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 24993
24994 PR tree-optimization/56150
24995 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
24996 mixed store non-store stmts.
24997
32887460
JJ		 249982013-01-30 Jakub Jelinek <jakub@redhat.com>
24999
e60e09a0
JJ		 25000 PR sanitizer/55374
25001 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25002 LIBASAN_EARLY_SPEC is defined.
25003 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25004 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25005 before %o.
25006 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25007
32887460
JJ		 25008 PR c++/55742
25009 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25010 invalid args instead of ICEing on it.
25011 (ix86_valid_target_attribute_tree): Return error_mark_node if
25012 ix86_valid_target_attribute_inner_p failed.
25013 (ix86_valid_target_attribute_p): Return false only if
25014 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25015 target("default") attribute.
25016 (sorted_attr_string): Change argument from const char * to tree,
25017 merge in all target attribute arguments rather than just one.
25018 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25019 instead of free. Avoid using strcat.
25020 (ix86_mangle_function_version_assembler_name): Mangle
25021 target("default") as if no target attribute is present. Adjust
25022 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25023 instead of xmalloc and XDELETEVEC instead of free.
25024 (ix86_function_versions): Don't return true if one of the decls
25025 doesn't have target attribute. If they don't and one of the decls
25026 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25027 sorted_attr_string caller. Use XDELETEVEC instead of free.
25028 (ix86_supports_function_versions): Remove.
25029 (make_name): Fix up formatting.
25030 (make_dispatcher_decl): Remove resolver_name and its initialization.
25031 Avoid leaking memory.
25032 (is_function_default_version): Return true if there is
25033 target("default") attribute rather than no target attribute at all.
25034 (make_resolver_func): Avoid leaking memory.
25035 (ix86_generate_version_dispatcher_body): Likewise.
25036 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25037 * target.def (supports_function_versions): Remove.
25038 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25039 * doc/tm.texi: Regenerated.
25040
73cca0cc
VM		 250412013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25042
25043 PR rtl-optimization/56144
25044 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25045 for values with side effects.
25046
11452e7b
RB		 250472013-01-30 Richard Biener <rguenther@suse.de>
25048
25049 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25050 (sparseset_pop): Likewise.
25051 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25052 to be able to use quick_push in the worker loop.
25053
01cb1ef5
MP		 250542013-01-30 Marek Polacek <polacek@redhat.com>
25055
25056 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25057
e892936e
RB		 250582013-01-30 Richard Biener <rguenther@suse.de>
25059
25060 PR lto/56147
6edc3e32 25061 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 25062
cc06c01d
GJL		 250632013-01-30 Georg-Johann Lay <avr@gjlay.de>
25064
25065 PR tree-optimization/56064
25066 * fixed-value.c (fixed_from_double_int): New function.
25067 * fixed-value.h (fixed_from_double_int): New prototype.
25068 (const_fixed_from_double_int): New static inline function.
25069 * fold-const.c (native_interpret_fixed): New static function.
25070 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25071 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25072 (native_encode_fixed): New static function.
25073 (native_encode_expr) <FIXED_CST>: Use it.
25074 (native_interpret_int): Move double_int worker code to...
25075 * double-int.c (double_int::from_buffer): ...this new static method.
25076 * double-int.h (double_int::from_buffer): Prototype it.
25077
d394a308
RB		 250782013-01-30 Richard Biener <rguenther@suse.de>
25079
25080 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25081 New pointer-map and obstack.
25082 (init_alias_vars): Allocate pointer-map and obstack.
25083 (delete_points_to_sets): Free them.
25084 (find_what_var_points_to): Cache result.
25085 (find_what_p_points_to): Adjust for changed interface of
25086 find_what_var_points_to.
25087 (compute_points_to_sets): Likewise.
25088 (ipa_pta_execute): Likewise.
25089
20804d96
RO		 250902013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25091
25092 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25093 * configure: Regenerate.
25094 * config.in: Regenerate.
25095 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25096 #nobits/#progbits if supported.
25097
a7ad88a2
OE		 250982013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25099
25100 PR target/56121
25101 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25102 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25103 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25104
7af79f92
GY		 251052013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25106
91bfca59
OE		 25107 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25108 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 25109
753bcf7b
GY		 251102013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25111
91bfca59
OE		 25112 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25113 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 25114
8cbc2ea8
GY		 251152013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25116
91bfca59
OE		 25117 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25118 declaration.
25119 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25120 * config/arm/cortex-a7.md: New bypasses using
25121 arm_mac_accumulator_is_result.
8cbc2ea8 25122
697a3325
GY		 251232013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25124
91bfca59 25125 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 25126 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 25127 (cortex_a7_fpfmad): New reservation.
25128 (cortex_a7_fpmacs): Use ffmas and update required units.
25129 (cortex_a7_fpmuld): Update required units and latency.
25130 (cortex_a7_fpmacd): Likewise.
25131 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25132 (cortex_a7_neon). Likewise.
25133 (bypass) Update participating units.
697a3325 25134
29637783
GY		 251352013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25136
91bfca59
OE		 25137 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25138 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25139 from fmac to ffma.
25140 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25141 (vfp_fmul): Use ffmad.
25142 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25143 (cortex_r4_fmacd): Use ffmad.
25144 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25145 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25146 (cortex_a9_fmacd): Use ffmad.
25147 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25148 (cortex_a8_vfp_macd): Use ffmad.
25149 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25150 (cortex_a5_fpmacd): Use ffmad.
25151 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25152 (cortex_a15_vfp_macd): Use ffmad.
25153 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 25154
2ee8a2d5
JM		 251552013-01-29 Jason Merrill <jason@redhat.com>
25156
25157 PR libstdc++/54314
25158 * varasm.c (default_assemble_visibility): Don't warn about
25159 visibility on artificial decls.
25160
78d087bc
RB		 251612013-01-29 Richard Biener <rguenther@suse.de>
25162
25163 PR tree-optimization/56113
25164 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25165 the bitmap leader.
25166 (label_visit): Free duplicate bitmaps and record the leader instead.
25167 (perform_var_substitution): Adjust.
25168
83ba4d6f
RB		 251692013-01-29 Richard Biener <rguenther@suse.de>
25170
25171 PR tree-optimization/55270
25172 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25173 the CFG, schedule loops for fixup.
25174
66dfe13f
NC		 251752013-01-29 Nick Clifton <nickc@redhat.com>
25176
25177 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25178 SP_REG.
25179
5a579c3b
LE		 251802013-01-28 Leif Ekblad <leif@rdos.net>
25181
25182 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25183 * config/i386/i386.h (TARGET_RDOS): New macro.
25184 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25185 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25186 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25187 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25188 DEFAULT_LARGE_SECTION_THRESHOLD.
25189 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25190 * config/i386/rdos.h: New file.
25191 * config/i386/rdos64.h: New file.
25192
51e44392
BS		 251932013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25194
25195 PR other/54814
25196 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25197 TEST_HARD_REG_BIT.
25198
db1fb332
JJ		 251992013-01-28 Jakub Jelinek <jakub@redhat.com>
25200
25201 PR rtl-optimization/56117
25202 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25203 call cselib_lookup_from_insn on the MEM before calling
25204 add_insn_mem_dependence.
25205
16917761
RB		 252062013-01-28 Richard Biener <rguenther@suse.de>
25207
25208 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25209 to a stmt that didn't have one.
25210 (copy_phis_for_bb): Likewise for PHI arguments.
25211 (copy_debug_stmt): Likewise for debug stmts.
25212
b9fc0497
RB		 252132013-01-28 Richard Biener <rguenther@suse.de>
25214
25215 PR tree-optimization/56034
6edc3e32 25216 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 25217 (partition_builtin_p): Adjust.
25218 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25219 it is the last partition.
25220 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25221 up the vertex for the definition.
25222 (classify_partition): Classify whether a partition is a
25223 PKIND_REDUCTION, thus has uses outside of the loop.
25224 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25225 Merge all PKIND_REDUCTION partitions into the last partition.
25226 (tree_loop_distribution): Seed partitions from reductions as well.
25227
aa710d25
JJ		 252282013-01-28 Jakub Jelinek <jakub@redhat.com>
25229
0bfbca58
JJ		 25230 PR tree-optimization/56125
25231 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25232 pow(x,c) into sqrt(x) * powi(x, n/2) or
25233 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25234 optimizing for size.
25235 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25236 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25237 integer.
25238
aa710d25
JJ		 25239 PR tree-optimization/56094
25240 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25241 to UNKNOWN_LOCATION while gimplifying expr.
25242
77dc5297
UB		 252432013-01-27 Uros Bizjak <ubizjak@gmail.com>
25244
25245 PR target/56114
25246 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25247 operand 0 in movabs insn template for -masm=intel asm alternative.
25248 (*movabs<mode>_2): Ditto for operand 1.
25249
0bfbca58 252502013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 25251
25252 PR target/54663
25253 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25254 of microblaze-c.o
25255
0bfbca58 252562013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 25257
25258 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25259 tm_file.
25260
0bfbca58 252612013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 25262
25263 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25264 Undef to avoid warning.
25265
478f60f9
MH		 252662013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25267
25268 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25269 * configure: Regenerate.
25270
d7fa6ee2
JJ		 252712013-01-25 Jakub Jelinek <jakub@redhat.com>
25272
25273 PR tree-optimization/56098
25274 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25275 for stmts with volatile ops.
25276 (cond_store_replacement): Don't optimize if assign has volatile ops.
25277 (cond_if_else_store_replacement_1): Don't optimize if either
25278 then_assign or else_assign have volatile ops.
25279 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25280 volatile ops.
25281
f8fe87bd
GJL		 252822013-01-25 Georg-Johann Lay <avr@gjlay.de>
25283
25284 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25285
28398d0d
GJL		 252862013-01-25 Georg-Johann Lay <avr@gjlay.de>
25287
25288 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25289 missing ':' in asm example.
25290
b7d7d917
TB		 252912013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25292
25293 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25294 entries into lane and laneq entries.
77dc5297
UB		 25295 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25296 Remove AdvSIMD scalar modes.
b7d7d917
TB		 25297 (aarch64_sq<r>dmulh_laneq<mode>): New.
25298 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25299 modes.
25300 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25301 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25302 * config/aarch64/iterators.md (VCOND): New.
25303 (VCONQ): New.
25304
556f9906
GJL		 253052013-01-25 Georg-Johann Lay <avr@gjlay.de>
25306
25307 PR target/54222
25308 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25309 Add NULL LIBNAME argument to existing definitions.
25310 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25311 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25312 * config/avr/avr.c (DEF_BUILTIN): Same.
25313 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25314 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25315 implementation is available (DECL_ASSEMBLER_NAME is set).
25316 (avr_fold_absfx): New static function.
25317 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25318 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25319 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25320 AVR_BUILTIN_ABSLLK.
25321 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25322 (abshk, absk, abslk, absllk): Provide as static inline functions.
25323
1f546bbb
MP		 253242013-01-25 Marek Polacek <polacek@redhat.com>
25325
25326 PR tree-optimization/56035
25327 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25328
7e184bd7
UB		 253292012-01-24 Uros Bizjak <ubizjak@gmail.com>
25330
25331 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25332 (*movtf_internal_rex64): Add (!o,C) alternative
25333 (*movxf_internal_rex64): Ditto.
25334 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25335
f5ba49ac
SM		 253362013-01-24 Shenghou Ma <minux.ma@gmail.com>
25337
25338 * doc/invoke.texi: fix typo.
25339 * doc/objc.texi: fix typo.
25340
a6343728
RS		 253412013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25342
25343 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25344 for the first two alternatives.
25345
cd030c07
DN		 253462013-01-24 Diego Novillo <dnovillo@google.com>
25347
77dc5297 25348 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 25349 (ggc-zone.o): Remove.
25350 * configure.ac: Remove option --with-gc.
25351 * configure: Re-generate.
25352 * doc/install.texi: Remove documentation for --with-gc.
25353 * gengtype.c (write_enum_defn): Remove. Update all users.
25354 (write_Types_process_field): Remove generation of gt_e_* argument.
25355 (output_type_enum): Remove. Update all users.
25356 (write_enum_defn): Remove. Update all users.
25357 (enum alloc_zone): Remove. Update all users.
77dc5297 25358 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 25359 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25360 Update all callers.
25361 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25362 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 25363 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25364 Update all users.
aaf1e810 25365 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 25366 (struct alloc_zone): Remove.
25367 (ggc_internal_alloc_zone_stat): Remove.
25368 (ggc_internal_cleared_alloc_zone_stat): Remove.
25369 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 25370 (ggc_pch_count_object): Remove last argument. Update all users.
25371 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 25372 (struct alloc_zone): Remove.
25373 * ggc-zone.c: Remove.
77dc5297 25374 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 25375 (struct alloc_zone): Remove.
25376 (ggc_alloc_typed_stat): Remove.
25377 (ggc_alloc_typed): Remove.
25378 (ggc_splay_alloc): Remove first argument.
25379 (rtl_zone): Remove. Update all users.
25380 (tree_zone): Remove. Update all users.
25381 (tree_id_zone): Remove. Update all users.
25382 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25383 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25384 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25385 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25386
a861ffa4
GJL		 253872013-01-24 Georg-Johann Lay <avr@gjlay.de>
25388
25389 * config/avr/avr.c (avr_out_fract): Make register numbers that
25390 might be outside of source operand signed.
25391
593c0ddd
UB		 253922013-01-24 Uros Bizjak <ubizjak@gmail.com>
25393
25394 * config/i386/constraints.md (Yf): New constraint.
25395 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25396 of f constraint to conditionaly disable x87 register preferences.
25397 (*movdf_internal): Ditto.
25398 (*movsf_internal): Ditto.
25399
e86c0101
SB		 254002013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25401
25402 PR inline-asm/55934
25403 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25404 that have operands with impossible constraints.
25405 Add a FIXME for a speed-up opportunity.
25406 * lra-constraints.c (process_alt_operands): Verify that a class
25407 selected from constraints on asms is valid for the operand mode.
25408 (curr_insn_transform): Remove incorrect comment.
25409
f6fee35f
DE		 254102013-01-23 David Edelsohn <dje.gcc@gmail.com>
25411
25412 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25413 TOC operand is a valid symbol ref in the constant pool.
25414
254152013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 25416
aaf1e810 25417 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 25418
dc62d7d1
GJL		 254192013-01-23 Georg-Johann Lay <avr@gjlay.de>
25420
25421 PR target/54222
25422 * config/avr/stdfix.h: New file.
25423 * t-avr (stdfix-gcc.h): New rule to build it.
25424 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25425
2660d12d
KS		 254262013-01-23 Kostya Serebryany <kcc@google.com>
25427
77dc5297
UB		 25428 * config/darwin.h: remove dependency on
25429 CoreFoundation (asan on Mac OS).
2660d12d 25430
a70418fc
JJ		 254312013-01-23 Jakub Jelinek <jakub@redhat.com>
25432
25433 PR target/49069
25434 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25435 instead of cmpdi_operand for first comparison operand.
25436 Don't assert that comparison operands aren't both constants.
25437
47876a2a
JW		 254382013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25439
25440 * doc/install.texi (Downloading the Source): Update references to
25441 downloading separate components.
25442
254432013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 25444
25445 * doc/extend.texi (__int128): Improve grammar.
25446
47876a2a 254472013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 25448
25449 PR target/56028
25450 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25451 alternative to (o,r).
25452 (*movdi_internal_rex64): Remove (!o,n) alternative.
25453 (DImode immediate->memory splitter): Remove.
25454 (DImode immediate->memory peephole2): Remove.
25455 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25456 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25457 alternative to (!o,*r).
25458 (*movtf_internal_sse): New pattern.
25459 (*movxf_internal_rex64): New pattern.
25460 (*movxf_internal): Disable for TARGET_64BIT.
25461 (*movdf_internal_rex64): Remove (!o,F) alternative.
25462
3a984f10
JJ		 254632013-01-22 Jakub Jelinek <jakub@redhat.com>
25464
502498d5
JJ		 25465 PR middle-end/56074
25466 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25467 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25468 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25469 stmt locations where LOCATION_LOCUS of the stmt location is
25470 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25471
3a984f10
JJ		 25472 PR target/55686
25473 * config/i386/i386.md (UNSPEC_STOS): New.
25474 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25475 *strsetqi_1): Add UNSPEC_STOS.
25476
fa817f7f
PC		 254772013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25478
25479 PR c++/56067
25480 * doc/invoke.texi: Remove left over -Wsynth example.
25481
8f498c1b
JJ		 254822013-01-21 Jakub Jelinek <jakub@redhat.com>
25483
25484 PR tree-optimization/56051
25485 * fold-const.c (fold_binary_loc): Don't fold
25486 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25487 a narrowing conversion, or widening conversion from signed
25488 to unsigned.
25489
47876a2a 254902013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 25491
25492 PR rtl-optimization/56023
25493 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25494 dependent on debug instruction.
25495
5b9db1bc
MJ		 254962013-01-21 Martin Jambor <mjambor@suse.cz>
25497
25498 PR middle-end/56022
25499 * function.c (allocate_struct_function): Call
25500 invoke_set_current_function_hook earlier.
25501
e8bb7d68
JJ		 255022013-01-21 Jakub Jelinek <jakub@redhat.com>
25503
25504 * reload1.c (init_reload): Only initialize reload_obstack
25505 during the first call.
25506
616a4e32
MP		 255072013-01-21 Marek Polacek <polacek@redhat.com>
25508
25509 * cfgloop.c (verify_loop_structure): Fix up grammar.
25510
4401981b
YHH		 255112013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25512
25513 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25514 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25515
8e87740b
RR		 255162013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25517
25518 PR target/56058
25519 * config/arm/marvell-pj4.md: Update copyright year.
25520 Fix up use of alu to alu_reg and simple_alu_imm.
25521
47876a2a 255222013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 25523
25524 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25525
89d56d79
VM		 255262013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25527
25528 PR target/55433
25529 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 25530 insn for secondary memory move when memory mode should be different.
89d56d79 25531
fe603553
JDA		 255322013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25533
25534 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25535 atomic_storedi_1): New patterns.
25536
01284895
VK		 255372013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25538
25539 btver2 pipeline descriptions.
25540 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25541 descriptions.
25542 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 25543 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 25544 type attributes.
25545 * config/i386/btver2.md: New file describing btver2 pipelines.
25546
5630e3e1
JL		 255472013-01-19 Andrew Pinski <apinski@cavium.com>
25548
25549 PR tree-optimization/52631
25550 * tree-ssa-sccvn (visit_use): Before looking up the original
25551 statement, try looking up the simplified expression.
25552
650ae806
AG		 255532013-01-19 Anthony Green <green@moxielogic.com>
25554
25555 * config/moxie/moxie.c (moxie_expand_prologue): Set
25556 current_function_static_stack_size.
25557
e300ec2d
JJ		 255582013-01-18 Jakub Jelinek <jakub@redhat.com>
25559
25560 PR tree-optimization/56029
25561 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25562 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25563
a3d7af04
SS		 255642013-01-18 Sharad Singhai <singhai@google.com>
25565
25566 PR tree-optimization/55995
25567 * dumpfile.c (dump_loc): Print location only if available.
25568 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25569
66aa7879
VM		 255702013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25571
25572 PR target/55433
25573 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25574 secondary memory move.
25575 (inherit_reload_reg): Use rclass instead of cl for
25576 check_secondary_memory_needed_p.
25577
3f0fee7b
JJ		 255782013-01-18 Jakub Jelinek <jakub@redhat.com>
25579
25580 PR middle-end/56015
25581 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 25582 the case where writing real complex part of target modifies op1.
3f0fee7b 25583
70c67693
JG		 255842013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25585
25586 * config/aarch64/aarch64-simd.md
25587 (aarch64_vcond_internal<mode>): Handle unordered cases.
25588 * config/aarch64/iterators.md (v_cmp_result): New.
25589
df8de9b3
YHH		 255902013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25591 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25592
25593 * config/arm/marvell-pj4.md: New file.
25594 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25595 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25596 (generic_vfp): Likewise.
25597 * config/arm/arm-cores.def: Add marvell-pj4.
25598 * config/arm/arm-tune.md: Regenerate.
25599 * config/arm/arm-tables.opt: Regenerate.
25600 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25601 * doc/invoke.texi: Document marvell-pj4.
25602
be30c356
TB		 256032013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25604
25605 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25606
0bfbca58 256072013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 25608
25609 PR debug/54114
25610 PR debug/54402
25611 PR debug/49888
25612 * var-tracking.c (negative_power_of_two_p): New.
25613 (global_get_addr_cache, local_get_addr_cache): New.
25614 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25615 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25616 heading comment.
25617 (vt_stack_offset_p): Remove.
25618 (vt_canon_true_dep): Always canonicalize loc's address.
25619 (clobber_overlapping_mems): Make sure we have a MEM.
25620 (local_get_addr_clear_given_value): New.
25621 (val_reset): Clear local cached entries.
25622 (compute_bb_dataflow): Create and release the local cache.
25623 Disable duplicate MEMs clobbering.
25624 (emit_notes_in_bb): Clobber MEMs likewise.
25625 (vt_emit_notes): Create and release the local cache.
25626 (vt_initialize, vt_finalize): Create and release the global
25627 cache, respectively.
1f6bc337 25628 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 25629
0bfbca58 256302013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 25631
25632 PR libmudflap/53359
25633 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25634 not found in the symtab.
25635
0bfbca58 256362013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 25637
c350ba53 25638 PR debug/56006
3aa03517
AO		 25639 PR rtl-optimization/55547
25640 PR rtl-optimization/53827
25641 PR debug/53671
25642 PR debug/49888
25643 * alias.c (offset_overlap_p): New, factored out of...
25644 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25645 the conservative special case for symbolic constants. Don't
25646 adjust zero sizes on alignment.
25647
c664546f
JL		 256482013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25649
25650 PR rtl-optimization/52573
25651 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25652 REG_UNUSED for the same register.
25653
1bd3f750
MP		 256542013-01-17 Richard Biener <rguenther@suse.de>
25655 Marek Polacek <polacek@redhat.com>
25656
25657 PR rtl-optimization/55833
25658 * loop-unswitch.c (unswitch_loops): Move loop verification...
25659 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25660 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25661 Set it to true when we're removing a loop from hierarchy tree in
25662 an irreducible region.
25663 (fix_bb_placements): Adjust caller.
25664 (fix_loop_placements): Likewise.
25665
e52a8b71
GJL		 256662013-01-17 Georg-Johann Lay <avr@gjlay.de>
25667
25668 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25669 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25670 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25671 Remove ID. Adjust comments.
25672 * config/avr/avr-c.c (avr_builtin_name): Remove.
25673 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25674 * config/avr/avr.c (avr_tolower): New static function.
25675 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25676 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25677 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25678 default expansion.
25679
8386a7ea
JH		 256802013-01-17 Jan Hubicka <jh@suse.cz>
25681
610fb637 25682 PR tree-optimization/55273
8386a7ea
JH		 25683 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25684
47876a2a 256852013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 25686
25687 PR target/55981
25688 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25689 store through atomic_store<mode>_1.
25690 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25691
8222c37e
MJ		 256922013-01-17 Martin Jambor <mjambor@suse.cz>
25693
25694 PR tree-optimizations/55264
25695 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25696 for virtual methods.
25697 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25698 virtual methods before inlining is over.
25699 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25700 virtual functions.
25701 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25702 non-virtual.
25703
79f01c76
VM		 257042013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25705
25706 PR rtl-optimization/56005
25707 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25708 pending reads for prefetch.
25709
d0b6bb1b
IB		 257102013-01-16 Ian Bolton <ian.bolton@arm.com>
25711
aaf1e810 25712 * config/aarch64/aarch64.md
d0b6bb1b
IB		 25713 (*cstoresi_neg_uxtw): New pattern.
25714 (*cmovsi_insn_uxtw): New pattern.
25715 (*<optab>si3_uxtw): New pattern.
25716 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25717 (*<optab>si3_insn_uxtw): New pattern.
25718 (*bswapsi2_uxtw): New pattern.
25719
cb9cf03b
RB		 257202013-01-16 Richard Biener <rguenther@suse.de>
25721
25722 * tree-inline.c (tree_function_versioning): Remove set but
25723 never used variable.
25724
2cfc56b9
RB		 257252013-01-16 Richard Biener <rguenther@suse.de>
25726
25727 PR tree-optimization/55964
25728 * tree-flow.h (rename_variables_in_loop): Remove.
25729 (rename_variables_in_bb): Likewise.
25730 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25731 (copy_loop_before): Adjust and delete update-ssa status.
25732 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 25733 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 25734 (rename_variables_in_loop): Remove.
25735 (slpeel_update_phis_for_duplicate_loop): Likewise.
25736 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25737 use available cfg machinery instead of duplicating it.
25738 Update PHI nodes and perform poor-mans SSA update here.
25739 (slpeel_tree_peel_loop_to_edge): Adjust.
25740
c25a0c60
RB		 257412013-01-16 Richard Biener <rguenther@suse.de>
25742
25743 PR tree-optimization/54767
25744 PR tree-optimization/53465
25745 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25746 (vrp_visit_phi_node): For PHI arguments coming via backedges
25747 drop all symbolical range information.
25748 (execute_vrp): Compute backedges.
25749
04b535af
RB		 257502013-01-16 Richard Biener <rguenther@suse.de>
25751
25752 * doc/install.texi: Update CLooG and ISL requirements to
25753 0.18.0 and 0.11.1.
25754
8b0a1e0b
CB		 257552013-01-16 Christian Bruel <christian.bruel@st.com>
25756
25757 PR target/55301
25758 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25759 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25760 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25761
257622013-01-16 DJ Delorie <dj@redhat.com>
25763
25764 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25765 (UNSPECV_SP_SWITCH_E): New.
25766 (sp_switch_1): Change to an unspec.
25767 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25768 replace $r15.
25769
257702013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 25771
25772 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25773 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25774 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25775 (expand_mem_thread_fence): Ditto.
25776 (expand_mem_signal_fence): Ditto.
25777 (expand_atomic_load): Ditto.
25778 (expand_atomic_store): Ditto.
25779
0bfbca58 257802013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 25781
25782 PR rtl-optimization/55547
25783 PR rtl-optimization/53827
25784 PR debug/53671
25785 PR debug/49888
25786 * alias.c (memrefs_conflict_p): Set sizes to negative after
25787 AND adjustments.
25788
305e3ac1
JJ		 257892013-01-15 Jakub Jelinek <jakub@redhat.com>
25790
25791 PR target/55940
25792 * function.c (thread_prologue_and_epilogue_insns): Always
25793 add crtl->drap_reg to set_up_by_prologue.set, even if
25794 stack_realign_drap is false.
25795
f78ac4f2
JBG		 257962013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25797
25798 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25799 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25800 *call): Fix indention.
25801
a78a8cc4
TV		 258022013-01-15 Tom de Vries <tom@codesourcery.com>
25803
25804 PR target/55876
25805 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25806 Update comment.
25807
0e80383f
VM		 258082013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25809
305e3ac1 25810 PR rtl-optimization/55153
0e80383f
VM		 25811 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25812
207b5956
MJ		 258132013-01-15 Martin Jambor <mjambor@suse.cz>
25814
25815 PR tree-optimization/55920
25816 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25817 accesses as grp_to_be_debug_replaced.
25818
a7818b54
JJ		 258192013-01-15 Jakub Jelinek <jakub@redhat.com>
25820
25821 PR tree-optimization/55920
25822 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25823 there is non-useless type conversion needed from debug rhs to lhs,
25824 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25825
b0fe107e
JM		 258262013-01-15 Joseph Myers <joseph@codesourcery.com>
25827 Mikael Pettersson <mikpe@it.uu.se>
25828
25829 PR target/43961
25830 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25831 Thumb.
25832 (ASM_OUTPUT_CASE_LABEL): Remove.
25833 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25834 * final.c (shorten_branches): Update alignment of labels before
25835 jump tables if CASE_VECTOR_SHORTEN_MODE.
25836
34ab62ee
RB		 258372013-01-15 Richard Biener <rguenther@suse.de>
25838
25839 PR bootstrap/55961
25840 * system.h: Do not include gmp.h for building host tools.
25841
783a3a05
RB		 258422013-01-15 Richard Biener <rguenther@suse.de>
25843
25844 PR middle-end/55882
25845 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25846 account for bitpos when computing alignment.
25847
3a579e09
VY		 258482013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25849
25850 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25851 (ix86_target_macros_internal): Likewise.
25852
25853 * config/i386/i386.c (m_CORE2I7): Removed.
25854 (m_CORE_HASWELL): New macro.
25855 (m_CORE_ALL): Likewise.
25856 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25857 (initial_ix86_arch_features): Likewise.
25858 (processor_target_table): Initializations for Core avx2.
25859 (cpu_names): New names "core-avx2".
25860 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25861 PROCESSOR_CORE_HASWELL.
25862 (ix86_issue_rate): New case.
25863 (ia32_multipass_dfa_lookahead): Likewise.
25864 (ix86_sched_init_global): Likewise.
25865
25866 * config/i386/i386.h (TARGET_HASWELL): New macro.
25867 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25868 (processor_type): New PROCESSOR_HASWELL.
25869
955f5a07
JJ		 258702013-01-15 Jakub Jelinek <jakub@redhat.com>
25871
ff784829
JJ		 25872 PR tree-optimization/55955
25873 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25874 *SHIFT_EXPR and *ROTATE_EXPR codes.
25875
955f5a07
JJ		 25876 PR tree-optimization/48766
25877 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25878 -ftrapv disable -fwrapv.
25879
aeb8b4e9
GJL		 258802013-01-14 Georg-Johann Lay <avr@gjlay.de>
25881
25882 PR target/55974
25883 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25884 etc. to 1 and not to __flash.
25885 Use LL suffix for __INT24_MAX__ with -mint8.
25886 Use ULL suffix for __UINT24_MAX__ with -mint8.
25887
1c494c6a
GJL		 258882013-01-14 Georg-Johann Lay <avr@gjlay.de>
25889
25890 * config/avr/avr-arch.h
25891 (struct base_arch_s): Use typedef avr_arch_t instead.
25892 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25893 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25894 * config/avr/avr.c: Same.
25895 * config/avr/avr-devices.c: Same.
25896 * config/avr/driver-avr.c: Same.
25897 * config/avr/gen-avr-mmcu-texi.c: Same.
25898 * config/avr/avr-mcus.def: Adjust comment.
25899
a50344cb
TB		 259002013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25901
88e784e6
UB		 25902 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25903 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 25904
47876a2a 259052013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 25906 Andi Kleen <ak@linux.intel.com>
25907
25908 PR target/55948
25909 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25910 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25911 memmodel flag.
25912
00892272
GJL		 259132013-01-14 Georg-Johann Lay <avr@gjlay.de>
25914
25915 * config/avr/avr-stdint.h: Remove trailing blanks.
25916 * config/avr/avr-log.h: Same.
25917 * config/avr/avr-arch.h: Same.
25918 * config/avr/avr-devices.c: Same.
25919 * config/avr/avr-dimode.md: Same.
25920 * config/avr/predicates.md: Same.
25921 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 25922
00892272
GJL		 25923 * config/avr/avr-protos.h: Same. And:
25924 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25925 (init_cumulative_args): Rename to avr_init_cumulative_args.
25926 (expand_prologue): Rename to avr_expand_prologue.
25927 (expand_epilogue): Rename to avr_expand_epilogue.
25928 (adjust_insn_length): Rename to avr_adjust_insn_length.
25929 (notice_update_cc): Rename to avr_notice_update_cc.
25930 (final_prescan_insn): Rename to avr_final_prescan_insn.
25931 * config/avr/avr.c: Same.
25932 * config/avr/avr.h: Same.
25933 * config/avr/avr.md: Remove trailing blanks.
25934 (prologue): Use avr_expand_prologue.
25935 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
25936
7bb01996
RB		 259372013-01-14 Richard Biener <rguenther@suse.de>
25938
25939 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
25940 verify_location, collect_subblocks): New functions.
25941 (verify_gimple_in_cfg): Verify that locations only reference
25942 BLOCKs in the functions BLOCK tree.
25943
2724573f
RB		 259442013-01-14 Richard Biener <rguenther@suse.de>
25945
25946 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
25947 PHI argument.
25948 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
25949 unshare reference.
25950 (insert_out_of_ssa_copy_on_edge): Likewise.
25951 (rewrite_close_phi_out_of_ssa): Likewise.
25952 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
25953 debug expressions.
25954 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
25955 propagated constants.
25956 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
25957 can not be shared.
25958
9a0bbab6
GJL		 259592013-01-14 Georg-Johann Lay <avr@gjlay.de>
25960
25961 * config/avr/avr-modes.def: Add GPL copyright notice.
25962
45805f17
UB		 259632013-01-13 Uros Bizjak <ubizjak@gmail.com>
25964
25965 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
25966 MEMMODEL_MASK to determine memory model.
25967 (atomic_store<mode>): Ditto from operands[2].
25968 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
25969
9d60be38
JJ		 259702013-01-13 Jakub Jelinek <jakub@redhat.com>
25971
25972 PR fortran/55935
45805f17 25973 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 25974 (fold_gimple_assign): Don't call unshare_expr here.
25975 (fold_ctor_reference): Call unshare_expr.
25976
e7f49d92
TG		 259772013-01-13 Terry Guo <terry.guo@arm.com>
25978
aaf1e810
EB		 25979 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
25980 * doc/fragments.texi: Document MULTILIB_REUSE.
25981 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 25982 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 25983 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 25984 (tmpmultilib4): Ditto.
25985 (multilib_reuse): New multilib argument.
25986
fbd03a27
RS		 259872013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
25988
25989 * Makefile.in: Update copyright.
25990
b3681f13
TV		 259912013-01-12 Tom de Vries <tom@codesourcery.com>
25992
25993 PR middle-end/55890
25994 * calls.c (expand_call): Check if arg_nr is valid.
25995
3f287e4b
MM		 259962013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
25997
25998 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
25999 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 26000 documentation. Add missing '__' in front of
26001 __builtin_ia32_packssdw256.
3f287e4b 26002
1abcd5eb
AK		 260032013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26004
26005 PR target/55719
26006 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26007 NO_REGS for larl operands.
26008 (s390_reload_larl_operand): Use s390_load_address instead of
26009 emit_move_insn.
26010
980d0812
RB		 260112013-01-11 Richard Biener <rguenther@suse.de>
26012
26013 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26014 (verify_node_sharing): ... here.
26015 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26016
7cb9fd07
EB		 260172013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26018
26019 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26020 Substitute TREECHECKING.
26021 * configure: Regenerate.
26022 * Makefile.in (TREECHECKING): New.
26023
47876a2a 260242013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 26025
26026 PR tree-optimization/44061
26027 * tree-vrp.c (extract_range_basic): Compute zero as
26028 value-range for __builtin_constant_p of function parameters.
26029
d1e082c2
RS		 260302013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26031
45805f17 26032 Update copyright years.
d1e082c2 26033
f85021b0
VM		 260342013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26035
9d60be38 26036 PR rtl-optimization/55672
0160442c
VM		 26037 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26038 const to be eliminable.
f85021b0 26039
7a8b1ec4
DE		 260402013-01-10 David Edelsohn <dje.gcc@gmail.com>
26041
26042 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26043 * configure: Regenerate.
26044
ad2e5b71
RB		 260452013-01-10 Richard Biener <rguenther@suse.de>
26046
ddf9322d 26047 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 26048
ffc5b2bb
RB		 260492013-01-10 Richard Biener <rguenther@suse.de>
26050
26051 PR bootstrap/55792
26052 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26053 locations for virtual PHI arguments.
26054 (rewrite_update_phi_arguments): Likewise.
26055
e1f674e4
JS		 260562013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26057
26058 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26059 on to assembler.
26060
a57fc743
JJ		 260612013-01-10 Jakub Jelinek <jakub@redhat.com>
26062
26063 PR tree-optimization/55921
26064 * tree-complex.c (expand_complex_asm): New function.
26065 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26066
0ff4390d
AK		 260672013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26068
26069 PR target/55718
26070 * config/s390/s390.c (s390_symref_operand_p)
26071 (s390_loadrelative_operand_p): Merge the two functions.
26072 (s390_check_qrst_address, print_operand_address): Add parameters
26073 to s390_loadrelative_operand_p invokation.
26074 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26075 (s390_reload_larl_operand, s390_secondary_reload): Use
26076 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26077 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26078
6f557e0e
MS		 260792013-01-09 Mike Stump <mikestump@comcast.net>
26080
26081 * dse.c (record_store): Remove unnecessary assert.
26082
7770c9e9
JH		 260832013-01-09 Jan Hubicka <jh@suse.cz>
26084
26085 PR tree-optimization/55569
26086 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26087 * cfgloop.h (scale_loop_profile): Likewise.
26088
a19b1432
JH		 260892013-01-09 Jan Hubicka <jh@suse.cz>
26090
26091 PR lto/45375
ddf9322d
UB		 26092 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26093 functions.
a19b1432
JH		 26094 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26095
9a002da8
RS		 260962013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26097
26098 PR middle-end/55114
26099 * expr.h (maybe_emit_group_store): Declare.
26100 * expr.c (maybe_emit_group_store): New function.
26101 * builtins.c (expand_builtin_int_roundingfn): Call it.
26102 (expand_builtin_int_roundingfn_2): Likewise.
26103
511dcace
VM		 261042013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26105
e1f2b729 26106 PR rtl-optimization/55829
511dcace
VM		 26107 * lra-constraints.c (match_reload): Add code for absent output.
26108 (curr_insn_transform): Add code for reloads of matched inputs
26109 without output.
26110
7b0fe4f4
UB		 261112013-01-09 Uros Bizjak <ubizjak@gmail.com>
26112
26113 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26114 attribute of movddup insn to DF.
26115 (*vec_interleave_lowv2df): Ditto.
26116 (vec_dupv2df): Ditto.
26117
870ca331
JH		 261182013-01-09 Jan Hubicka <jh@suse.cz>
26119
26120 PR tree-optimiation/55875
26121 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26122 EVERY_ITERATION parameter.
7b0fe4f4 26123 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 26124 (idx_infer_loop_bounds): Similarly here.
26125 (n_of_executions_at_most): Simplify
26126 to only test for cases where statement is dominated by the
7b0fe4f4 26127 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 26128 (scev_probably_wraps_p): Use max loop iterations info
26129 as a global bound first.
26130
6f575fe4 261312013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 26132 Nick Clifton <nickc@redhat.com>
26133
26134 * config/v850/v850.md (cbranchsf4): New pattern.
26135 (cstoresf4): New pattern.
26136 (cbranchdf4): New pattern.
26137 (cstoredf4): New pattern.
26138 (movsicc): Disallow floating point comparisons.
26139 (cmpsf_le_insn): Fix order of operators.
26140 (cmpsf_lt_insn): Likewise.
26141 (cmpsf_eq_insn): Likewise.
26142 (cmpdf_le_insn): Likewise.
26143 (cmpdf_lt_insn): Likewise.
26144 (cmpdf_eq_insn): Likewise.
26145 (cmpsf_ge_insn): Use LE comparison.
26146 (cmpdf_ge_insn): Likewise.
26147 (cmpsf_gt_insn): Use LT comparison.
26148 (cmpdf_gt_insn): Likewise.
26149 (cmpsf_ne_insn): Delete pattern.
26150 (cmpdf_ne_insn): Delete pattern.
26151 * config/v850/v850.c (v850_gen_float_compare): Use
26152 gen_cmpdf_eq_insn for NE comparison.
26153 (v850_float_z_comparison_operator)
26154 (v850_float_nz_comparison_operator): Move from here ...
26155 * config/v850/predicates.md: ... to here. Move GT and GE
26156 comparisons into v850_float_z_comparison_operator.
26157 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26158 Delete prototype.
26159 (v850_float_nz_comparison_operator): Likewise.
26160
f0d54148
JDA		 261612013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26162
26163 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26164 with calls to gen_insvsi/gen_insvdi.
26165
8f01beca
VK		 261662013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26167
26168 * config/i386/i386.c (initial_ix86_tune_features): Set up
26169 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26170
2aa579ad
JJ		 261712013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26172 Jakub Jelinek <jakub@redhat.com>
26173
26174 PR tree-optimization/48189
26175 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26176 If nitercst is 0, don't predict the exit edge.
26177
6edc3e32 261782013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 26179
26180 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26181 in asm_fprintf with reg_names.
26182 (aarch64_print_operand_address): Likewise.
26183 (aarch64_return_addr): Likewise.
26184 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26185
f6f94d94
JDA		 261862013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26187
26188 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26189 (INT_U6_BITS): Likewise.
26190 * config/pa/predicates.md (uint6_operand): New predicate.
26191 (shift5_operand, shift6_operand): Likewise.
26192 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26193 arith32_operand.
26194 (lshrdi3): Use shift6_operand.
26195 (shrpsi4, shrpdi4): New insn patterns.
26196 (extzv): Delete expander.
26197 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26198 predicates in unamed zero extract patterns. Tighten common constraint.
26199 (extv): Delete expander.
26200 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26201 predicates in unamed sign extract patterns. Tighten common constraint.
26202 (insv): Delete expander.
26203 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26204 predicates in unamed insert patterns. Tighten common constraint.
26205 Change uint32_operand predicate to uint6_operand predicate in unamed
26206 DImode pattern to insert constant values of type 1...1xxxx.
26207
36b72910
JH		 262082013-01-04 Jan Hubicka <jh@suse.cz>
26209
26210 PR tree-optimization/55823
7b0fe4f4
UB		 26211 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26212 issue.
36b72910 26213
47876a2a 262142013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 26215 Uros Bizjak <ubizjak@gmail.com>
26216
26217 PR rtl-optimization/55845
26218 * df-problems.c (can_move_insns_across): Stop scanning at
26219 volatile_insn_p source instruction or give up if
26220 across_from .. across_to range contains any volatile_insn_p
26221 instructions.
26222
4369c11e
TB		 262232013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26224
7b0fe4f4
UB		 26225 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26226 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26227 Declare.
4369c11e 26228 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26229 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26230
aeb7e7c1
JJ		 262312013-01-08 Jakub Jelinek <jakub@redhat.com>
26232
26233 PR fortran/55341
26234 * asan.c (asan_clear_shadow): New function.
26235 (asan_emit_stack_protection): Use it.
26236
a02ad1aa
TB		 262372013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26238
26239 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26240 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26241 with tab instead of space.
26242
f90d87f5
NC		 262432013-01-08 Nick Clifton <nickc@redhat.com>
26244
26245 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26246 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 26247 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26248 MDBH registers.
f90d87f5 26249
385eb93d
JG		 262502013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26251
26252 * config/aarch64/aarch64-simd.md
26253 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26254 (aarch64_simd_bsl): Likewise.
26255 (aarch64_vcond_internal<mode>): Likewise.
26256 (vcond<mode><mode>): Likewise.
26257 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26258 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26259
4dcd1054
JG		 262602013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26261
26262 * config/aarch64/aarch64-builtins.c
26263 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26264
4267a4a6
MJ		 262652013-01-08 Martin Jambor <mjambor@suse.cz>
26266
26267 PR debug/55579
26268 * tree-sra.c (analyze_access_subtree): Return true also after
26269 potentially creating a debug-only replacement.
26270
5f4e6de3
JJ		 262712013-01-08 Jakub Jelinek <jakub@redhat.com>
26272
3138f224
JJ		 26273 PR middle-end/55890
26274 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26275
5f4e6de3
JJ		 26276 PR tree-optimization/54120
26277 * tree-vrp.c (range_fits_type_p): Don't allow
26278 src_precision < precision from signed vr to unsigned_p
26279 if vr->min or vr->max is negative.
26280 (simplify_float_conversion_using_ranges): Test can_float_p
26281 against CODE_FOR_nothing.
26282
f3ef18ff
JJ		 262832013-01-08 Jakub Jelinek <jakub@redhat.com>
26284 Richard Biener <rguenther@suse.de>
26285
26286 PR middle-end/55851
26287 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26288 types instead of just INTEGER_TYPE types.
26289
25c210f9
MK		 262902013-01-07 Mark Kettenis <kettenis@openbsd.org>
26291
26292 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26293 TF_SIZE): Define.
f90d87f5 26294
81826a7b
SE		 262952013-01-07 Steve Ellcey <sellcey@mips.com>
26296
26297 PR target/42661
26298 * config/mips/mips.opt: Change mad to mmad to match documentation.
26299
35678514
GJL		 263002013-01-07 Georg-Johann Lay <avr@gjlay.de>
26301
26302 PR target/55897
26303 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26304 .progmemx.data now.
26305
e5669488
GJL		 263062013-01-07 Georg-Johann Lay <avr@gjlay.de>
26307
26308 PR target/55897
26309 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26310 (avr_addrspace_t): Add .section_name field.
26311 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26312 array size.
26313 (avr_addrspace): Same. Initialize .section_name. Remove last
26314 NULL entry. Put __memx into .progmemx.data.
26315 (progmem_section_prefix): Remove.
26316 (avr_asm_init_sections): No need to initialize progmem_section.
26317 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26318 section name prefix.
26319 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26320 retrieve the progmem section.
26321 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26322 boundary to run over avr_addrspace[].
26323 (avr_register_target_pragmas): Ditto.
26324
656e6f37
JJ		 263252013-01-06 Jakub Jelinek <jakub@redhat.com>
26326
26327 * varasm.c (output_constant_def_contents): For asan_protect_global
26328 protected strings, adjust DECL_ALIGN if needed, before testing for
26329 anchored symbols.
26330 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26331 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26332 normal decls.
26333 (output_object_block): For asan protected decls, emit asan padding
26334 after their contents.
26335 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26336 (asan_finish_file): Test it here instead.
26337
6edc3e32
UB		 263382013-01-07 Nick Clifton <nickc@redhat.com>
26339 Matthias Klose <doko@debian.org>
26340 Doug Kwan <dougkwan@google.com>
26341 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 26342
26343 PR driver/55470
26344 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26345
26346 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26347
26348 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26349
7b0fe4f4 26350 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 26351
26352 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26353
2a095093
GJL		 263542013-01-07 Georg-Johann Lay <avr@gjlay.de>
26355
f3b503f4 26356 PR target/54461
2a095093
GJL		 26357 * doc/install.texi (Cross-Compiler-Specific Options): Document
26358 --with-avrlibc.
26359
383f9b34
TB		 263602013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26361
26362 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26363 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26364 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26365 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26366 vqmovun_high_s64): Fix source operand number and update copyright.
26367
3626621a
RB		 263682013-01-07 Richard Biener <rguenther@suse.de>
26369
26370 PR middle-end/55890
26371 * gimple.h (gimple_call_builtin_p): New overload.
26372 * gimple.c (validate_call): New function.
26373 (gimple_call_builtin_p): Likewise.
26374 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26375 Use gimple_call_builtin_p.
26376 (find_func_clobbers): Likewise.
26377 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26378 (strlen_optimize_stmt): Likewise.
26379
8b2ea410
JG		 263802013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26381
26382 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26383 (vld1q_dup_*): Likewise.
26384 (vld1_*): Likewise.
26385 (vld1q_*): Likewise.
26386 (vld1_lane_*): Likewise.
26387 (vld1q_lane_*): Likewise.
26388
9713d329
RB		 263892013-01-07 Richard Biener <rguenther@suse.de>
26390
26391 * lto-streamer.h (LTO_minor_version): Bump to 2.
26392
3520f7cc
JG		 263932013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26394
26395 * config/aarch64/aarch64-protos.h
26396 (aarch64_const_double_zero_rtx_p): Rename to...
26397 (aarch64_float_const_zero_rtx_p): ...this.
26398 (aarch64_float_const_representable_p): New.
26399 (aarch64_output_simd_mov_immediate): Likewise.
26400 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26401 move immediate case.
26402 * config/aarch64/aarch64.c
26403 (aarch64_const_double_zero_rtx_p): Rename to...
26404 (aarch64_float_const_zero_rtx_p): ...this.
26405 (aarch64_print_operand): Allow printing of new constants.
26406 (aarch64_valid_floating_const): New.
26407 (aarch64_legitimate_constant_p): Check for valid floating-point
26408 constants.
26409 (aarch64_simd_valid_immediate): Likewise.
26410 (aarch64_vect_float_const_representable_p): New.
26411 (aarch64_float_const_representable_p): Likewise.
26412 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26413 (aarch64_output_simd_mov_immediate): New.
26414 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26415 (*movdf_aarch64): Likewise.
26416 * config/aarch64/constraints.md (Ufc): New.
26417 (Y): call aarch64_float_const_zero_rtx.
26418 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26419
e167c04d
RB		 264202013-01-07 Richard Biener <rguenther@suse.de>
26421
26422 PR tree-optimization/55888
26423 PR tree-optimization/55862
26424 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26425 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26426 not if it is contained therein.
26427
0139aaab
GJL		 264282013-01-07 Georg-Johann Lay <avr@gjlay.de>
26429
26430 * config/avr/t-avr: Typo.
26431
4a176b23
GJL		 264322013-01-07 Georg-Johann Lay <avr@gjlay.de>
26433
26434 PR55243
26435 * config/avr/t-avr: Don't automatically rebuild
26436 $(srcdir)/config/avr/t-multilib
26437 $(srcdir)/config/avr/avr-tables.opt
26438 $(srcdir)/doc/avr-mmcu.texi
26439 (avr-mcus): New phony target to build them on request.
26440 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26441 * avr/avr-mcus.def: Adjust comments.
26442
c7afdc98
UB		 264432013-01-07 Uros Bizjak <ubizjak@gmail.com>
26444
26445 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26446
1ab05c31
RS		 264472013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26448
26449 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26450
488e3acc
RS		 264512013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26452
26453 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26454
a89599d2
DE		 264552013-01-05 David Edelsohn <dje.gcc@gmail.com>
26456
26457 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26458 to generate profiling.
26459 * config/rs6000/aix64.h (LIB_SPEC): Same.
26460
70f09188
AP		 264612013-01-04 Andrew Pinski <apinski@cavium.com>
26462
26463 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26464 New function.
26465 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26466
918d445f
UB		 264672013-01-04 Uros Bizjak <ubizjak@gmail.com>
26468
26469 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26470 unconditionally.
26471 (ix86_expand_move): Ditto.
26472 (ix86_zero_extend_to_Pmode): Ditto.
26473 (ix86_expand_call): Ditto.
26474 (ix86_expand_special_args_builtin): Ditto.
26475 (ix86_expand_builtin): Ditto.
26476
361618ec
RB		 264772013-01-04 Richard Biener <rguenther@suse.de>
26478
26479 PR tree-optimization/55862
26480 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26481 translating them through PHI nodes.
26482
87eab554
MJ		 264832013-01-04 Martin Jambor <mjambor@suse.cz>
26484
26485 PR tree-optimization/55755
26486 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26487 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26488
33879b9f
MP		 264892013-01-04 Marek Polacek <polacek@redhat.com>
26490
26491 PR middle-end/55859
26492 * opts.c (default_options_optimization): Clarify error message.
26493
3068819a
RB		 264942013-01-04 Richard Biener <rguenther@suse.de>
26495
26496 PR middle-end/55863
26497 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26498 reassociation.
26499
0e4ae794
JDA		 265002013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26501
26502 PR target/53789
26503 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 26504 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26505 references.
0e4ae794 26506
65c88cef
RH		 265072013-01-03 Richard Henderson <rth@redhat.com>
26508
26509 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26510 after eliminating TLS symbols.
26511
8ac16127
MG		 265122013-01-03 Marc Glisse <marc.glisse@inria.fr>
26513
26514 PR bootstrap/50167
26515 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26516 * graphite-poly.c (debug_gmp_value): Likewise.
26517
bb664f09
UB		 265182013-01-03 Uros Bizjak <ubizjak@gmail.com>
26519
26520 PR target/55712
26521 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26522 selected code model, define __code_mode_small__, __code_model_medium__,
26523 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26524 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26525 xchg temporary register with %k. Declare temporary register as
26526 early clobbered.
26527 [__x86_64__]: For medium and large code models, preserve %rbx register.
26528
bcf1ef00
RB		 265292013-01-03 Richard Biener <rguenther@suse.de>
26530
0506634a 26531 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 26532 (dump_subscript): Adjust.
26533 (finalize_ddr_dependent): Do not dump redundant info.
26534 (analyze_siv_subscript): Adjust.
26535 (subscript_dependence_tester): Likewise.
26536 (compute_affine_dependence): Likewise.
26537
59fd17e3
RB		 265382013-01-03 Richard Biener <rguenther@suse.de>
26539
26540 Revert
26541 2013-01-03 Richard Biener <rguenther@suse.de>
26542
26543 PR tree-optimization/55857
26544 * tree-vect-stmts.c (vectorizable_load): Do not setup
26545 re-alignment for invariant loads.
26546
26547 2013-01-02 Richard Biener <rguenther@suse.de>
26548
26549 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26550 invariant load do not generate a vector load from the scalar location.
59fd17e3 26551
595c2679
RB		 265522013-01-03 Richard Biener <rguenther@suse.de>
26553
26554 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26555 for not vectorizing.
26556 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26557 not build INDIRECT_REFs, call get_name once only.
26558 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26559 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26560
90bb7d7a
RB		 265612013-01-03 Richard Biener <rguenther@suse.de>
26562
26563 PR tree-optimization/55857
26564 * tree-vect-stmts.c (vectorizable_load): Do not setup
26565 re-alignment for invariant loads.
26566
f09b77ca
RB		 265672013-01-03 Richard Biener <rguenther@suse.de>
26568
26569 PR lto/55848
26570 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26571 prefer a built-in decl.
26572
3461a16e
JJ		 265732013-01-03 Jakub Jelinek <jakub@redhat.com>
26574
df8e2b4f
JJ		 26575 * gcc.c (process_command): Update copyright notice dates.
26576 * gcov.c (print_version): Likewise.
26577 * gcov-dump.c (print_version): Likewise.
26578
3461a16e
JJ		 26579 PR rtl-optimization/55838
26580 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26581 iv0.step, iv1.step and step.
26582
8b5546d6
JJ		 265832013-01-03 Jakub Jelinek <jakub@redhat.com>
26584 Marc Glisse <marc.glisse@inria.fr>
26585
26586 PR tree-optimization/55832
26587 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26588 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26589 integer_{one,zero}_node.
26590
8ab1d2e9
JJ		 265912013-01-03 Jakub Jelinek <jakub@redhat.com>
26592
26593 PR debug/54402
26594 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26595 * var-tracking.c (reverse_op): Don't add reverse ops to
26596 VALUEs that have already
0506634a 26597 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 26598
5b9ad1d4
GP		 265992013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26600
26601 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26602
e25a6711
TJ		 266032013-01-02 Teresa Johnson <tejohnson@google.com>
26604
26605 * dumpfile.c (dump_loc): Print filename with location.
26606 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26607 new location_t parameter to emit complete unroll message with
26608 new dump framework.
26609 (canonicalize_loop_induction_variables): Compute loops location
26610 and pass to try_unroll_loop_completely.
26611 * loop-unroll.c (report_unroll_peel): New function.
26612 (peel_loops_completely): Use new dump format with location
26613 for main dumpfile message, and invoke report_unroll_peel on success.
26614 (decide_unrolling_and_peeling): Ditto.
26615 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26616 by report_unroll_peel.
26617 (decide_peel_completely): Ditto.
26618 (decide_unroll_constant_iterations): Ditto.
26619 (decide_unroll_runtime_iterations): Ditto.
26620 (decide_peel_simple): Ditto.
26621 (decide_unroll_stupid): Ditto.
26622 * cfgloop.c (get_loop_location): New function.
26623 * cfgloop.h (get_loop_location): Declare.
26624
77878621
ST		 266252013-01-02 Sriraman Tallam <tmsriram@google.com>
26626
26627 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26628 NULL.
26629
9e65d03e
JDA		 266302013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26631
26632 PR middle-end/55198
26633 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26634 BLKmode objects when EXPAND_MEMORY is specified.
26635
6a7da30f
ST		 266362013-01-02 Sriraman Tallam <tmsriram@google.com>
26637
26638 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26639 in loop predicate.
26640 (fold_builtin_cpu): Do not share cpu model decls across statements.
26641
e78167f2
JM		 266422013-01-02 Jason Merrill <jason@redhat.com>
26643
26644 PR c++/55804
26645 * tree.c (build_array_type_1): Revert earlier change.
26646
8c075fb4
YZ		 266472013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26648
26649 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26650 "cortex-a57".
26651 * config/aarch64/aarch64-tune.md: Re-generate.
26652
0682ed3e
RB		 266532013-01-02 Richard Biener <rguenther@suse.de>
26654
26655 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26656 invariant load do not generate a vector load from the scalar location.
0682ed3e 26657
b8f6a302
RB		 266582013-01-02 Richard Biener <rguenther@suse.de>
26659
26660 PR bootstrap/55784
26661 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26662 * configure: Regenerate.
26663
04b80dbb
RS		 266642013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26665
26666 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26667 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26668 (expand_builtin_int_roundingfn_2): Keep the original target around
26669 for the fallback case.
26670
635b0b0c
RS		 266712013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26672
26673 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26674 to be clear for sign changes.
26675
42d57399
JH		 266762013-01-01 Jan Hubicka <jh@suse.cz>
26677
26678 * ipa-inline-analysis.c: Fix formatting.
26679
5bb6669d
JJ		 266802013-01-01 Jakub Jelinek <jakub@redhat.com>
26681
26682 PR tree-optimization/55831
26683 * tree-vect-loop.c (get_initial_def_for_induction): Use
26684 gsi_after_labels instead of gsi_start_bb.
ad41bd84 26685\f
86a2db33 26686Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 26687
26688Copying and distribution of this file, with or without modification,
26689are permitted in any medium without royalty provided the copyright
26690notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git