1 2013-09-05 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58137
4 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
5 Do not create vectors of pointers.
6 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7 types for the components of the vector initializer.
8 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
11 2013-09-05 Martin Jambor <mjambor@suse.cz>
13 * ipa-prop.c (remove_described_reference): Accept missing references,
14 return false if that hppens, otherwise return true.
15 (cgraph_node_for_jfunc): New function.
16 (try_decrement_rdesc_refcount): Likewise.
17 (try_make_edge_direct_simple_call): Use them.
18 (ipa_edge_removal_hook): Remove references from rdescs.
19 (ipa_edge_duplication_hook): Clone rdescs and their references
20 when the new edge has the same caller as the old one.
21 * cgraph.c (cgraph_resolve_speculation): Remove speculative
22 reference before removing any edges.
24 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
26 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
28 * thumb2.md (thumb2_storewb_parisi): New pattern.
30 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
32 * config/aarch64/aarch64-option-extensions.def: Add
33 AARCH64_OPT_EXTENSION of 'crc'.
34 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
35 (AARCH64_ISA_CRC): Ditto.
36 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
37 description of the CRC extension.
39 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
41 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
42 * config/rs6000/linux.h: Ditto.
43 * alpha/linux.h: Ditto.
44 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
45 no_c99_libc_has_function.
46 * config/c6x/uclinux-elf.h: Ditto.
47 * config/lm32/uclinux-elf.h: Ditto.
48 * config/m68k/uclinux.h: Ditto.
49 * config/moxie/uclinux.h: Ditto.
50 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
51 (crisv32-*-linux*, cris-*-linux*): Ditto.
52 * config/bfin/bfin.c: Include "tm_p.h".
54 2013-09-05 Richard Biener <rguenther@suse.de>
56 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
57 check for a definition without a basic-block.
59 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
60 Sofiane Naci <sofiane.naci@arm.com>
62 * config/aarch64/aarch64.md
63 (*movti_aarch64): Rename r_2_f and f_2_r.
64 (*movsf_aarch64): Likewise.
65 (*movdf_aarch64): Likewise.
66 (*movtf_aarch64): Likewise.
67 (aarch64_movdi_<mode>low): Likewise.
68 (aarch64_movdi_<mode>high): Likewise.
69 (aarch64_mov<mode>high_di): Likewise.
70 (aarch64_mov<mode>low_di): Likewise.
71 (aarch64_movtilow_tilow): Likewise.
72 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
73 values to config/arm/types.md
74 (attribute "conds"): Update for attribute change.
75 (anddi3_insn): Likewise.
76 (iordi3_insn): Likewise.
77 (xordi3_insn): Likewise.
78 (one_cmpldi2): Likewise.
79 * config/arm/types.md (type): Add Neon types.
80 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
82 (movmisalign<mode>_neon_store): Likewise.
83 (movmisalign<mode>_neon_load): Likewise.
84 (vec_set<mode>_internal): Likewise.
85 (vec_setv2di_internal): Likewise.
86 (vec_extract<mode>): Likewise.
87 (vec_extractv2di): Likewise.
88 (add<mode>3_neon): Likewise.
89 (adddi3_neon): Likewise.
90 (sub<mode>3_neon): Likewise.
91 (subdi3_neon): Likewise.
92 (mul<mode>3_neon): Likewise.
93 (mul<mode>3add<mode>_neon): Likewise.
94 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
95 (fma<VCVTF:mode>4)): Likewise.
96 (fma<VCVTF:mode>4_intrinsic): Likewise.
97 (fmsub<VCVTF:mode>4)): Likewise.
98 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
99 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
100 (ior<mode>3): Likewise.
101 (and<mode>3): Likewise.
102 (anddi3_neon): Likewise.
103 (orn<mode>3_neon): Likewise.
104 (orndi3_neon): Likewise.
105 (bic<mode>3_neon): Likewise.
106 (bicdi3_neon): Likewise.
107 (xor<mode>3): Likewise.
108 (one_cmpl<mode>2): Likewise.
109 (abs<mode>2): Likewise.
110 (neg<mode>2): Likewise.
111 (umin<mode>3_neon): Likewise.
112 (umax<mode>3_neon): Likewise.
113 (smin<mode>3_neon): Likewise.
114 (smax<mode>3_neon): Likewise.
115 (vashl<mode>3): Likewise.
116 (vashr<mode>3_imm): Likewise.
117 (vlshr<mode>3_imm): Likewise.
118 (ashl<mode>3_signed): Likewise.
119 (ashl<mode>3_unsigned): Likewise.
120 (neon_load_count): Likewise.
121 (ashldi3_neon_noclobber): Likewise.
122 (signed_shift_di3_neon): Likewise.
123 (unsigned_shift_di3_neon): Likewise.
124 (ashrdi3_neon_imm_noclobber): Likewise.
125 (lshrdi3_neon_imm_noclobber): Likewise.
126 (widen_ssum<mode>3): Likewise.
127 (widen_usum<mode>3): Likewise.
128 (quad_halves_<code>v4si): Likewise.
129 (quad_halves_<code>v4sf): Likewise.
130 (quad_halves_<code>v8hi): Likewise.
131 (quad_halves_<code>v16qi): Likewise.
132 (reduc_splus_v2di): Likewise.
133 (neon_vpadd_internal<mode>): Likewise.
134 (neon_vpsmin<mode>): Likewise.
135 (neon_vpsmax<mode>): Likewise.
136 (neon_vpumin<mode>): Likewise.
137 (neon_vpumax<mode>): Likewise.
138 (ss_add<mode>_neon): Likewise.
139 (us_add<mode>_neon): Likewise.
140 (ss_sub<mode>_neon): Likewise.
141 (us_sub<mode>_neon): Likewise.
142 (neon_vadd<mode>_unspec): Likewise.
143 (neon_vaddl<mode>): Likewise.
144 (neon_vaddw<mode>): Likewise.
145 (neon_vhadd<mode>): Likewise.
146 (neon_vqadd<mode>): Likewise.
147 (neon_vaddhn<mode>): Likewise.
148 (neon_vmul<mode>): Likewise.
149 (neon_vmla<mode>): Likewise.
150 (neon_vmlal<mode>): Likewise.
151 (neon_vmls<mode>): Likewise.
152 (neon_vmlsl<mode>): Likewise.
153 (neon_vqdmulh<mode>): Likewise.
154 (neon_vqdmlal<mode>): Likewise.
155 (neon_vqdmlsl<mode>): Likewise.
156 (neon_vmull<mode>): Likewise.
157 (neon_vqdmull<mode>): Likewise.
158 (neon_vsub<mode>_unspec): Likewise.
159 (neon_vsubl<mode>): Likewise.
160 (neon_vsubw<mode>): Likewise.
161 (neon_vqsub<mode>): Likewise.
162 (neon_vhsub<mode>): Likewise.
163 (neon_vsubhn<mode>): Likewise.
164 (neon_vceq<mode>): Likewise.
165 (neon_vcge<mode>): Likewise.
166 (neon_vcgeu<mode>): Likewise.
167 (neon_vcgt<mode>): Likewise.
168 (neon_vcgtu<mode>): Likewise.
169 (neon_vcle<mode>): Likewise.
170 (neon_vclt<mode>): Likewise.
171 (neon_vcage<mode>): Likewise.
172 (neon_vcagt<mode>): Likewise.
173 (neon_vtst<mode>): Likewise.
174 (neon_vabd<mode>): Likewise.
175 (neon_vabdl<mode>): Likewise.
176 (neon_vaba<mode>): Likewise.
177 (neon_vabal<mode>): Likewise.
178 (neon_vmax<mode>): Likewise.
179 (neon_vmin<mode>): Likewise.
180 (neon_vpaddl<mode>): Likewise.
181 (neon_vpadal<mode>): Likewise.
182 (neon_vpmax<mode>): Likewise.
183 (neon_vpmin<mode>): Likewise.
184 (neon_vrecps<mode>): Likewise.
185 (neon_vrsqrts<mode>): Likewise.
186 (neon_vqabs<mode>): Likewise.
187 (neon_vqneg<mode>): Likewise.
188 (neon_vcls<mode>): Likewise.
189 (clz<mode>2): Likewise.
190 (popcount<mode>2): Likewise.
191 (neon_vrecpe): Likewise.
192 (neon_vrsqrte): Likewise.
193 (neon_vget_lane<mode>_sext_internal): Likewise.
194 (neon_vget_lane<mode>_zext_internal): Likewise.
195 (neon_vdup_n<mode>): Likewise.
196 (neon_vdup_nv2di): Likewise.
197 (neon_vdpu_lane<mode>_internal): Likewise.
198 (neon_vswp<mode>): Likewise.
199 (float<mode><V_cvtto>2): Likewise.
200 (floatuns<mode><V_cvtto>2): Likewise.
201 (fix_trunc<mode><V_cvtto>)2): Likewise
202 (fixuns_trunc<mode><V_cvtto)2): Likewise.
203 (neon_vcvt<mode>): Likewise.
204 (neon_vcvtv4sfv4hf): Likewise.
205 (neon_vcvtv4hfv4sf): Likewise.
206 (neon_vcvt_n<mode>): Likewise.
207 (neon_vmovn<mode>): Likewise.
208 (neon_vqmovn<mode>): Likewise.
209 (neon_vqmovun<mode>): Likewise.
210 (neon_vmovl<mode>): Likewise.
211 (neon_vmul_lane<mode>): Likewise.
212 (neon_vmull_lane<mode>): Likewise.
213 (neon_vqdmull_lane<mode>): Likewise.
214 (neon_vqdmulh_lane<mode>): Likewise.
215 (neon_vmla_lane<mode>): Likewise.
216 (neon_vmlal_lane<mode>): Likewise.
217 (neon_vqdmlal_lane<mode>): Likewise.
218 (neon_vmls_lane<mode>): Likewise.
219 (neon_vmlsl_lane<mode>): Likewise.
220 (neon_vqdmlsl_lane<mode>): Likewise.
221 (neon_vext<mode>): Likewise.
222 (neon_vrev64<mode>): Likewise.
223 (neon_vrev32<mode>): Likewise.
224 (neon_vrev16<mode>): Likewise.
225 (neon_vbsl<mode>_internal): Likewise.
226 (neon_vshl<mode>): Likewise.
227 (neon_vqshl<mode>): Likewise.
228 (neon_vshr_n<mode>): Likewise.
229 (neon_vshrn_n<mode>): Likewise.
230 (neon_vqshrn_n<mode>): Likewise.
231 (neon_vqshrun_n<mode>): Likewise.
232 (neon_vshl_n<mode>): Likewise.
233 (neon_vqshl_n<mode>): Likewise.
234 (neon_vqshlu_n<mode>): Likewise.
235 (neon_vshll_n<mode>): Likewise.
236 (neon_vsra_n<mode>): Likewise.
237 (neon_vsri_n<mode>): Likewise.
238 (neon_vsli_n<mode>): Likewise.
239 (neon_vtbl1v8qi): Likewise.
240 (neon_vtbl2v8qi): Likewise.
241 (neon_vtbl3v8qi): Likewise.
242 (neon_vtbl4v8qi): Likewise.
243 (neon_vtbx1v8qi): Likewise.
244 (neon_vtbx2v8qi): Likewise.
245 (neon_vtbx3v8qi): Likewise.
246 (neon_vtbx4v8qi): Likewise.
247 (neon_vtrn<mode>_internal): Likewise.
248 (neon_vzip<mode>_internal): Likewise.
249 (neon_vuzp<mode>_internal): Likewise.
250 (neon_vld1<mode>): Likewise.
251 (neon_vld1_lane<mode>): Likewise.
252 (neon_vld1_dup<mode>): Likewise.
253 (neon_vld1_dupv2di): Likewise.
254 (neon_vst1<mode>): Likewise.
255 (neon_vst1_lane<mode>): Likewise.
256 (neon_vld2<mode>): Likewise.
257 (neon_vld2_lane<mode>): Likewise.
258 (neon_vld2_dup<mode>): Likewise.
259 (neon_vst2<mode>): Likewise.
260 (neon_vst2_lane<mode>): Likewise.
261 (neon_vld3<mode>): Likewise.
262 (neon_vld3qa<mode>): Likewise.
263 (neon_vld3qb<mode>): Likewise.
264 (neon_vld3_lane<mode>): Likewise.
265 (neon_vld3_dup<mode>): Likewise.
266 (neon_vst3<mode>): Likewise.
267 (neon_vst3qa<mode>): Likewise.
268 (neon_vst3qb<mode>): Likewise.
269 (neon_vst3_lane<mode>): Likewise.
270 (neon_vld4<mode>): Likewise.
271 (neon_vld4qa<mode>): Likewise.
272 (neon_vld4qb<mode>): Likewise.
273 (neon_vld4_lane<mode>): Likewise.
274 (neon_vld4_dup<mode>): Likewise.
275 (neon_vst4<mode>): Likewise.
276 (neon_vst4qa<mode>): Likewise.
277 (neon_vst4qb<mode>): Likewise.
278 (neon_vst4_lane<mode>): Likewise.
279 (neon_vec_unpack<US>_lo_<mode>): Likewise.
280 (neon_vec_unpack<US>_hi_<mode>): Likewise.
281 (neon_vec_<US>mult_lo_<mode>): Likewise.
282 (neon_vec_<US>mult_hi_<mode>): Likewise.
283 (neon_vec_<US>shiftl_<mode>): Likewise.
284 (neon_unpack<US>_<mode>): Likewise.
285 (neon_vec_<US>mult_<mode>): Likewise.
286 (vec_pack_trunc_<mode>): Likewise.
287 (neon_vec_pack_trunk_<mode>): Likewise.
288 (neon_vabd<mode>_2): Likewise.
289 (neon_vabd<mode>_3): Likewise.
290 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
291 (thumb2_movsi_vfp): Likewise.
292 (movdi_vfp): Likewise.
293 (movdi_vfp_cortexa8): Likewise.
294 (movhf_vfp_neon): Likewise.
295 (movhf_vfp): Likewiwse.
296 (movsf_vfp): Likewiwse.
297 (thumb2_movsf_vfp): Likewiwse.
298 (movdf_vfp): Likewise.
299 (thumb2_movdf_vfp): Likewise.
300 (movsfcc_vfp): Likewise.
301 (thumb2_movsfcc_vfp): Likewise.
302 (movdfcc_vfp): Likewise.
303 (thumb2_movdfcc_vfp): Likewise.
304 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
305 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
307 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
309 (cortex_a15_neon_int_2): Likewise.
310 (cortex_a15_neon_int_3): Likewise.
311 (cortex_a15_neon_int_4): Likewise.
312 (cortex_a15_neon_int_5): Likewise.
313 (cortex_a15_neon_vqneg_vqabs): Likewise.
314 (cortex_a15_neon_vmov): Likewise.
315 (cortex_a15_neon_vaba): Likewise.
316 (cortex_a15_neon_vaba_qqq): Likewise.
317 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
318 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
319 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
320 scalar_64_32_long_scalar): Likewise.
321 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
322 (cortex_a15_neon_mla_qqq_8_16): Likewise.
323 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
324 lotype_qdd_64_32_long): Likewise.
325 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
326 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
327 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
328 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
329 (cortex_a15_neon_shift_1): Likewise.
330 (cortex_a15_neon_shift_2): Likewise.
331 (cortex_a15_neon_shift_3): Likewise.
332 (cortex_a15_neon_vshl_ddd): Likewise.
333 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
334 (cortex_a15_neon_vsra_vrsra): Likewise.
335 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
336 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
337 (cortex_a15_neon_fp_vmul_ddd): Likewise.
338 (cortex_a15_neon_fp_vmul_qqd): Likewise.
339 (cortex_a15_neon_fp_vmla_ddd): Likewise.
340 (cortex_a15_neon_fp_vmla_qqq): Likewise.
341 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
342 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
343 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
344 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
345 (cortex_a15_neon_bp_simple): Likewise.
346 (cortex_a15_neon_bp_2cycle): Likewise.
347 (cortex_a15_neon_bp_3cycle): Likewise.
348 (cortex_a15_neon_vld1_1_2_regs): Likewise.
349 (cortex_a15_neon_vld1_3_4_regs): Likewise.
350 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
351 (cortex_a15_neon_vld2_4_regs): Likewise.
352 (cortex_a15_neon_vld3_vld4): Likewise.
353 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
354 (cortex_a15_neon_vst1_3_4_regs): Likewise.
355 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
356 (cortex_a15_neon_vst3_vst4): Likewise.
357 (cortex_a15_neon_vld1_vld2_lane): Likewise.
358 (cortex_a15_neon_vld3_vld4_lane" 10
359 (cortex_a15_neon_vst1_vst2_lane): Likewise.
360 (cortex_a15_neon_vst3_vst4_lane): Likewise.
361 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
362 (cortex_a15_neon_ldm_2): Likewise.0
363 (cortex_a15_neon_stm_2): Likewise.
364 (cortex_a15_neon_mcr): Likewise.
365 (cortex_a15_neon_mcr_2_mcrr): Likewise.
366 (cortex_a15_neon_mrc): Likewise.
367 (cortex_a15_neon_mrrc): Likewise.
368 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
370 (cortex_a15_alu_shift): Likewise.
371 (cortex_a15_alu_shift_reg): Likewise.
372 (cortex_a15_mult32): Likewise.
373 (cortex_a15_mult64): Likewise.
374 (cortex_a15_block): Likewise.
375 (cortex_a15_branch): Likewise.
376 (cortex_a15_load1): Likewise.
377 (cortex_a15_load3): Likewise.
378 (cortex_a15_store1): Likewise.
379 (cortex_a15_store3): Likewise.
380 (cortex_a15_call): Likewise.
381 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute
383 (cortex_a5_f2r): Likewise.
384 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
386 (cortex_a53_f2r): Likewise.
387 * config/arm/cortex-a7.md
388 (cortex_a7_branch): Update for attribute change.
389 (cortex_a7_call): Likewise.
390 (cortex_a7_alu_imm): Likewise.
391 (cortex_a7_alu_reg): Likewise.
392 (cortex_a7_alu_shift): Likewise.
393 (cortex_a7_mul): Likewise.
394 (cortex_a7_load1): Likewise.
395 (cortex_a7_store1): Likewise.
396 (cortex_a7_load2): Likewise.
397 (cortex_a7_store2): Likewise.
398 (cortex_a7_load3): Likewise.
399 (cortex_a7_store3): Likewise.
400 (cortex_a7_load4): Likewise.
401 (cortex_a7_store4): Likewise.
402 (cortex_a7_fpalu): Likewise.
403 (cortex_a7_fconst): Likewise.
404 (cortex_a7_fpmuls): Likewise.
405 (cortex_a7_neon_mul): Likewise.
406 (cortex_a7_fpmacs): Likewise.
407 (cortex_a7_neon_mla: Likewise.
408 (cortex_a7_fpmuld: Likewise.
409 (cortex_a7_fpmacd: Likewise.
410 (cortex_a7_fpfmad: Likewise.
411 (cortex_a7_fdivs: Likewise.
412 (cortex_a7_fdivd: Likewise.
413 (cortex_a7_r2f: Likewise.
414 (cortex_a7_f2r: Likewise.
415 (cortex_a7_f_flags: Likewise.
416 (cortex_a7_f_loads: Likewise.
417 (cortex_a7_f_loadd: Likewise.
418 (cortex_a7_f_stores: Likewise.
419 (cortex_a7_f_stored: Likewise.
420 (cortex_a7_neon): Likewise.
421 * config/arm/cortex-a8-neon.md
422 (cortex_a8_neon_mrc): Update for attribute change.
423 (cortex_a8_neon_mrrc): Likewise.
424 (cortex_a8_neon_int_1): Likewise.
425 (cortex_a8_neon_int_2): Likewise.
426 (cortex_a8_neon_int_3): Likewise.
427 (cortex_a8_neon_int_4): Likewise.
428 (cortex_a8_neon_int_5): Likewise.
429 (cortex_a8_neon_vqneg_vqabs): Likewise.
430 (cortex_a8_neon_vmov): Likewise.
431 (cortex_a8_neon_vaba): Likewise.
432 (cortex_a8_neon_vaba_qqq): Likewise.
433 (cortex_a8_neon_vsma): Likewise.
434 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
435 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
436 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
438 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
439 (cortex_a8_neon_mla_qqq_8_16): Likewise.
440 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
441 long_scalar_qdd_64_32_long): Likewise.
442 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
443 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
444 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
445 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
446 (cortex_a8_neon_shift_1): Likewise.
447 (cortex_a8_neon_shift_2): Likewise.
448 (cortex_a8_neon_shift_3): Likewise.
449 (cortex_a8_neon_vshl_ddd): Likewise.
450 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
451 (cortex_a8_neon_vsra_vrsra): Likewise.
452 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
453 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
454 (cortex_a8_neon_fp_vsum): Likewise.
455 (cortex_a8_neon_fp_vmul_ddd): Likewise.
456 (cortex_a8_neon_fp_vmul_qqd): Likewise.
457 (cortex_a8_neon_fp_vmla_ddd): Likewise.
458 (cortex_a8_neon_fp_vmla_qqq): Likewise.
459 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
460 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
461 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
462 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
463 (cortex_a8_neon_bp_simple): Likewise.
464 (cortex_a8_neon_bp_2cycle): Likewise.
465 (cortex_a8_neon_bp_3cycle): Likewise.
466 (cortex_a8_neon_ldr): Likewise.
467 (cortex_a8_neon_str): Likewise.
468 (cortex_a8_neon_vld1_1_2_regs): Likewise.
469 (cortex_a8_neon_vld1_3_4_regs): Likewise.
470 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
471 (cortex_a8_neon_vld2_4_regs): Likewise.
472 (cortex_a8_neon_vld3_vld4): Likewise.
473 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
474 (cortex_a8_neon_vst1_3_4_regs): Likewise.
475 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
476 (cortex_a8_neon_vst3_vst4): Likewise.
477 (cortex_a8_neon_vld1_vld2_lane): Likewise.
478 (cortex_a8_neon_vld3_vld4_lane): Likewise.
479 (cortex_a8_neon_vst1_vst2_lane): Likewise.
480 (cortex_a8_neon_vst3_vst4_lane): Likewise.
481 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
482 (cortex_a8_neon_mcr): Likewise.
483 (cortex_a8_neon_mcr_2_mcrr): Likewise.
484 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
486 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
488 (ca9_neon_mrrc): Likewise.
489 (cortex_a9_neon_int_1): Likewise.
490 (cortex_a9_neon_int_2): Likewise.
491 (cortex_a9_neon_int_3): Likewise.
492 (cortex_a9_neon_int_4): Likewise.
493 (cortex_a9_neon_int_5): Likewise.
494 (cortex_a9_neon_vqneg_vqabs): Likewise.
495 (cortex_a9_neon_vmov): Likewise.
496 (cortex_a9_neon_vaba): Likewise.
497 (cortex_a9_neon_vaba_qqq): Likewise.
498 (cortex_a9_neon_vsma): Likewise.
499 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
500 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
501 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
503 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
504 (cortex_a9_neon_mla_qqq_8_16): Likewise.
505 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
506 long_scalar_qdd_64_32_long): Likewise.
507 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
508 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
509 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
510 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
511 (cortex_a9_neon_shift_1): Likewise.
512 (cortex_a9_neon_shift_2): Likewise.
513 (cortex_a9_neon_shift_3): Likewise.
514 (cortex_a9_neon_vshl_ddd): Likewise.
515 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
516 (cortex_a9_neon_vsra_vrsra): Likewise.
517 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
518 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
519 (cortex_a9_neon_fp_vsum): Likewise.
520 (cortex_a9_neon_fp_vmul_ddd): Likewise.
521 (cortex_a9_neon_fp_vmul_qqd): Likewise.
522 (cortex_a9_neon_fp_vmla_ddd): Likewise.
523 (cortex_a9_neon_fp_vmla_qqq): Likewise.
524 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
525 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
526 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
527 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
528 (cortex_a9_neon_bp_simple): Likewise.
529 (cortex_a9_neon_bp_2cycle): Likewise.
530 (cortex_a9_neon_bp_3cycle): Likewise.
531 (cortex_a9_neon_ldr): Likewise.
532 (cortex_a9_neon_str): Likewise.
533 (cortex_a9_neon_vld1_1_2_regs): Likewise.
534 (cortex_a9_neon_vld1_3_4_regs): Likewise.
535 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
536 (cortex_a9_neon_vld2_4_regs): Likewise.
537 (cortex_a9_neon_vld3_vld4): Likewise.
538 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
539 (cortex_a9_neon_vst1_3_4_regs): Likewise.
540 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
541 (cortex_a9_neon_vst3_vst4): Likewise.
542 (cortex_a9_neon_vld1_vld2_lane): Likewise.
543 (cortex_a9_neon_vld3_vld4_lane): Likewise.
544 (cortex_a9_neon_vst1_vst2_lane): Likewise.
545 (cortex_a9_neon_vst3_vst4_lane): Likewise.
546 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
547 (cortex_a9_neon_mcr): Likewise.
548 (cortex_a9_neon_mcr_2_mcrr): Likewise.
549 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
550 (cortex_a9_fps): Likewise.
551 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
553 (cortex_m4_fmuls): Likewise.
554 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
556 (cortex_r4_mrc): Likewise.
557 * config/arm/iterators.md: Update comment referring to neon_type.
558 * config/arm/iwmmxt.md
559 (iwmmxt_arm_movdi): Update for attribute change.
560 (iwmmxt_movsi_insn): Likewise.
561 * config/arm/marvell-pj4.md
562 (pj4_vfp_to_core): Update for attribute change.
563 (pj4_core_to_vfp): Likewise.
564 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
566 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
567 (vfp_fstore): Likewise.
568 * doc/md.texi: Change references to neon_type to refer to type.
570 2013-09-04 Dodji Seketeli <dodji@redhat.com>
572 * tree.h (DECL_BUILT_IN): Fix typo in comment.
574 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
576 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
579 2013-09-04 Easwaran Raman <eraman@google.com>
582 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
583 (build_and_add_sum): Use it.
584 (appears_later_in_bb): Simplify code.
586 2013-09-04 Teresa Johnson <tejohnson@google.com>
588 * dumpfile.c (dump_finish): Don't close stderr/stdout.
590 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
592 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
594 2013-09-04 Jan Hubicka <jh@suse.cz>
596 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
597 * ipa-devirt.c: Include diganostic.h
598 (odr_type_d): Add types and types_set.
599 (hash_type_name): Work for types with vtables during LTO.
600 (odr_hasher::remove): Fix comment; destroy types_set.
601 (add_type_duplicate): New function,
602 (get_odr_type): Use it.
603 (dump_type_inheritance_graph): Dump type duplicates.
604 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
606 * tree.c (types_same_for_odr): Give exact answers on types with
609 2013-09-04 Dodji Seketeli <dodji@redhat.com>
611 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
612 explaining their differences.
614 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
616 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
618 2013-09-03 Jeff Law <law@redhat.com>
620 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
621 when not threading through a joiner block. Pass joiner/no joiner
622 state to register_jump_thread.
623 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
624 state from argument rather than implying on path length.
625 Dump the entire jump thread path into debugging dump.
626 * tree-flow.h (register_jump_thread): Update prototype.
628 2013-08-29 Xinliang David Li <davidxl@google.com>
630 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
631 Remove a trivial gcc_assert.
633 2013-08-29 Xinliang David Li <davidxl@google.com>
635 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
636 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
637 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
638 Delay base decl alignment adjustment.
639 * tree-vectorizer.c (vect_destroy_datarefs): New function.
640 * tree-vectorizer.h: New data structure.
641 (set_dr_misalignment): New function.
642 (dr_misalignment): Ditto.
643 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
644 (vectorizable_load): Ditto.
645 (ensure_base_align): New function.
646 (vectorize_loops): Add dbg_cnt support.
647 (execute_vect_slp): Ditto.
648 * dbgcnt.def: New debug counter.
649 * Makefile: New dependency.
651 2013-09-03 Meador Inge <meadori@codesourcery.com>
655 2013-08-30 Meador Inge <meadori@codesourcery.com>
657 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
659 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
661 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
664 2013-09-03 Richard Biener <rguenther@suse.de>
666 * tree-affine.c (add_elt_to_tree): Fix association issue,
667 avoid useless converts and make sure to always return a
668 properly typed result.
670 2013-09-03 Richard Biener <rguenther@suse.de>
673 * fold-const.c (negate_expr_p): Fix division case.
674 (negate_expr): Likewise.
676 2013-09-03 Richard Biener <rguenther@suse.de>
679 * tree-streamer-out.c: Include tm.h.
680 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
682 2013-09-03 Jan Hubicka <jh@suse.cz>
684 * tree-profile.c (tree_profiling): Cleanup CFG when done.
686 2013-09-03 Alan Modra <amodra@gmail.com>
688 * config.gcc (powerpc*-*-linux*): Add support for little-endian
689 multilibs to big-endian target and vice versa.
690 * config/rs6000/t-linux64: Use := assignment on all vars.
691 (MULTILIB_EXTRA_OPTS): Remove fPIC.
692 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
693 * config/rs6000/t-linux64le: New file.
694 * config/rs6000/t-linux64bele: New file.
695 * config/rs6000/t-linux64lebe: New file.
697 2013-09-02 Jan Hubicka <jh@suse.cz>
699 * ipa-inline-transform.c (inline_transform): Do not
700 optimize_inline_calls when not optimizing.
702 2013-09-02 Jan Hubicka <jh@suse.cz>
704 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
705 duplicated nodes for assembler names.
706 * symtab.c (symtab_unregister_node): Do not attempt to unlink
707 hard registers from assembler name hash.
709 2013-09-02 Jan Hubicka <jh@suse.cz>
711 * ipa-split.c (execute_split_functions): Split externally visible
712 functions called once.
714 2013-09-02 Martin Jambor <mjambor@suse.cz>
717 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
718 linked list. When finding the correct duplicate, also consider also
719 the caller in additon to its inlined_to node.
721 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
723 * config/aarch64/aarch64-simd-builtins.def
724 (dup_lane_scalar): Remove.
725 * config/aarch64/aarch64-simd.md
726 (aarch64_simd_dup): Add 'w->w' alternative.
727 (aarch64_dup_lane<mode>): Allow for VALL.
728 (aarch64_dup_lane_scalar<mode>): Remove.
729 (aarch64_dup_lane_<vswap_width_name><mode>): New.
730 (aarch64_get_lane_signed<mode>): Add w->w altenative.
731 (aarch64_get_lane_unsigned<mode>): Likewise.
732 (aarch64_get_lane<mode>): Likewise.
733 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
734 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
735 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
736 (VCON): Change container of V2SF.
737 (vswap_width_name): Likewise.
738 * config/aarch64/arm_neon.h
739 (__aarch64_vdup_lane_any): New.
740 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
741 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
742 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
744 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
747 * expr.c (emit_move_complex): Do not move complex FP values as parts if
748 the source or the destination is a single hard register.
750 2013-09-02 Richard Biener <rguenther@suse.de>
753 * tree-scalar-evolution.c (instantiate_scev_name): Allow
756 2013-09-02 Richard Biener <rguenther@suse.de>
758 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
759 arithmetic to sizetype.
761 2013-09-02 Bin Cheng <bin.cheng@arm.com>
763 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
764 Find auto-increment use both before and after candidate.
766 2013-09-02 Marek Polacek <polacek@redhat.com>
768 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
770 2013-09-01 Jan Hubicka <jh@suse.cz>
772 * Makefile.in: Add ipa-profile.o
773 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
774 * cgraph.c (struct cgraph_propagate_frequency_data,
775 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
776 ipa-profile.c; replace cgraph_ by ipa_ prefix.
777 * cgraph.h (cgraph_propagate_frequency): Remove.
778 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
779 (inline_update_callee_summaries): Update.
780 * ipa-profile.c: New file.
781 * ipa-utils.h (ipa_propagate_frequency): Declare.
782 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
783 data-streamer.h, value-prof.h
784 (symtab_remove_unreachable_nodes): Update profile.
785 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
786 account_time_size, cmp_counts, dump_histogram,
787 ipa_profile_generate_summary, ipa_profile_write_summary,
788 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
789 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
790 Move to ipa-profile.c
792 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
794 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
796 2013-09-01 Jan Hubicka <jh@suse.cz>
798 * common.opt (fdevirtualize-speculatively): New function.
799 * invoke.texi (fdevirtualize-speculatively): Document.
800 * ipa-devirt.c: Include ipa-inline.h
801 (likely_target_p): New function.
802 (ipa_devirt): New function.
803 (gate_ipa_devirt): New function.
804 (pass_data_ipa_devirt): New static var.
805 (pass_ipa_devirt): Likewise.
806 (make_pass_ipa_devirt): New function.
807 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
808 (common_handle_option): Disable devirtualization when
809 value range profiling is available.
810 * passes.def (pass_ipa_devirt): Add.
811 * timever.def (TV_IPA_DEVIRT): New timevar.
812 * tree-pass.h (make_pass_ipa_devirt):
814 2013-09-01 Iain Sandoe <iain@codesourcery.com>
816 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
817 include sanitize(undefined).
819 2013-08-31 Diego Novillo <dnovillo@google.com>
821 * Makefile.in (TREE_CORE_H): Define.
823 (GTFILES): Add tree-core.h.
824 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
826 * tree-core.h: New file.
827 Move all data structures, enum, typedefs, global
828 declarations and constants from ...
831 2013-08-31 Jan Hubicka <jh@suse.cz>
833 * bulitins.c (expand_builtin): Do not early exit for gcov
834 instrumented functions.
836 2013-08-31 Marek Polacek <polacek@redhat.com>
838 * ubsan.c: Include tm_p.h.
840 2013-08-31 Jan Hubicka <jh@suse.cz>
842 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
845 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
846 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
847 * tree-cfg.c (verify_gimple_label): ... here.
848 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
849 (ipa_merge_profiles): New function.
850 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
851 (lto_input_function_body): Likewise.
852 * ipa-utils.h (ipa_merge_profiles): Declare.
853 * lto-streamer.h (lto_input_function_body): Update prototype.
854 (emit_label_in_global_context_p): Remove.
855 * lto-symtab.c: Include ipa-utils.h
856 (lto_cgraph_replace_node): Use ipa_merge_profiles.
858 2013-08-31 Jan Hubicka <jh@suse.cz>
860 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
862 2013-08-31 Jan Hubicka <jh@suse.cz>
864 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
866 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
868 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
869 "cmp" RTX before signed_comparison_operator check to account
872 2013-08-30 Jan Hubicka <jh@suse.cz>
874 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
875 (detect_type_change_1): Rename to ...
876 (detect_type_change): ... this one; early return on non-polymorphic
878 (detect_type_change_ssa): Add comp_type parameter; update
879 use of detect_type_change.
880 (compute_complex_assign_jump_func): Add param_type parameter;
881 update use of detect_type_change_ssa.
882 (compute_complex_ancestor_jump_func): Likewise.
883 (ipa_get_callee_param_type): New function.
884 (ipa_compute_jump_functions_for_edge): Compute parameter type;
885 update calls to the jump function computation functions.
887 2013-08-30 Teresa Johnson <tejohnson@google.com>
888 Steven Bosscher <steven@gcc.gnu.org>
890 * cfgrtl.c (fixup_new_cold_bb): New routine.
891 (commit_edge_insertions): Invoke fixup_partitions.
892 (find_partition_fixes): New routine.
893 (fixup_partitions): Ditto.
894 (verify_hot_cold_block_grouping): Update comments.
895 (rtl_verify_edges): Invoke find_partition_fixes.
896 (rtl_verify_bb_pointers): Update comments.
897 (rtl_verify_bb_layout): Ditto.
898 * basic-block.h (probably_never_executed_edge_p): Declare.
899 (fixup_partitions): Ditto.
900 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
901 * bb-reorder.c (sanitize_hot_paths): New function.
902 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
904 * predict.c (probably_never_executed_edge_p): New routine.
905 * cfg.c (check_bb_profile): Add partition insanity warnings.
907 2013-08-30 Meador Inge <meadori@codesourcery.com>
909 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
911 2013-08-30 Marek Polacek <polacek@redhat.com>
913 * Makefile.in (ubsan.o): Add.
914 (c-family/c-ubsan.o): Add.
915 (builtins.o): Add ubsan.h dependency.
918 * common.opt: Add -fsanitize=undefined option.
919 (flag_sanitize): Add variable.
920 (fsanitize=): Add option. Add Driver.
921 (fsanitize=thread): Remove option.
922 (fsanitize=address): Likewise.
923 (static-libubsan): New option.
924 * doc/invoke.texi: Document the new flag and -static-libubsan.
925 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
926 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
927 * builtin-attrs.def (ATTR_COLD): Define.
928 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
929 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
930 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
931 * flag-types.h (sanitize_code): New enum.
932 * opts.c (common_handle_option): Parse command line arguments
933 of -fsanitize=. Add -fsanitize=unreachable option.
934 * varasm.c (get_variable_section): Adjust.
935 (assemble_noswitch_variable): Likewise.
936 (assemble_variable): Likewise.
937 (output_constant_def_contents): Likewise.
938 (categorize_decl_for_section): Likewise.
939 (place_block_symbol): Likewise.
940 (output_object_block): Likewise.
941 * builtins.def: Likewise.
942 * toplev.c (compile_file): Likewise.
943 (process_options): Likewise.
944 * cppbuiltin.c: Likewise.
945 * tsan.c (tsan_pass): Likewise.
946 (tsan_gate): Likewise.
947 (tsan_gate_O0): Likewise.
948 * cfgexpand.c (partition_stack_vars): Likewise.
949 (expand_stack_vars): Likewise.
950 (defer_stack_allocation): Likewise.
951 (expand_used_vars): Likewise.
952 * cfgcleanup.c (old_insns_match_p): Likewise.
953 * asan.c (asan_finish_file): Likewise.
954 (asan_instrument): Likewise.
955 (gate_asan): Likewise.
956 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
957 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
958 (asan_global_struct): Use pointer_sized_int_node instead
959 calling build_nonstandard_integer_type.
960 (initialize_sanitizer_builtins): Likewise.
961 (asan_finish_file): Likewise.
962 * gcc.c: Document %{%:function(args):X}.
963 (static_spec_functions): Add sanitize.
964 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
965 store funcval != NULL there.
966 (do_spec_1): Adjust handle_spec_function caller.
967 (handle_braces): Allow %:function(args) as condition.
968 (sanitize_spec_function): New function.
969 (ADD_STATIC_LIBUBSAN_LIBS): Define.
970 (LIBUBSAN_SPEC): Likewise.
971 (LIBUBSAN_EARLY_SPEC): Likewise.
972 (SANITIZER_SPEC): Handle libubsan.
973 (SANITIZER_EARLY_SPEC): Likewise.
974 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
975 instead of fsanitize=address.
976 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
977 instead of fsanitize=address*.
978 * builtins.c: Include ubsan.h.
979 (fold_builtin_0): Instrument __builtin_unreachable.
980 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
981 instead of flag_asan.
982 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
983 (pointer_sized_int_node): Define.
984 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
986 2013-08-30 Mike Stump <mikestump@comcast.net>
988 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
991 2013-08-29 Jan Hubicka <jh@suse.cz>
993 * cgraph.c (cgraph_function_body_availability): Handle weakref
995 * passes.def: Remove pass_fixup_cfg.
996 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
997 track when we need to remove functions.
998 (gate_ipa_inline): Execute inlining always; add comment why.
999 (pass_data_ipa_inline): Remove TODO_remove_functions.
1000 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
1001 do not produce summaries.
1002 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
1003 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
1004 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
1005 constant pool and vtable.
1007 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
1009 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
1010 New arm_neon.h's internal macros to specify 64-bit constants.
1011 Avoid using stdint.h's macros.
1013 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
1015 * recog.c (verify_changes): Verify that changes[i].old is non-zero
1016 before applying REG_P.
1018 2013-08-30 Jakub Jelinek <jakub@redhat.com>
1020 PR tree-optimization/58277
1021 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
1022 after seeing too many stmts with vdef in between dombb and current
1023 bb, invalidate everything.
1025 2013-08-30 Richard Biener <rguenther@suse.de>
1027 * fold-const.c (fold_single_bit_test): Fix overflow test.
1029 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
1031 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
1032 and which can live in a register, always retrieve the value on entry.
1033 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
1034 passed by invisible reference specially.
1035 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
1036 (vt_add_function_parameter): Correctly deal with a parameter passed by
1037 invisible reference.
1039 2013-08-30 Jan Hubicka <jh@suse.cz>
1041 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
1043 2013-08-30 Richard Biener <rguenther@suse.de>
1045 PR tree-optimization/58228
1046 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
1047 allow invariant loads in nested loop vectorization.
1049 2013-08-30 Richard Biener <rguenther@suse.de>
1051 PR tree-optimization/58223
1052 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
1053 (has_anti_or_output_dependence): ... this and adjust to also
1054 look for output dependences.
1055 (mark_nodes_having_upstream_mem_writes): Adjust.
1056 (rdg_flag_uses): Likewise.
1058 2013-08-30 Richard Biener <rguenther@suse.de>
1060 PR tree-optimization/58010
1061 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
1062 assert that we have a loop-closed PHI.
1064 2013-08-29 Jan Hubicka <jh@suse.cz>
1066 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
1067 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
1068 * lto-section-in.c (lto_free_function_in_decl_state): New function.
1069 (lto_free_function_in_decl_state_for_node): New function.
1071 2013-08-29 Xinliang David Li <davidxl@google.com>
1073 * loop-unroll.c (report_unroll_peel): Minor message change.
1074 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
1075 Emit alignment peeling message with default -fopt-info.
1076 (vect_loop_versioning): Emit loop version info message.
1077 * tree-vectorizer.c (vectorize_loops): Minor message change.
1078 (execute_vect_slp): Ditto.
1080 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
1082 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
1083 of the clone from the DECL_NAME of the original function.
1085 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
1087 * passes.c (register_pass): Add overload.
1088 * tree-pass.h (register_pass): Forward declare it. Add comment.
1090 2013-08-29 Jan Hubicka <jh@suse.cz>
1092 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
1093 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
1094 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
1095 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1096 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
1097 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
1098 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1099 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
1101 2013-08-29 Teresa Johnson <tejohnson@google.com>
1103 * dumpfile.c (dump_loc): Output column number.
1104 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
1105 * doc/invoke.texi: Document optall -fopt-info flag.
1106 * profile.c (read_profile_edge_counts): Use new dump framework.
1107 (compute_branch_probabilities): Ditto.
1108 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
1109 when pass not in any opt group.
1110 * pass_manager.h (pass_manager::get_pass_profile): New method.
1111 * value-prof.c (check_counter): Use new dump framework.
1112 (check_ic_target): Ditto.
1113 * coverage.c (get_coverage_counts): Ditto.
1114 (coverage_init): Setup new dump framework.
1116 2013-08-29 Richard Biener <rguenther@suse.de>
1118 PR tree-optimization/58246
1119 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
1120 handle the dominance check inside a basic-block.
1122 2013-08-29 Richard Biener <rguenther@suse.de>
1125 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
1126 of default defs that appear in abnormal PHI nodes.
1128 2013-08-29 Richard Biener <rguenther@suse.de>
1130 PR tree-optimization/57685
1131 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
1132 single-use operands to avoid exponential complexity.
1134 2013-08-28 Dehao Chen <dehao@google.com>
1136 * ipa-inline.c (edge_badness): Fix integer underflow.
1138 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
1140 * gtm-builtins.def (_ITM_free): Declare leaf.
1142 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1145 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
1146 (*tls_local_dynamic_base_64_largepic): Likewise.
1147 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
1148 Remove predicate from call operand.
1149 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
1150 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
1152 2013-08-28 Jeff Law <law@redhat.com>
1154 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
1155 checks for the number of predecessors and successors allowed.
1156 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
1157 which require copying a joiner block if there is a request which
1158 is a subpath that requires no joiner block copying.
1160 2013-08-28 Jan Hubicka <jh@suse.cz>
1162 * lto-streamer-out.c (DFS_write_tree_body): Drop
1163 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
1164 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
1165 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
1166 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
1167 Do not read DECL_ERROR_ISSUED.
1168 (unpack_ts_decl_with_vis_value_fields): Do not read
1170 (lto_input_ts_binfo_tree_pointers): Do not read
1171 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
1172 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
1173 write DECL_ERROR_ISSUED..
1174 (pack_ts_decl_with_vis_value_fields): Do not write
1176 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
1177 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
1178 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
1179 * tree.h (tree_decl_common): Update comment.
1180 (DECL_ERROR_ISSUED): Remove.
1182 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1185 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
1187 2013-08-28 Jan Hubicka <jh@suse.cz>
1189 * builtins.def (free): Declare leaf.
1191 2013-08-27 David Malcolm <dmalcolm@redhat.com>
1194 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
1195 * configure: Regenerate.
1197 2013-08-27 Martin Jambor <mjambor@suse.cz>
1199 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
1200 (ipa_ancestor_jf_data): Likewise.
1201 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
1202 (ipa_get_jf_pass_through_type_preserved): New function.
1203 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
1204 (ipa_get_jf_ancestor_type_preserved): New function.
1205 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
1206 (ipa_get_jf_ancestor_result): Likewise.
1207 (propagate_vals_accross_pass_through): Use
1208 ipa_get_jf_pass_through_result to do all the value mappings.
1209 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
1210 type_preserved flag.
1211 (ipa_set_jf_cst_copy): New function.
1212 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
1213 (ipa_set_jf_arith_pass_through): Likewise.
1214 (ipa_set_ancestor_jf): Likewise.
1215 (compute_complex_assign_jump_func): Set type_preserved instead of
1217 (ipa_compute_jump_functions_for_edge): Likewise.
1218 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
1219 (update_jump_functions_after_inlining): Update type_preserved.
1220 Explicitely create jump functions when combining one with pass_through.
1221 (ipa_write_jump_function): Stream the type_preserved flags.
1222 (ipa_read_jump_function): Likewise.
1224 2013-08-27 Jakub Jelinek <jakub@redhat.com>
1225 Aldy Hernandez <aldyh@redhat.com>
1227 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
1228 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
1229 * function.h (struct function): Add has_force_vect_loops and
1231 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
1232 * gimple.c (gimple_build_omp_critical): Add KIND argument and
1234 * gimple.def: Update CLAUSES comments.
1235 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
1236 (gimple_build_omp_for): Add argument to prototype.
1237 (gimple_omp_for_kind): New.
1238 (gimple_omp_for_set_kind): New.
1239 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
1240 GOVD_DATA_SHARE_CLASS.
1241 (enum omp_region_type): Add ORT_SIMD.
1242 (gimple_add_tmp_var): Handle ORT_SIMD.
1243 (gimplify_var_or_parm_decl): Same.
1244 (is_gimple_stmt): Same.
1245 (omp_firstprivatize_variable): Same.
1246 (omp_add_variable): Only use splay_tree_insert if lookup failed.
1247 (omp_notice_variable): Handle ORT_SIMD.
1248 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
1249 (omp_check_private): Handle ORT_SIMD.
1250 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1252 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
1253 Handle OMP_CLAUSE_LASTPRIVATE.
1254 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1256 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
1257 (gimplify_expr): Handle OMP_SIMD.
1258 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
1259 (expand_GOMP_SIMD_VF): New.
1260 (expand_GOMP_SIMD_LAST_LANE): New.
1261 * internal-fn.def (GOMP_SIMD_LANE): New.
1262 (GOMP_SIMD_VF): New.
1263 (GOMP_SIMD_LAST_LANE): New.
1264 * omp-low.c: Include target.h.
1265 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
1267 (check_omp_nesting_restrictions): Same.
1269 (lower_rec_simd_input_clauses): New.
1270 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
1272 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
1273 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
1274 (expand_omp_build_assign): New.
1275 (expand_omp_for_init_counts): New.
1276 (expand_omp_for_init_vars): New.
1277 (extract_omp_for_update_vars): New.
1278 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
1279 and rewrite accordingly.
1280 (expand_omp_simd): New.
1281 (expand_omp_for): Use expand_omp_simd.
1282 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
1283 (lower_omp_for): Do not lower the body.
1284 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
1286 * tree-flow.h (find_omp_clause): Remove prototype.
1287 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
1288 forcing vectorization of the loop, or if flag_tree_vectorize.
1289 (gate_tree_if_conversion): Similarly.
1290 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
1291 gimple_build_omp_for.
1292 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
1293 * tree-parloops (create_parallel_loop): Pass kind argument to
1294 gimple_build_omp_for.
1295 * tree-pretty-print.c (dump_omp_clause): Add cases for
1296 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
1297 OMP_CLAUSE__SIMDUID_.
1298 (dump_generic_node): Handle OMP_SIMD.
1299 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
1300 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
1301 unroll OMP_SIMD loops here.
1302 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
1303 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
1305 (vect_analyze_data_refs): Handle simd loops.
1306 * tree-vect-loop.c (vectorizable_live_operation): Handle
1308 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
1309 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
1310 (vectorizable_load): Same.
1311 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
1312 (struct simduid_to_vf): New.
1313 (simduid_to_vf::hash): New.
1314 (simduid_to-vf::equal): New.
1315 (struct simd_array_to_simduid): New.
1316 (simd_array_to_simduid::hash): New.
1317 (simd_array_to_simduid::equal): New.
1318 (adjust_simduid_builtins): New.
1319 (struct note_simd_array_uses_struct): New.
1320 (note_simd_array_uses_cb): New.
1321 (note_simd_array_uses): New.
1322 (vectorize_loops): Handle simd hints and adjust simd builtins
1324 * tree-vectorizer.h (struct _stmt_vec_info): Add
1325 simd_lane_access_p field.
1326 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
1327 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
1328 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
1329 (omp_clause_code_name): Same.
1330 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
1331 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
1332 * tree.def (OMP_SIMD): New entry.
1333 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
1334 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
1335 (OMP_CLAUSE_DECL): Adjust range for new clauses.
1336 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
1337 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
1338 (OMP_CLAUSE_LINEAR_STEP): New.
1339 (OMP_CLAUSE_SAFELEN_EXPR): New.
1340 (OMP_CLAUSE__SIMDUID__DECL): New.
1341 (find_omp_clause): New prototype.
1343 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
1345 * config/i386/driver-i386.c (host_detect_local_cpu): Update
1346 Haswell processor detection.
1348 2013-08-27 Christian Widmer <shadow@umbrox.de>
1351 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
1352 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
1353 AVX2 capable processors.
1355 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
1357 * config/aarch64/arm_neon.h: Replace all inline asm implementations
1358 of vget_low_* with implementations in terms of other intrinsics.
1360 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
1363 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
1364 values to -1, 0 and 1.
1366 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
1368 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
1369 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
1370 (<optab><mode>3_insn): Remove.
1371 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
1372 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
1373 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
1374 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
1375 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
1376 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
1377 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
1378 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
1379 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
1380 (ror<mode>3_insn): Likewise.
1381 * config/aarch64/predicates.md (aarch64_simd_register): New.
1383 2013-08-27 Richard Biener <rguenther@suse.de>
1385 PR tree-optimization/57521
1386 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
1387 one edge is non-critical.
1388 (find_phi_replacement_condition): Make sure to use a non-critical
1389 edge. Cleanup and remove old bug workarounds.
1390 (bb_postdominates_preds): Remove.
1391 (if_convertible_loop_p_1): Do not compute post-dominators.
1392 (combine_blocks): Do not free post-dominators.
1393 (main_tree_if_conversion): Likewise.
1394 (pass_data_if_conversion): Add TODO_verify_ssa.
1396 2013-08-27 DJ Delorie <dj@redhat.com>
1398 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
1400 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
1402 * function.c (assign_parm_find_data_types): Set passed_mode and
1403 nominal_mode to the TYPE_MODE of nominal_type for the built
1404 pointer type in case of the struct-pass-by-reference.
1406 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1408 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
1409 (UINT16_TYPE): Change default to "unsigned int".
1411 * config/avr/avr.opt (mfract-convert-truncate): New option.
1412 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
1413 is set, round negative fractional integers according to n1169
1414 when converting to integer types.
1416 2013-08-26 Jan Hubicka <jh@suse.cz>
1418 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
1419 methods can not be called indirectly when their address is not taken.
1421 2013-08-26 Jan Hubicka <jh@suse.cz>
1423 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
1426 2013-08-26 Jan Hubicka <jh@suse.cz>
1428 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
1431 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1433 * reload.c (find_valid_class): Allow classes that do not include
1434 FIRST_PSEUDO_REGISTER - 1.
1436 2013-08-26 Jan Hubicka <jh@suse.cz>
1438 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
1439 fix edge count/frequency when speculation failed; fix type check
1440 for the direct call.
1442 2013-08-26 Jan Hubicka <jh@suse.cz>
1444 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
1446 2013-08-26 Jan Hubicka <jh@suse.cz>
1448 * ipa-inline-transform.c (inline_transform): Be ready for basic block
1449 to be changed by edge redirection.
1451 2013-08-26 Jan Hubicka <jh@suse.cz>
1453 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
1454 formating; add sanity check.
1455 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
1456 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
1457 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
1458 (resolve_noninline_speculation): Update callee keys, too.
1460 2013-08-26 Jan Hubicka <jh@suse.cz>
1462 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
1463 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
1465 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1467 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
1470 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
1472 * config/i386/i386.c (ix86_debug_options): Remove prototype.
1473 (x86_64_elf_select_section): Ditto.
1474 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
1476 (ix86_pass_by_reference): Ditto.
1477 (output_set_got): Ditto.
1478 (ix86_unary_operator_ok): Ditto.
1479 (ix86_expand_builtin): Ditto.
1481 2013-08-23 Jan Hubicka <jh@suse.cz>
1483 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
1485 2013-08-23 Jan Hubicka <jh@suse.cz>
1487 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
1488 (tree_decl_with_vis): Add FINAL field.
1490 2013-08-23 Jeff Law <law@redhat.com>
1492 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
1493 the debugging dump when the expression is fully redundant.
1495 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1497 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
1498 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
1499 * pretty-print.c (pp_formatted_text_data): Likewise.
1500 (pp_write_text_to_stream): Likewise.
1501 (pp_write_text_as_dot_label_to_stream): Likewise.
1502 (pp_append_r): Likewise.
1503 (pp_format): Likewise.
1504 (pp_flush): Likewise.
1505 (pp_clear_output_area): Likewise.
1506 (pp_append_text): Likewise.
1507 (pp_formatted_text): Likewise.
1508 (pp_remaining_character_count_for_line): Likewise.
1509 (pp_newline): Likewise.
1510 (pp_character): Likewise.
1511 (output_buffer::~output_buffer): Define.
1512 (pretty_printer::~pretty_printer): Destruct output buffer.
1513 * pretty-print.h (output_buffer::~output_buffer): Declare.
1514 (pretty_printer::~pretty_printer): Declare virtual.
1516 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
1519 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
1520 HOST_WIDE_INT_M1U): New macros.
1521 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
1522 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
1523 unsigned -1 for lshift.
1524 * cse.c (cse_insn): Likewise.
1525 * double-int.c (rshift_double, lshift_double): Likewise.
1526 * builtins.c (fold_builtin_bitop): Likewise.
1527 * combine.c (force_to_mode): Likewise.
1528 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
1529 * simplify-rtx.c (simplify_const_unary_operation,
1530 simplify_const_binary_operation): Likewise.
1531 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
1532 check_va_list_escapes): Likewise.
1533 * rtlanal.c (nonzero_bits1): Likewise.
1534 * expmed.c (expand_smod_pow2): Likewise.
1535 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
1537 2013-08-23 Jan Hubicka <jh@suse.cz>
1539 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
1540 as having address taken.
1542 2013-08-23 Jan Hubicka <jh@suse.cz>
1544 * ipa-utils.h (method_class_type): Declare.
1545 * ipa-devirt.c (method_class_type): Export.
1547 * cgraphunit.c (analyze_functions): Do basic devirtualization;
1548 do not walk base classes of anonymous types.
1550 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
1552 PR rtl-optimization/58220
1554 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
1555 handle SEQUENCE insns properly.
1557 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1559 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
1561 (pp_newline_and_indent): Likewise.
1562 (pp_separate_with): Likewise.
1563 * pretty-print.c (pp_newline_and_flush): Define.
1564 (pp_newline_and_indent): Likewise.
1565 (pp_separate_with): Likewise.
1567 2013-08-23 Jakub Jelinek <jakub@redhat.com>
1570 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
1571 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
1573 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
1575 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
1576 * gcc/config/i386/i386.md (*movti_internal): Use
1577 predicate to determine if EVEX is needed.
1578 (*movsi_internal): Ditto.
1579 (*movdf_internal): Ditto.
1580 (*movsf_internal): Ditto.
1581 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
1583 2013-08-23 Jakub Jelinek <jakub@redhat.com>
1585 PR tree-optimization/58209
1586 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
1587 (find_tail_calls): Give up for pointer result types if m is non-NULL.
1588 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
1589 emit POINTER_PLUS_EXPR.
1590 (create_tailcall_accumulator): For pointer result type accumulate in
1593 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
1595 * configure.ac: Add backslashes missing from the last change.
1596 * configure: Regenerate.
1598 2013-08-22 Jan Hubicka <jh@susue.cz>
1600 * ipa.c (function_and_variable_visibility): First remember function
1601 was global and then make it local.
1603 2013-08-22 Julian Brown <julian@codesourcery.com>
1605 * configure.ac: Add aarch64 to list of arches which use "nop" in
1607 * configure: Regenerate.
1609 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1611 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
1612 gnu_libc_has_function.
1613 * config/s390/tpf.h: Likewise.
1615 2013-08-22 Jan Hubicka <jh@susue.cz>
1617 * timevar.c (validate_phases): Add cast.
1619 2013-08-22 Jan Hubicka <jh@susue.cz>
1621 * timevar.c (validate_phases): Use size_t for memory.
1622 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
1624 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
1626 * pretty-print.h (output_buffer::output_buffer): Declare.
1627 (pretty_printer::pretty_printer): Likewise.
1628 (pp_construct): Remove.
1629 * pretty-print.c (output_buffer::output_buffer): Define.
1630 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
1631 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
1632 (print_gimple_expr): Likewise.
1633 (print_gimple_seq): Likewise.
1634 (gimple_dump_bb): Likewise.
1635 * sched-vis.c (dump_value_slim): Likewise.
1636 (dump_insn_slim): Likewise.
1637 (dump_rtl_slim): Likewise.
1638 (str_pattern_slim): Likewise.
1639 * tree-mudflap.c (mf_varname_tree): Likewise.
1640 * graph.c (print_graph_cfg): Likewise.
1641 (start_graph_dump): Likewise.
1642 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
1644 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
1645 pretty printer initialization.
1646 * coretypes.h (diagnostic_context): Remove superflous type alias
1648 (pretty_printer): Likewise. Declare directly as a class.
1649 (pretty_print_info): Remove declaration as class.
1650 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
1651 and pp_clear_output_area.
1652 (asan_add_global): Likewise.
1654 2013-08-22 Jan Hubicka <jh@suse.cz>
1656 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
1657 * ipa-utils.h (update_type_inheritance_graph): Declare.
1658 (possible_polymorphic_call_target_p): Declare.
1659 (possible_polymorphic_call_target_p): New.
1660 * ipa-devirt.c: Update toplevel comments.
1661 (cached_polymorphic_call_targets): Move up.
1662 (odr_type_d): Move ID down.
1663 (polymorphic_type_binfo_p): Update comment.
1664 (odr_hasher::remove): Likewise;
1665 (get_odr_type): Set anonymous_namespace.
1666 (dump_odr_type): Dump it.
1667 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1668 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1669 (record_binfo): Add comment.
1670 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1672 (devirt_node_removal_hook): Do not iCE when cache is freed.
1673 (possible_polymorphic_call_target_p): New predicate.
1674 (update_type_inheritance_graph): New function.
1676 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1677 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1678 Sergey Lega <sergey.s.lega@intel.com>
1679 Anna Tikhonova <anna.tikhonova@intel.com>
1680 Ilya Tocar <ilya.tocar@intel.com>
1681 Andrey Turetskiy <andrey.turetskiy@intel.com>
1682 Ilya Verbin <ilya.verbin@intel.com>
1683 Kirill Yukhin <kirill.yukhin@intel.com>
1684 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1686 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1687 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1688 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1689 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1690 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1691 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1692 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1693 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1694 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1695 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1696 OPT_mavx512pf, OPT_mavx512er cases.
1697 * config/i386/constraints.md (v): New constraint.
1698 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1699 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1700 (bit_AVX512CD): New.
1701 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1702 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1703 * config/i386/i386-c.c (ix86_target_macros_internal):
1704 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1706 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1707 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1708 * config/i386/i386.c (regclass_map, dbx_register_map)
1709 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1710 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1711 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1713 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1714 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1715 -mavx512pf options. Fix formatting.
1716 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1717 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1718 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1719 -mavx512cd, -mavx512pf options.
1720 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1721 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1722 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1724 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1725 (ix86_set_reg_reg_cost): Ditto.
1726 (x86_order_regs_for_local_alloc): Ditto.
1727 (MAX_VECT_LEN): Extend to 64-byte.
1728 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1729 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1730 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1731 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1732 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1733 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1734 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1735 (SSE_REG_MODE_P): Support new modes.
1736 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1737 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1738 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1739 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1740 (REG_CLASS_CONTENTS): Add new registers.
1741 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1742 (EXT_REX_SSE_REGNO_P): New.
1743 (HI_REGISTER_NAMES): Add new registers.
1744 * config/i386/i386.md: Define constants for new registers.
1745 (mode): Add new 512-bit modes.
1746 (prefix): Support evex prefix.
1747 (isa): Support avx512f, noavx512f, fma_avx512f.
1748 (ssemodesuffix): Add new 512-bit modes.
1750 (*movxi_internal_avx512f): Ditto.
1751 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1753 (*movsi_internal): Likewise.
1754 (*movdf_internal): Likewise.
1755 (*movsf_internal): Likewise.
1756 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1757 (<code><mode>3): Likewise.
1758 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1760 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1761 with the new constraint "v".
1762 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1764 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1765 with the new constraint "v".
1766 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1767 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1768 (*<plusminus_insn><mode>3): Likewise.
1769 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1770 (*mul<mode>3): Likewise.
1771 (<sse>_vmmul<mode>3): Likewise.
1772 (<sse>_div<mode>3): Likewise.
1773 (<sse>_vmdiv<mode>3): Likewise.
1774 (<sse>_sqrt<mode>2): Likewise.
1775 (<sse>_vmsqrt<mode>2): Likewise.
1776 (*<code><mode>3_finite): Likewise.
1777 (*<code><mode>3) <smaxmin>: Likewise.
1778 (<sse>_vm<code><mode>3): Likewise.
1779 (*<code><mode>3) <any_logic>: Likewise.
1780 (*fma_fmadd_<mode>): Likewise.
1781 (*fma_fmsub_<mode>): Likewise.
1782 (*fma_fnmadd_<mode>): Likewise.
1783 (*fma_fnmsub_<mode>): Likewise.
1784 (*fma_fmaddsub_<mode>): Likewise.
1785 (*fma_fmsubadd_<mode>): Likewise.
1786 (*fmai_fmadd_<mode>): Likewise.
1787 (*fmai_fmsub_<mode>): Likewise.
1788 (*fmai_fnmadd_<mode>): Likewise.
1789 (*fmai_fnmsub_<mode>): Likewise.
1790 (sse_cvtsi2ss): Likewise.
1791 (sse_cvtsi2ssq): Likewise.
1792 (sse_cvtss2si): Likewise.
1793 (sse_cvtss2si_2): Likewise.
1794 (sse_cvtss2siq): Likewise.
1795 (sse_cvtss2siq_2): Likewise.
1796 (sse_cvttss2si): Likewise.
1797 (sse_cvtss2siq_2): Likewise.
1798 (float<sseintvecmodelower><mode>2): Likewise.
1799 (sse2_cvtsd2si_2): Likewise.
1800 (sse2_cvtsd2siq_2): Likewise.
1801 (*<plusminus_insn><mode>3): Likewise.
1802 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1803 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1804 (ashr<mode>3): Likewise.
1805 (<shift_insn><mode>3): Likewise.
1806 (avx2_<code><mode>3): Likewise.
1807 (*avx2_<code><mode>3): Likewise.
1808 (*andnot<mode>3): Likewise.
1809 (*<code><mode>3) <any_logic>: Likewise.
1810 (abs<mode>2): Likewise.
1811 (avx2_permvar<mode>): Likewise.
1812 (avx2_perm<mode>_1): Likewise.
1813 (*avx_vpermilp<mode>): Likewise.
1814 (avx_vpermilvar<mode>3): Likewise.
1815 (avx2_ashrv<mode>): Likewise.
1816 (avx2_<shift_insn>v<mode>): Likewise.
1817 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1819 * doc/rtl.texi: Document XImode.
1821 2013-08-21 Jeff Law <law@redhat.com>
1823 * tree-flow.h (register_jump_thread): Pass vector of edges
1824 instead of each important edge.
1825 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1826 thread path into a vector and pass that to register_jump_thread.
1827 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1828 passed in edge vector to the current 3-edge form.
1831 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1833 * dce.c (fini_dce): Call df_analyze again just in case
1834 delete_unmarked_insns removed anything.
1836 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1838 * reload.h (struct reg_equivs): Rename to ..
1839 (struct reg_equivs_s): .. this.
1841 2013-08-20 Martin Liska <marxin.liska@gmail.com>
1843 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
1845 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1847 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
1849 2013-08-21 Jeff Law <law@redhat.com>
1851 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
1852 simplify assignments too. If the RHS collapses to a singleton
1853 range, then return the value for the range.
1855 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
1857 * config/i386/sse.md (V16): Rename to...
1859 (mov<mode>): Update iterator name.
1860 (*mov<mode>_internal): Ditto.
1861 (push<mode>1): Ditto.
1862 (movmisalign<mode>): Ditto.
1864 2013-08-20 Jan Hubicka <jh@suse.cz>
1867 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
1868 entry for direct edges.
1869 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
1871 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1873 Revert my last two changes, r201865 and r201864:
1876 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1878 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1879 instances can own GC refs.
1881 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1882 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1883 (gcc::context::gt_pch_nx): Likewise.
1884 (gcc::context::gt_pch_nx): Likewise.
1885 * ggc.h (gt_ggc_mx <T>): New.
1886 (gt_pch_nx_with_op <T>): New.
1887 (gt_pch_nx <T>): New.
1888 * passes.c (opt_pass::gt_ggc_mx): New.
1889 (opt_pass::gt_pch_nx): New.
1890 (opt_pass::gt_pch_nx_with_op): New.
1891 (pass_manager::gt_ggc_mx): New.
1892 (pass_manager::gt_pch_nx): New.
1893 (pass_manager::gt_pch_nx_with_op): New.
1894 (pass_manager::operator new): Use
1895 ggc_internal_cleared_alloc_stat rather than xcalloc.
1896 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1897 (pass_manager::gt_ggc_mx): New.
1898 (pass_manager::gt_pch_nx): New.
1899 (pass_manager::gt_pch_nx_with_op): New.
1900 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1901 (opt_pass::operator new): New.
1902 (opt_pass::gt_ggc_mx): New.
1903 (opt_pass::gt_pch_nx): New.
1904 (opt_pass::gt_pch_nx_with_op): New.
1907 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1909 * Makefile.in (GTFILES): Add context.h.
1910 * context.c (gcc::context::operator new): New.
1911 (gcc::context::gt_ggc_mx): New.
1912 (gcc::context::gt_pch_nx): New.
1913 (gcc::context::gt_pch_nx): New.
1914 * context.h (gcc::context): Add GTY((user)) marking.
1915 (gcc::context::operator new): New.
1916 (gcc::context::gt_ggc_mx): New.
1917 (gcc::context::gt_pch_nx): New.
1918 (gcc::context::gt_pch_nx): New.
1919 (g): Add GTY marking.
1920 (gt_ggc_mx (gcc::context *)): New.
1921 (gt_pch_nx (gcc::context *)): New.
1922 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1923 void *cookie)): New.
1924 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1926 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1928 * dce.c (fini_dce): Call df_analyze again just in case
1929 delete_unmarked_insns removed anything.
1931 2013-08-20 Teresa Johnson <tejohnson@google.com>
1933 PR rtl-optimizations/57451
1934 * final.c (reemit_insn_block_notes): Prevent lexical blocks
1935 from crossing split section boundaries.
1937 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1939 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
1940 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
1941 with MULTLIB_DEFAULTS.
1943 2013-08-20 Nick Clifton <nickc@redhat.com>
1945 * target.def (narrow_volatile_bitfield): Note that the default
1946 value is false, not !TARGET_STRICT_ALIGN.
1947 * doc/tm.texi: Regenerate.
1949 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
1951 Fix LIB_SPEC for systems without libpthread.
1953 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
1954 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
1956 * config/i386/linux-common.h: Likewise.
1957 * config/mips/linux-common.h: Likewise.
1959 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1961 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
1964 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1966 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1967 instances can own GC refs.
1969 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1970 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1971 (gcc::context::gt_pch_nx): Likewise.
1972 (gcc::context::gt_pch_nx): Likewise.
1973 * ggc.h (gt_ggc_mx <T>): New.
1974 (gt_pch_nx_with_op <T>): New.
1975 (gt_pch_nx <T>): New.
1976 * passes.c (opt_pass::gt_ggc_mx): New.
1977 (opt_pass::gt_pch_nx): New.
1978 (opt_pass::gt_pch_nx_with_op): New.
1979 (pass_manager::gt_ggc_mx): New.
1980 (pass_manager::gt_pch_nx): New.
1981 (pass_manager::gt_pch_nx_with_op): New.
1982 (pass_manager::operator new): Use
1983 ggc_internal_cleared_alloc_stat rather than xcalloc.
1984 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1985 (pass_manager::gt_ggc_mx): New.
1986 (pass_manager::gt_pch_nx): New.
1987 (pass_manager::gt_pch_nx_with_op): New.
1988 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1989 (opt_pass::operator new): New.
1990 (opt_pass::gt_ggc_mx): New.
1991 (opt_pass::gt_pch_nx): New.
1992 (opt_pass::gt_pch_nx_with_op): New.
1994 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1996 * Makefile.in (GTFILES): Add context.h.
1997 * context.c (gcc::context::operator new): New.
1998 (gcc::context::gt_ggc_mx): New.
1999 (gcc::context::gt_pch_nx): New.
2000 (gcc::context::gt_pch_nx): New.
2001 * context.h (gcc::context): Add GTY((user)) marking.
2002 (gcc::context::operator new): New.
2003 (gcc::context::gt_ggc_mx): New.
2004 (gcc::context::gt_pch_nx): New.
2005 (gcc::context::gt_pch_nx): New.
2006 (g): Add GTY marking.
2007 (gt_ggc_mx (gcc::context *)): New.
2008 (gt_pch_nx (gcc::context *)): New.
2009 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2010 void *cookie)): New.
2011 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2013 2013-08-20 Alan Modra <amodra@gmail.com>
2016 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
2017 (rs6000_emit_epilogue): Likewise.
2019 2013-08-19 Dehao Chen <dehao@google.com>
2021 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
2023 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
2024 Jakub Jelinek <jakub@redhat.com>
2026 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
2027 (BUILT_IN_FABSD64): Likewise.
2028 (BUILT_IN_FABSD128): Likewise.
2029 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
2030 (fold_builtin_1): Likewise.
2031 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
2032 destination and source operands.
2033 (*abstd2_fpr): Likewise.
2034 (*nabstd2_fpr): Likewise.
2036 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
2038 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
2041 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
2043 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
2046 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
2048 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
2049 * builtins.c (default_libc_has_function): New.
2050 (gnu_libc_has_function): Ditto.
2051 (no_c99_libc_has_function): Ditto.
2052 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
2053 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
2054 (fold_builtin_sincos): Likewise.
2055 (fold_builtin_cexp): Likewise.
2056 * builtins.def (DEF_C94_BUILTIN): Likewise.
2057 (DEF_C99_BUILTIN): Likewise.
2058 (DEF_C99_C90RES_BUILTIN): Likewise.
2059 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
2060 definitions to using this define.
2061 * config/darwin-protos.h (darwin_libc_has_function): New.
2062 * config/darwin.c (darwin_libc_has_function): Ditto.
2063 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
2064 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
2065 * config/darwin.h: Ditto.
2066 * config/elfos.h: Ditto.
2067 * config/freebsd.h: Ditto.
2068 * config/i386/cygming.h: Ditto.
2069 * config/i386/djgpp.h: Ditto.
2070 * config/i386/i386-interix.h: Ditto.
2071 * config/microblaze/microblaze.h: Ditto.
2072 * config/mmix/mmix.h: Ditto.
2073 * config/gnu-user.h: Ditto.
2074 * config/ia64/hpux.h: Ditto.
2075 * config/pa/pa-hpux.h: Ditto.
2076 * config/pdp11/pdp11.h: Ditto.
2077 * config/picochip/picochip.h: Ditto.
2078 * config/linux.h: Ditto.
2079 * config/netbsd.h: Ditto.
2080 * config/openbsd.h: Ditto.
2081 * config/rs6000/aix43.h: Ditto.
2082 * config/rs6000/aix51.h: Ditto.
2083 * config/rs6000/aix52.h: Ditto.
2084 * config/rs6000/aix53.h: Ditto.
2085 * config/rs6000/aix61.h: Ditto.
2086 * config/rs6000/darwin.h: Ditto.
2087 * config/rs6000/linux.h: Ditto.
2088 * config/rs6000/linux64.h: Ditto.
2089 * config/s390/tpf.h: Ditto.
2090 * config/sol2-10.h: Ditto.
2091 * config/sol2.h: Ditto.
2092 * config/vms/vms.h: Ditto.
2093 * config/vxworks.h: Ditto.
2094 * config/linux-android.c (linux_android_libc_has_function):
2095 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
2096 * config/linux-protos.h (linux_android_libc_has_function):
2098 * config/i386/i386.c (ix86_libc_has_function): New.
2099 * config/i386/i386-protos.h
2100 (ix86_libc_has_function): New declaration.
2101 * config/i386/i386.md
2102 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
2103 ("isinf<mode>2): Likewise.
2104 * convert.c (convert_to_integer): Using new target hook
2105 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
2106 TARGET_C99_FUNCTIONS.
2107 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
2108 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
2109 * coretypes.h (function_class): New enum for different
2110 classes of functions.
2111 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
2112 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
2113 (TARGET_HAS_SINCOS): Likewise.
2114 (TARGET_LIBC_HAS_FUNCTION): New.
2115 * doc/tm.texi: Regenerated.
2116 * targhooks.h (default_libc_has_function): New declaration.
2117 (no_c99_libc_has_function): Ditto.
2118 (gnu_libc_has_function): Ditto.
2119 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
2120 and TARGET_HAS_SINCOS.
2122 2013-08-18 Jan Hubicka <jh@suse.cz>
2124 * Makeifle-in (ipa-devirt.o): New.
2125 (GTFILES): Add ipa-utils.h and ipa-devirt.c
2126 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
2127 (analyze_functions): Look into possible targets of polymorphic call.
2128 * dumpfile.c (dump_files): Add type-inheritance dump.
2129 * dumpfile.h (TDI_inheritance): New.
2130 * ipa-devirt.c: New file.
2131 * ipa-utils.h (odr_type_d): Forward declare.
2132 (odr_type): New type.
2133 (build_type_inheritance_graph): Declare.
2134 (possible_polymorphic_call_targets): Declare and introduce inline
2135 variant when only edge is pased.
2136 (dump_possible_polymorphic_call_targets): Likewise.
2137 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
2138 * tree.c (type_in_anonymous_namespace_p): Break out from ...
2139 (types_same_for_odr): ... here.
2140 * tree.h (type_in_anonymous_namespace_p): Declare.
2142 2013-08-18 Jakub Jelinek <jakub@redhat.com>
2144 PR tree-optimization/58006
2145 * tree-parloops.c (take_address_of): Don't ICE if get_name
2147 (eliminate_local_variables_stmt): Remove clobber stmts.
2149 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
2151 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
2152 error message is issued for an alias to undefined symbol.
2154 2013-08-18 Jan Hubicka <jh@suse.cz>
2156 * cgraph.c (cgraph_create_indirect_edge): Discover
2157 polymorphic calls and record basic info into indirect_info.
2158 * gimple-fold.c (gimple_fold_call): When doing BINFO based
2159 devirtualization, ignore objc function calls.
2160 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
2161 call with no parm index info.
2162 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
2163 * tree.c (virtual_method_call_p): New function.
2164 * tree.h (virtual_method_call_p): Declare.
2166 2013-08-16 Jan Hubicka <jh@suse.cz>
2169 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
2171 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
2173 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
2176 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2178 * gengtype.c (type_for_name): Add special-case support for
2179 locating types within the "gcc::" namespace.
2180 (open_base_files): Emit a "using namespace gcc" directive.
2182 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2185 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
2186 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
2188 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
2189 array instead of each individual operand as a separate argument.
2190 (emit_fusion_gpr_load): Likewise.
2191 (expand_fusion_gpr_load): Add new function declaration.
2193 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
2194 signature to have the operands passed as an array, instead of as
2195 separate arguments. Allow ZERO_EXTEND to be in the memory
2196 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
2197 depend on the register live/dead flags when peepholes are run.
2198 (expand_fusion_gpr_load): New function to be called from the
2199 peephole2 pass, to change the register that addis sets to be the
2201 (emit_fusion_gpr_load): Change the calling signature to have the
2202 operands passed as an array, instead of as separate arguments.
2203 Allow ZERO_EXTEND to be in the memory address, and also
2204 SIGN_EXTEND if -mpower8-fusion-sign.
2206 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
2208 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
2209 adjust the register addis loads up in the peephole2 pass. Do not
2210 depend on the register live/dead state when the peephole pass is done.
2212 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2214 * gengtype.c (create_user_defined_type): Ensure that the kind
2215 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
2216 declaration is seen before the GTY((user)) marking.
2218 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
2221 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
2223 2013-08-16 Jan Hubicka <jh@suse.cz>
2225 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
2226 arugment expected_type.
2227 (gimple_fold_call): Use it.
2228 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
2229 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
2230 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
2231 (try_make_edge_direct_virtual_call): Likewise.
2232 * tree.c (obj_type_ref_class): New.
2233 * tree.h (obj_type_ref_class): Use it.
2235 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
2237 * sched-vis.c (rtl_slim_pp_initialized): Remove.
2238 (rtl_slim_pp): Likewise.
2239 (init_rtl_slim_pretty_print): Likewise.
2240 (dump_value_slim): Don't call it. Use local pretty printer.
2241 (dump_insn_slim): Likewise.
2242 (dump_rtl_slim): Likewise.
2243 (str_pattern_slim): Likewise.
2244 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
2247 2013-08-16 Jakub Jelinek <jakub@redhat.com>
2249 PR tree-optimization/58164
2250 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
2251 walk gimple_goto_dest of GIMPLE_GOTO.
2253 PR tree-optimization/58165
2254 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
2255 bi_call must be the last stmt in a bb, don't split_block, instead
2256 use fallthru edge from it and give up if there is none.
2257 Release conds vector when returning early.
2259 2013-08-14 Xinliang David Li <davidxl@google.com>
2261 * config/i386/i386.c (ix86_option_override_internal):
2262 Remove unused variable and field.
2264 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2267 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
2268 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
2269 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
2270 and Linux, correct BLKmode alignment when 128-bit alignment is
2271 required and compatibility flag is not set.
2272 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
2273 for zero-size arguments when compatibility flag is not set.
2275 2013-08-14 Jakub Jelinek <jakub@redhat.com>
2277 PR tree-optimization/58145
2278 * tree-sra.c (build_ref_for_offset): If prev_base has
2279 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
2281 2013-08-14 Xinliang David Li <davidxl@google.com>
2283 * config/i386/i386.c (ix86_option_override_internal):
2284 Fix uninitialized variable error.
2286 2013-08-14 Xinliang David Li <davidxl@google.com>
2288 * config/i386/i386.opt: Define two new options.
2289 * config/i386/x86-tune.def: Add arch selector field in macros.
2290 * config/i386/i386.h: Adjust macro definition.
2291 * config/i386/i386.c (ix86_option_override_internal):
2293 (parse_mtune_ctrl_str): New function.
2294 (set_ix86_tune_features): New function.
2295 (ix86_function_specific_restore): Call the new helper function.
2297 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
2299 PR rtl-optimization/57662
2300 * sel-sched.c (code_motion_process_successors): When the current insn
2301 is removed after the recursive traversal, break from the loop.
2302 Add comments and debug printouts.
2304 2013-08-14 Jakub Jelinek <jakub@redhat.com>
2305 Alexandre Oliva <aoliva@redhat.com>
2308 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
2309 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
2310 there also UNSPEC_PLTOFF.
2312 2013-08-14 Marek Polacek <polacek@redhat.com>
2314 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
2315 NUM_CONDITIONS bit positions.
2317 2013-08-13 Cary Coutant <ccoutant@google.com>
2319 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
2320 (attr_checksum): Hash vector contents instead of pointer.
2321 (attr_checksum_ordered): Likewise.
2323 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
2325 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
2326 when Pmode != word_mode. Add length_address attribute.
2327 (sse3_monitor_<mode>): Merge from sse3_monitor and
2328 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
2329 Pmode != word_mode. Update insn length attribute.
2330 * config/i386/i386.c (ix86_option_override_internal): Update
2331 ix86_gen_monitor selection for merged sse3_monitor insn.
2333 2013-08-13 Julian Brown <julian@codesourcery.com>
2335 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
2336 perform invalid legitimization on greater-than-word-size modes for
2339 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
2341 * ira.c (setup_class_translate_array): Use aclass instead of cl
2342 for classes not fully covered by allocno classes.
2344 2013-08-13 Jakub Jelinek <jakub@redhat.com>
2346 PR tree-optimization/57661
2347 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
2348 * tree-inline.c (tree_function_versioning): Initialize it.
2349 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
2350 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
2351 that is not being copied.
2354 * asan.c (instrument_strlen_call): Fix typo in comment.
2355 Use char * type even for the lhs of POINTER_PLUS_EXPR.
2357 2013-08-13 Steve Ellcey <sellcey@mips.com>
2359 * config/mips/mips.md (prefetch): Use lw instead of ld on
2360 loongson in 32bit mode.
2362 2013-08-13 Nick Clifton <nickc@redhat.com>
2364 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
2366 2013-08-13 Jan Hubicka <jh@suse.cz>
2368 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
2369 introduced edge; fix typo in sanity check.
2370 (cgraph_resolve_speculation): Export; improve diagnostic.
2371 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
2372 speculation at type mismatch.
2373 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
2374 (cgraph_resolve_speculation): Declare.
2375 (symtab_can_be_discarded): New function.
2376 * value-prof.c (gimple_ic_transform): Remove actual transform code.
2377 * ipa-inline-transform.c (speculation_removed): New global var.
2378 (clone_inlined_nodes): See if speculation can be removed.
2379 (inline_call): If speculations was removed, we growths may not match.
2380 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
2381 (speculation_useful_p): New function.
2382 (resolve_noninline_speculation): New function.
2383 (inline_small_functions): Resolve useless speculations.
2384 * ipa-inline.h (speculation_useful_p): Declare
2385 * ipa.c (can_replace_by_local_alias): Simplify.
2386 (ipa_profile): Produce speculative calls in non-lto, too;
2387 add simple cost model; produce local aliases.
2389 2013-08-13 David Malcolm <dmalcolm@redhat.com>
2391 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
2394 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
2396 * config/i386/i386.c (ix86_function_versions): Use error + inform.
2398 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
2400 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
2401 iterator instead of X87MODEF.
2403 2013-08-12 Perez Read <netfirewall@gmail.com>
2406 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
2407 operand 0 for intel asm alternative.
2408 (*movabs<mode>_2): Ditto for operand 1.
2410 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
2412 * config/aarch64/arm_none.h
2413 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
2415 2013-08-12 Nick Clifton <nickc@redhat.com>
2417 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
2419 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2421 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
2422 expand for QI/HImode operand to produce more effictive code for
2423 unsigned char(short) --> float(double) conversion.
2425 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
2427 * doc/invoke.texi: Mention that -ftls-model does not force the final
2430 2013-08-12 Marek Polacek <polacek@redhat.com>
2431 Marc Glisse <marc.glisse@inria.fr>
2433 PR tree-optimization/57980
2434 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
2435 when creating -1 constant.
2437 2013-08-10 Jan Hubicka <jh@suse.cz>
2439 Workaround binutils PR14342.
2440 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
2441 (gimple_init_edge_profiler): Likewise.
2442 (gimple_gen_ic_func_profiler): Likewise.
2444 2013-08-09 Jan Hubicka <jh@suse.cz>
2446 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
2448 2013-08-09 Xinliang David Li <davidxl@google.com>
2450 * config/i386/stringop.def: New file.
2451 * config/i386/stringop.opt: New file.
2452 * config/i386/i386-opts.h: Include stringopt.def.
2453 * config/i386/i386.opt: Include stringopt.opt.
2454 * config/i386/i386.c (ix86_option_override_internal):
2455 Override default size based stringop inline strategies with options.
2456 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
2459 2013-08-09 Jan Hubicka <jh@suse.cz>
2461 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
2463 2013-08-09 Jan Hubicka <jh@suse.cz>
2465 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
2467 (dump_cgraph_node): Dump profile-id.
2468 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
2469 and common_target_probability.
2470 * lto-cgraph.c (lto_output_edge): Stream common targets.
2471 (lto_output_node): Stream profile ids.
2472 (input_node): Stream profile ids.
2473 (input_edge): Stream common targets.
2474 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
2475 * ipa.c: Include value-prof.h
2476 (ipa_profile_generate_summary): Turn indirect call statement histograms
2477 into common targets.
2478 (ipa_profile): Turn common targets into speculative edges.
2480 2013-08-09 Jan Hubicka <jh@suse.cz>
2482 * cgraph.h (cgraph_node): Add profile_id.
2483 * value-prof.c (cgraph_node_map): Turn into pointer_map.
2484 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2485 (del_node_map): Update.
2486 (find_func_by_funcdef_no): Replace by ...
2487 (find_func_by_profile_id): ... this one.
2488 (gimple_ic_transform): Do not remove useful histograms when
2489 speculation is not done; dump info when indirect call removal
2491 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
2492 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
2493 (__gcov_indirect_call_profiler_v2): .. this one.
2494 * profile.h (init_node_map): Update.
2495 * coverage.c (coverage_compute_profile_id): New function.
2496 * coverage.h (coverage_compute_profile_id): Declare.
2497 * tree-profile.c (init_ic_make_global_vars): Make
2498 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
2499 (gimple_init_edge_profiler): Update prototype of
2500 __gcov_indirect_call_profiler.
2501 (gimple_gen_ic_func_profiler): Simplify.
2502 (tree_profiling): Use init_node_map
2504 2013-08-09 Jan Hubicka <jh@suse.cz>
2506 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
2507 non-speculative refs.
2508 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
2509 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
2510 (cgraph_set_call_stmt): Likewise.
2511 (cgraph_create_edge_1): Fix release checking compilatoin;
2513 (cgraph_free_edge): Free indirect info.
2514 (cgraph_turn_edge_to_speculative): New function.
2515 (cgraph_speculative_call_info): New function.
2516 (cgraph_make_edge_direct): Return direct edge; handle speculation.
2517 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
2518 (dump_cgraph_node): Dump speculation.
2519 (verify_edge_count_and_frequency): Accept speculative edges.
2520 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
2521 (verify_cgraph_node): Handle speculation.
2522 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
2523 (cgraph_set_call_stmt): Update prototype.
2524 (cgraph_make_edge_direct): Update prototype.
2525 (cgraph_speculative_call_info): Declare.
2526 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
2527 to change; update call of ipa_find_references.
2528 * ipa-ref.c (ipa_record_reference): Fix return value; clear
2529 lto_stmt_uid and speculative flags.
2530 (ipa_dump_references): Dump speculation.
2531 (ipa_clone_references): Clone speculative flag.
2532 (ipa_clone_referring): Likewise.
2533 (ipa_clone_ref): New function.
2534 (ipa_find_reference): Look into lto_stmt_uids
2535 (ipa_clear_stmts_in_references): Do not clear speculative calls.
2536 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
2537 (ipa_find_reference): Update declaration.
2538 (ipa_clone_ref): Declare.
2539 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
2540 stream speculative flag.
2541 (lto_output_ref): Stream statements uids and speculation.
2542 (input_ref): Likewise.
2543 (input_edge): Stream speuclation.
2544 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
2545 (cgraph_set_call_stmt_including_clones): Handle speculation.
2546 * ipa-inline.c (heap_edge_removal_hook): New function.
2547 (inline_small_functions): Register it.
2548 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
2549 also initialize refs.
2550 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
2552 (try_make_edge_direct_simple_call): Likewise.
2553 (try_make_edge_direct_simple_call): Likewise.
2554 (update_indirect_edges_after_inlining): Likewise.
2555 (remove_described_reference): Look proper lto_stmt_uid.
2556 (propagate_controlled_uses): Likewise.
2557 (propagate_controlled_uses): Liekwise.
2558 * tree-inline.c (copy_bb): Copy speculative edges.
2559 (redirect_all_calls): New function.
2560 (copy_cfg_body): Do redirection after loop info is updated.
2561 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
2563 2013-08-09 Jan Hubicka <jh@suse.cz>
2565 * lto-streamer-out.c (output_function): Renumber PHIs.
2566 * lto-streamer-in.c (input_function): Likewise.
2568 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
2570 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
2571 (get_lane_unsigned): Likewise.
2572 (dup_lane_scalar): Likewise.
2573 (get_lane): enable for VALL.
2574 * config/aarch64/aarch64-simd.md
2575 (aarch64_dup_lane_scalar<mode>): Remove.
2576 (aarch64_get_lane_signed<mode>): Likewise.
2577 (aarch64_get_lane_unsigned<mode>): Likewise.
2578 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
2579 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
2580 (aarch64_get_lane<mode>): Enable for all vector modes.
2581 (aarch64_get_lanedi): Remove misleading constraints.
2582 * config/aarch64/arm_neon.h
2583 (__aarch64_vget_lane_any): Define.
2584 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
2585 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
2586 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
2587 * config/aarch64/iterators.md (VDQQH): New.
2591 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
2593 * configure.ac: Add GAS check for LEON instructions on SPARC.
2594 * configure: Regenerate.
2595 * config.in: Likewise.
2596 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
2598 * config/sparc/sparc.opt (LEON, LEON3): New masks.
2599 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
2601 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
2602 (AS_LEON_FLAG): New macro.
2603 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
2604 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
2605 Deal with LEON and LEON3 for the memory model.
2606 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
2607 (atomic_compare_and_swap<mode>_1): Likewise.
2608 (*atomic_compare_and_swap<mode>_1): Likewise.
2610 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2612 * config/arm/neon.md (vcond): Fix floating-point vector
2613 comparisons against 0.
2615 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
2617 * lra-constraints.c (emit_spill_move): Remove assert.
2618 (process_alt_operands): Add more debugging
2619 output. Increase reject for spilling into memory. Decrease
2620 reject for reloading scratch.
2621 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
2623 2013-08-08 Steve Ellcey <sellcey@mips.com>
2625 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
2626 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
2627 micromips incompatible. Add nan2008.
2628 (MULTILIB_DIRNAMES): Add nan2008.
2629 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2630 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
2631 and micromips incompatible. Add nan2008.
2632 (MULTILIB_DIRNAMES): Add nan2008.
2633 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2635 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
2637 PR rtl-optimization/58079
2638 * combine.c (combine_simplify_rtx): Avoid using SUBST if
2639 simplify_comparison has widened a comparison with an integer.
2641 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2643 * config/arm/neon.md (movmisalign<mode>): Disable when we
2644 don't allow unaligned accesses.
2645 (*movmisalign<mode>_neon_store): Likewise.
2646 (*movmisalign<mode>_neon_load): Likewise.
2647 (*movmisalign<mode>_neon_store): Likewise.
2648 (*movmisalign<mode>_neon_load): Likewise.
2650 2013-08-08 Jan Hubicka <jh@suse.cz>
2652 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
2653 (make_pass_rebuild_cgraph_edges): Also clear references.
2654 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
2655 * ipa-inline-transform.c (inline_transform): Remove all references
2657 * cgraphunit.c (expand_function): Remove all references after
2659 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2660 (ipa_find_reference): Rewrite to iterator.
2661 (remove_stmt_references): Likewise.
2662 (ipa_clear_stmts_in_references): New function.
2663 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
2664 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2666 * ipa-split.c (split_function): Remove references in split function.
2668 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2671 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
2672 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2674 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2677 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2678 suggested mode for the assignment isn't compatible with the
2681 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2684 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2686 2013-08-07 Xinliang David Li <davidxl@google.com>
2688 * config/i386/i386.opt: New option -mtune-ctrl=.
2689 * config/i386/x86-tune.def: New file.
2690 * config/i386/i386.h: include x86-tune.def.
2691 * config/i386/i386.c (ix86_option_override_internal):
2692 Parsing -mtune-ctrl= option and set tune features.
2694 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2697 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2698 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2701 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2703 * rtl.h (update_alignments): Declare.
2704 * final.c (grow_label_align): New function extracted from...
2705 (shorten_branches): ...here. Call it.
2706 (update_alignments): New function.
2707 * reorg.c (sibling_labels): New variable.
2708 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2709 the new label along with it onto the sibling_labels vector.
2710 (fill_simple_delay_slots): Adjust call to get_label_before.
2711 (fill_slots_from_thread): Likewise.
2712 (relax_delay_slots): Likewise.
2713 (make_return_insns): Likewise.
2714 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2716 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2718 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2719 document its semantics.
2720 (diagnostic_report_diagnostic): Adjust accordingly.
2722 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2724 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2725 (sparc_option_override): ...and port to new C++ pass API.
2726 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2728 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2730 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2732 2013-08-06 Caroline Tice <cmtice@google.com>
2734 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2735 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2736 * tree-pass.h: Add pass_vtable_verify.
2737 * varasm.c (assemble_variable): Add code to properly set the comdat
2738 section and name for the .vtable_map_vars section.
2739 (assemble_vtyv_preinit_initializer): New function.
2740 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2742 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2743 * vtable-verify.c: New file.
2744 * vtable-verify.h: New file.
2745 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2746 initialiation levels.
2747 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2748 * passes.def: Insert pass_vtable_verify.
2749 * aclocal.m4: Reorder includes.
2750 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2751 -fvtv-counts options.
2752 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2753 as appropriate, if -fvtable-verify=... is used.
2754 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2755 -fvtable-verify=... is used.
2756 * Makefile.in (OBJS): Add vtable-verify.o to list.
2757 (vtable-verify.o): Add new build rule.
2758 (GTFILES): Add vtable-verify.c to list.
2759 * common.opt (fvtable-verify=): New flag.
2760 (vtv_priority): Values for fvtable-verify= flag.
2761 (fvtv-counts): New flag.
2762 (fvtv-debug): New flag.
2763 * tree.h (save_vtable_map_decl): New extern function decl.
2765 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2767 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2768 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2769 (pass_data_rl78_devirt): ...new pass_data instance and...
2770 (make_pass_rl78_devirt): ...new function.
2771 (rl78_asm_file_start): Port pass registration to new C++ API.
2773 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2775 * coretypes.h (rtl_opt_pass): Add.
2776 (gcc::context): Add.
2777 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2778 (epiphany_init): Port to new C++ pass API.
2779 (epiphany_optimize_mode_switching): Likewise.
2780 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2781 (pass_manager::get_pass_mode_switching): New.
2782 (pass_manager::get_pass_peephole2): New.
2783 * mode-switching.c (pass_mode_switching): Add clone method.
2784 * recog.c (pass_peephole2): Add clone method.
2785 (pass_split_all_insns): Add clone method.
2787 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2789 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2790 (mips_option_override): ...here, porting to new C++ API for passes.
2792 2013-08-06 Jan Hubicka <jh@suse.cz>
2794 * cgraph.c (cgraph_get_body): New function based on lto.c
2796 * cgraph.h (cgraph_get_body): Declare.
2797 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2799 * cgraphunit.c (expand_function): Get body prior expanding.
2800 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2801 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2803 * passes.c (do_per_function_toporder): Get body.
2804 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2805 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2807 2013-08-06 Martin Jambor <mjambor@suse.cz>
2810 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2811 is not re-finalized. Rename second parameter to no_collect.
2813 2013-08-06 Martin Jambor <mjambor@suse.cz>
2816 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2817 MEM_REF has proper alignment information.
2819 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2822 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2824 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2825 rtx (*) (rtx, ...) with insn_gen_fn.
2826 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2827 insn_gen_fn::stored_funcptr. Add initializer braces.
2829 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2831 Rewrite how instances of passes are cloned to remove assumptions
2832 about their sizes (thus allowing pass subclasses to have
2833 additional data fields, albeit non-GC-managed ones at this point).
2835 * passes.c (make_pass_instance): Now that passes have clone
2836 methods, rewrite this function to eliminate XNEW and memcpy
2837 calls that used hardcoded sizes. Since this function no longer
2838 creates pass instances, rename it to...
2839 (add_pass_instance): ...this. Document the old way that passes were
2840 numbered and flagged, and rework this function to continue using it.
2841 (next_pass_1): Add an initial_pass argument for use by
2843 (position_pass): When adding multiple instances of a pass, use
2844 the pass's clone method, rather than relying on the XNEW/memcpy
2845 within the former make_pass_instance (now add_pass_instance).
2846 (pass_manager::pass_manager): When invoking next_pass_1, also supply
2847 the initial instance of the current pass within the pass manager.
2849 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2851 This is the automated part of the conversion of passes from C
2852 structs to C++ classes.
2854 Patch autogenerated by refactor_passes.py from
2855 https://github.com/davidmalcolm/gcc-refactoring-scripts
2856 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
2858 * asan.c (pass_asan): Convert from a global struct to a subclass of
2859 gimple_opt_pass along with...
2860 (pass_data_asan): ...new pass_data instance and...
2861 (make_pass_asan): ...new function.
2862 (pass_asan_O0): Convert from a global struct to a subclass of
2863 gimple_opt_pass along with...
2864 (pass_data_asan_O0): ...new pass_data instance and...
2865 (make_pass_asan_O0): ...new function.
2866 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
2867 subclass of rtl_opt_pass along with...
2868 (pass_data_inc_dec): ...new pass_data instance and...
2869 (make_pass_inc_dec): ...new function.
2870 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
2871 a subclass of rtl_opt_pass along with...
2872 (pass_data_reorder_blocks): ...new pass_data instance and...
2873 (make_pass_reorder_blocks): ...new function.
2874 (pass_duplicate_computed_gotos): Convert from a global struct to a
2875 subclass of rtl_opt_pass along with...
2876 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
2877 (make_pass_duplicate_computed_gotos): ...new function.
2878 (pass_partition_blocks): Convert from a global struct to a subclass of
2879 rtl_opt_pass along with...
2880 (pass_data_partition_blocks): ...new pass_data instance and...
2881 (make_pass_partition_blocks): ...new function.
2882 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
2883 struct to a subclass of rtl_opt_pass along with...
2884 (pass_data_branch_target_load_optimize1): ...new pass_data instance
2886 (make_pass_branch_target_load_optimize1): ...new function.
2887 (pass_branch_target_load_optimize2): Convert from a global struct to a
2888 subclass of rtl_opt_pass along with...
2889 (pass_data_branch_target_load_optimize2): ...new pass_data instance
2891 (make_pass_branch_target_load_optimize2): ...new function.
2892 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
2893 of rtl_opt_pass along with...
2894 (pass_data_jump): ...new pass_data instance and...
2895 (make_pass_jump): ...new function.
2896 (pass_jump2): Convert from a global struct to a subclass of
2897 rtl_opt_pass along with...
2898 (pass_data_jump2): ...new pass_data instance and...
2899 (make_pass_jump2): ...new function.
2900 * cfgexpand.c (pass_expand): Convert from a global struct to a
2901 subclass of rtl_opt_pass along with...
2902 (pass_data_expand): ...new pass_data instance and...
2903 (make_pass_expand): ...new function.
2904 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
2905 of rtl_opt_pass along with...
2906 (pass_data_free_cfg): ...new pass_data instance and...
2907 (make_pass_free_cfg): ...new function.
2908 (pass_into_cfg_layout_mode): Convert from a global struct to a
2909 subclass of rtl_opt_pass along with...
2910 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
2911 (make_pass_into_cfg_layout_mode): ...new function.
2912 (pass_outof_cfg_layout_mode): Convert from a global struct to a
2913 subclass of rtl_opt_pass along with...
2914 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
2915 (make_pass_outof_cfg_layout_mode): ...new function.
2916 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
2917 struct to a subclass of gimple_opt_pass along with...
2918 (pass_data_build_cgraph_edges): ...new pass_data instance and...
2919 (make_pass_build_cgraph_edges): ...new function.
2920 (pass_rebuild_cgraph_edges): Convert from a global struct to a
2921 subclass of gimple_opt_pass along with...
2922 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
2923 (make_pass_rebuild_cgraph_edges): ...new function.
2924 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
2925 subclass of gimple_opt_pass along with...
2926 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
2928 (make_pass_remove_cgraph_callee_edges): ...new function.
2929 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
2930 struct to a subclass of rtl_opt_pass along with...
2931 (pass_data_stack_adjustments): ...new pass_data instance and...
2932 (make_pass_stack_adjustments): ...new function.
2933 * combine.c (pass_combine): Convert from a global struct to a subclass
2934 of rtl_opt_pass along with...
2935 (pass_data_combine): ...new pass_data instance and...
2936 (make_pass_combine): ...new function.
2937 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
2938 global struct to a subclass of rtl_opt_pass along with...
2939 (pass_data_compare_elim_after_reload): ...new pass_data instance
2941 (make_pass_compare_elim_after_reload): ...new function.
2942 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
2943 of rtl_opt_pass along with...
2944 (pass_data_rtl_cprop): ...new pass_data instance and...
2945 (make_pass_rtl_cprop): ...new function.
2946 * cse.c (pass_cse): Convert from a global struct to a subclass of
2947 rtl_opt_pass along with...
2948 (pass_data_cse): ...new pass_data instance and...
2949 (make_pass_cse): ...new function.
2950 (pass_cse2): Convert from a global struct to a subclass of
2951 rtl_opt_pass along with...
2952 (pass_data_cse2): ...new pass_data instance and...
2953 (make_pass_cse2): ...new function.
2954 (pass_cse_after_global_opts): Convert from a global struct to a
2955 subclass of rtl_opt_pass along with...
2956 (pass_data_cse_after_global_opts): ...new pass_data instance and...
2957 (make_pass_cse_after_global_opts): ...new function.
2958 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
2959 of rtl_opt_pass along with...
2960 (pass_data_ud_rtl_dce): ...new pass_data instance and...
2961 (make_pass_ud_rtl_dce): ...new function.
2962 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
2963 rtl_opt_pass along with...
2964 (pass_data_fast_rtl_dce): ...new pass_data instance and...
2965 (make_pass_fast_rtl_dce): ...new function.
2966 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
2967 a subclass of rtl_opt_pass along with...
2968 (pass_data_df_initialize_opt): ...new pass_data instance and...
2969 (make_pass_df_initialize_opt): ...new function.
2970 (pass_df_initialize_no_opt): Convert from a global struct to a
2971 subclass of rtl_opt_pass along with...
2972 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
2973 (make_pass_df_initialize_no_opt): ...new function.
2974 (pass_df_finish): Convert from a global struct to a subclass of
2975 rtl_opt_pass along with...
2976 (pass_data_df_finish): ...new pass_data instance and...
2977 (make_pass_df_finish): ...new function.
2978 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
2979 rtl_opt_pass along with...
2980 (pass_data_rtl_dse1): ...new pass_data instance and...
2981 (make_pass_rtl_dse1): ...new function.
2982 (pass_rtl_dse2): Convert from a global struct to a subclass of
2983 rtl_opt_pass along with...
2984 (pass_data_rtl_dse2): ...new pass_data instance and...
2985 (make_pass_rtl_dse2): ...new function.
2986 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
2987 subclass of rtl_opt_pass along with...
2988 (pass_data_dwarf2_frame): ...new pass_data instance and...
2989 (make_pass_dwarf2_frame): ...new function.
2990 * except.c (pass_set_nothrow_function_flags): Convert from a global
2991 struct to a subclass of rtl_opt_pass along with...
2992 (pass_data_set_nothrow_function_flags): ...new pass_data instance
2994 (make_pass_set_nothrow_function_flags): ...new function.
2995 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
2996 subclass of rtl_opt_pass along with...
2997 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
2999 (make_pass_convert_to_eh_region_ranges): ...new function.
3000 * final.c (pass_compute_alignments): Convert from a global struct to a
3001 subclass of rtl_opt_pass along with...
3002 (pass_data_compute_alignments): ...new pass_data instance and...
3003 (make_pass_compute_alignments): ...new function.
3004 (pass_final): Convert from a global struct to a subclass of
3005 rtl_opt_pass along with...
3006 (pass_data_final): ...new pass_data instance and...
3007 (make_pass_final): ...new function.
3008 (pass_shorten_branches): Convert from a global struct to a subclass of
3009 rtl_opt_pass along with...
3010 (pass_data_shorten_branches): ...new pass_data instance and...
3011 (make_pass_shorten_branches): ...new function.
3012 (pass_clean_state): Convert from a global struct to a subclass of
3013 rtl_opt_pass along with...
3014 (pass_data_clean_state): ...new pass_data instance and...
3015 (make_pass_clean_state): ...new function.
3016 * function.c (pass_instantiate_virtual_regs): Convert from a global
3017 struct to a subclass of rtl_opt_pass along with...
3018 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
3019 (make_pass_instantiate_virtual_regs): ...new function.
3020 (pass_leaf_regs): Convert from a global struct to a subclass of
3021 rtl_opt_pass along with...
3022 (pass_data_leaf_regs): ...new pass_data instance and...
3023 (make_pass_leaf_regs): ...new function.
3024 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
3025 subclass of rtl_opt_pass along with...
3026 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
3028 (make_pass_thread_prologue_and_epilogue): ...new function.
3029 (pass_match_asm_constraints): Convert from a global struct to a
3030 subclass of rtl_opt_pass along with...
3031 (pass_data_match_asm_constraints): ...new pass_data instance and...
3032 (make_pass_match_asm_constraints): ...new function.
3033 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
3034 subclass of rtl_opt_pass along with...
3035 (pass_data_rtl_fwprop): ...new pass_data instance and...
3036 (make_pass_rtl_fwprop): ...new function.
3037 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
3038 rtl_opt_pass along with...
3039 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
3040 (make_pass_rtl_fwprop_addr): ...new function.
3041 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
3042 rtl_opt_pass along with...
3043 (pass_data_rtl_pre): ...new pass_data instance and...
3044 (make_pass_rtl_pre): ...new function.
3045 (pass_rtl_hoist): Convert from a global struct to a subclass of
3046 rtl_opt_pass along with...
3047 (pass_data_rtl_hoist): ...new pass_data instance and...
3048 (make_pass_rtl_hoist): ...new function.
3049 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
3050 subclass of gimple_opt_pass along with...
3051 (pass_data_lower_cf): ...new pass_data instance and...
3052 (make_pass_lower_cf): ...new function.
3053 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
3054 from a global struct to a subclass of gimple_opt_pass along with...
3055 (pass_data_strength_reduction): ...new pass_data instance and...
3056 (make_pass_strength_reduction): ...new function.
3057 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
3058 of rtl_opt_pass along with...
3059 (pass_data_rtl_ifcvt): ...new pass_data instance and...
3060 (make_pass_rtl_ifcvt): ...new function.
3061 (pass_if_after_combine): Convert from a global struct to a subclass of
3062 rtl_opt_pass along with...
3063 (pass_data_if_after_combine): ...new pass_data instance and...
3064 (make_pass_if_after_combine): ...new function.
3065 (pass_if_after_reload): Convert from a global struct to a subclass of
3066 rtl_opt_pass along with...
3067 (pass_data_if_after_reload): ...new pass_data instance and...
3068 (make_pass_if_after_reload): ...new function.
3069 * init-regs.c (pass_initialize_regs): Convert from a global struct to
3070 a subclass of rtl_opt_pass along with...
3071 (pass_data_initialize_regs): ...new pass_data instance and...
3072 (make_pass_initialize_regs): ...new function.
3073 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
3074 of ipa_opt_pass_d along with...
3075 (pass_data_ipa_cp): ...new pass_data instance and...
3076 (make_pass_ipa_cp): ...new function.
3077 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
3078 global struct to a subclass of gimple_opt_pass along with...
3079 (pass_data_inline_parameters): ...new pass_data instance and...
3080 (make_pass_inline_parameters): ...new function.
3081 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
3082 subclass of gimple_opt_pass along with...
3083 (pass_data_early_inline): ...new pass_data instance and...
3084 (make_pass_early_inline): ...new function.
3085 (pass_ipa_inline): Convert from a global struct to a subclass of
3086 ipa_opt_pass_d along with...
3087 (pass_data_ipa_inline): ...new pass_data instance and...
3088 (make_pass_ipa_inline): ...new function.
3089 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
3090 struct to a subclass of gimple_opt_pass along with...
3091 (pass_data_local_pure_const): ...new pass_data instance and...
3092 (make_pass_local_pure_const): ...new function.
3093 (pass_ipa_pure_const): Convert from a global struct to a subclass of
3094 ipa_opt_pass_d along with...
3095 (pass_data_ipa_pure_const): ...new pass_data instance and...
3096 (make_pass_ipa_pure_const): ...new function.
3097 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
3098 to a subclass of ipa_opt_pass_d along with...
3099 (pass_data_ipa_reference): ...new pass_data instance and...
3100 (make_pass_ipa_reference): ...new function.
3101 * ipa-split.c (pass_split_functions): Convert from a global struct to
3102 a subclass of gimple_opt_pass along with...
3103 (pass_data_split_functions): ...new pass_data instance and...
3104 (make_pass_split_functions): ...new function.
3105 (pass_feedback_split_functions): Convert from a global struct to a
3106 subclass of gimple_opt_pass along with...
3107 (pass_data_feedback_split_functions): ...new pass_data instance and...
3108 (make_pass_feedback_split_functions): ...new function.
3109 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
3110 global struct to a subclass of simple_ipa_opt_pass along with...
3111 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
3113 (make_pass_ipa_function_and_variable_visibility): ...new function.
3114 (pass_ipa_free_inline_summary): Convert from a global struct to a
3115 subclass of simple_ipa_opt_pass along with...
3116 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
3117 (make_pass_ipa_free_inline_summary): ...new function.
3118 (pass_ipa_whole_program_visibility): Convert from a global struct to a
3119 subclass of ipa_opt_pass_d along with...
3120 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
3122 (make_pass_ipa_whole_program_visibility): ...new function.
3123 (pass_ipa_profile): Convert from a global struct to a subclass of
3124 ipa_opt_pass_d along with...
3125 (pass_data_ipa_profile): ...new pass_data instance and...
3126 (make_pass_ipa_profile): ...new function.
3127 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
3128 ipa_opt_pass_d along with...
3129 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
3130 (make_pass_ipa_cdtor_merge): ...new function.
3131 * ira.c (pass_ira): Convert from a global struct to a subclass of
3132 rtl_opt_pass along with...
3133 (pass_data_ira): ...new pass_data instance and...
3134 (make_pass_ira): ...new function.
3135 (pass_reload): Convert from a global struct to a subclass of
3136 rtl_opt_pass along with...
3137 (pass_data_reload): ...new pass_data instance and...
3138 (make_pass_reload): ...new function.
3139 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
3140 subclass of rtl_opt_pass along with...
3141 (pass_data_cleanup_barriers): ...new pass_data instance and...
3142 (make_pass_cleanup_barriers): ...new function.
3143 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
3144 of rtl_opt_pass along with...
3145 (pass_data_loop2): ...new pass_data instance and...
3146 (make_pass_loop2): ...new function.
3147 (pass_rtl_loop_init): Convert from a global struct to a subclass of
3148 rtl_opt_pass along with...
3149 (pass_data_rtl_loop_init): ...new pass_data instance and...
3150 (make_pass_rtl_loop_init): ...new function.
3151 (pass_rtl_loop_done): Convert from a global struct to a subclass of
3152 rtl_opt_pass along with...
3153 (pass_data_rtl_loop_done): ...new pass_data instance and...
3154 (make_pass_rtl_loop_done): ...new function.
3155 (pass_rtl_move_loop_invariants): Convert from a global struct to a
3156 subclass of rtl_opt_pass along with...
3157 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
3158 (make_pass_rtl_move_loop_invariants): ...new function.
3159 (pass_rtl_unswitch): Convert from a global struct to a subclass of
3160 rtl_opt_pass along with...
3161 (pass_data_rtl_unswitch): ...new pass_data instance and...
3162 (make_pass_rtl_unswitch): ...new function.
3163 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
3164 subclass of rtl_opt_pass along with...
3165 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
3167 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3168 (pass_rtl_doloop): Convert from a global struct to a subclass of
3169 rtl_opt_pass along with...
3170 (pass_data_rtl_doloop): ...new pass_data instance and...
3171 (make_pass_rtl_doloop): ...new function.
3172 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
3173 a subclass of rtl_opt_pass along with...
3174 (pass_data_lower_subreg): ...new pass_data instance and...
3175 (make_pass_lower_subreg): ...new function.
3176 (pass_lower_subreg2): Convert from a global struct to a subclass of
3177 rtl_opt_pass along with...
3178 (pass_data_lower_subreg2): ...new pass_data instance and...
3179 (make_pass_lower_subreg2): ...new function.
3180 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
3181 struct to a subclass of ipa_opt_pass_d along with...
3182 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
3183 (make_pass_ipa_lto_gimple_out): ...new function.
3184 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
3185 of ipa_opt_pass_d along with...
3186 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
3187 (make_pass_ipa_lto_finish_out): ...new function.
3188 * mode-switching.c (pass_mode_switching): Convert from a global struct
3189 to a subclass of rtl_opt_pass along with...
3190 (pass_data_mode_switching): ...new pass_data instance and...
3191 (make_pass_mode_switching): ...new function.
3192 * modulo-sched.c (pass_sms): Convert from a global struct to a
3193 subclass of rtl_opt_pass along with...
3194 (pass_data_sms): ...new pass_data instance and...
3195 (make_pass_sms): ...new function.
3196 * omp-low.c (pass_expand_omp): Convert from a global struct to a
3197 subclass of gimple_opt_pass along with...
3198 (pass_data_expand_omp): ...new pass_data instance and...
3199 (make_pass_expand_omp): ...new function.
3200 (pass_lower_omp): Convert from a global struct to a subclass of
3201 gimple_opt_pass along with...
3202 (pass_data_lower_omp): ...new pass_data instance and...
3203 (make_pass_lower_omp): ...new function.
3204 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
3205 of gimple_opt_pass along with...
3206 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
3207 (make_pass_diagnose_omp_blocks): ...new function.
3208 * passes.c (pass_early_local_passes): Convert from a global struct to
3209 a subclass of simple_ipa_opt_pass along with...
3210 (pass_data_early_local_passes): ...new pass_data instance and...
3211 (make_pass_early_local_passes): ...new function.
3212 (pass_all_early_optimizations): Convert from a global struct to a
3213 subclass of gimple_opt_pass along with...
3214 (pass_data_all_early_optimizations): ...new pass_data instance and...
3215 (make_pass_all_early_optimizations): ...new function.
3216 (pass_all_optimizations): Convert from a global struct to a subclass
3217 of gimple_opt_pass along with...
3218 (pass_data_all_optimizations): ...new pass_data instance and...
3219 (make_pass_all_optimizations): ...new function.
3220 (pass_all_optimizations_g): Convert from a global struct to a subclass
3221 of gimple_opt_pass along with...
3222 (pass_data_all_optimizations_g): ...new pass_data instance and...
3223 (make_pass_all_optimizations_g): ...new function.
3224 (pass_rest_of_compilation): Convert from a global struct to a subclass
3225 of rtl_opt_pass along with...
3226 (pass_data_rest_of_compilation): ...new pass_data instance and...
3227 (make_pass_rest_of_compilation): ...new function.
3228 (pass_postreload): Convert from a global struct to a subclass of
3229 rtl_opt_pass along with...
3230 (pass_data_postreload): ...new pass_data instance and...
3231 (make_pass_postreload): ...new function.
3232 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
3233 subclass of rtl_opt_pass along with...
3234 (pass_data_gcse2): ...new pass_data instance and...
3235 (make_pass_gcse2): ...new function.
3236 * postreload.c (pass_postreload_cse): Convert from a global struct to
3237 a subclass of rtl_opt_pass along with...
3238 (pass_data_postreload_cse): ...new pass_data instance and...
3239 (make_pass_postreload_cse): ...new function.
3240 * predict.c (pass_profile): Convert from a global struct to a subclass
3241 of gimple_opt_pass along with...
3242 (pass_data_profile): ...new pass_data instance and...
3243 (make_pass_profile): ...new function.
3244 (pass_strip_predict_hints): Convert from a global struct to a subclass
3245 of gimple_opt_pass along with...
3246 (pass_data_strip_predict_hints): ...new pass_data instance and...
3247 (make_pass_strip_predict_hints): ...new function.
3248 * recog.c (pass_peephole2): Convert from a global struct to a subclass
3249 of rtl_opt_pass along with...
3250 (pass_data_peephole2): ...new pass_data instance and...
3251 (make_pass_peephole2): ...new function.
3252 (pass_split_all_insns): Convert from a global struct to a subclass of
3253 rtl_opt_pass along with...
3254 (pass_data_split_all_insns): ...new pass_data instance and...
3255 (make_pass_split_all_insns): ...new function.
3256 (pass_split_after_reload): Convert from a global struct to a subclass
3257 of rtl_opt_pass along with...
3258 (pass_data_split_after_reload): ...new pass_data instance and...
3259 (make_pass_split_after_reload): ...new function.
3260 (pass_split_before_regstack): Convert from a global struct to a
3261 subclass of rtl_opt_pass along with...
3262 (pass_data_split_before_regstack): ...new pass_data instance and...
3263 (make_pass_split_before_regstack): ...new function.
3264 (pass_split_before_sched2): Convert from a global struct to a subclass
3265 of rtl_opt_pass along with...
3266 (pass_data_split_before_sched2): ...new pass_data instance and...
3267 (make_pass_split_before_sched2): ...new function.
3268 (pass_split_for_shorten_branches): Convert from a global struct to a
3269 subclass of rtl_opt_pass along with...
3270 (pass_data_split_for_shorten_branches): ...new pass_data instance
3272 (make_pass_split_for_shorten_branches): ...new function.
3273 * ree.c (pass_ree): Convert from a global struct to a subclass of
3274 rtl_opt_pass along with...
3275 (pass_data_ree): ...new pass_data instance and...
3276 (make_pass_ree): ...new function.
3277 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
3278 subclass of rtl_opt_pass along with...
3279 (pass_data_stack_regs): ...new pass_data instance and...
3280 (make_pass_stack_regs): ...new function.
3281 (pass_stack_regs_run): Convert from a global struct to a subclass of
3282 rtl_opt_pass along with...
3283 (pass_data_stack_regs_run): ...new pass_data instance and...
3284 (make_pass_stack_regs_run): ...new function.
3285 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
3286 subclass of rtl_opt_pass along with...
3287 (pass_data_cprop_hardreg): ...new pass_data instance and...
3288 (make_pass_cprop_hardreg): ...new function.
3289 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
3290 subclass of rtl_opt_pass along with...
3291 (pass_data_reginfo_init): ...new pass_data instance and...
3292 (make_pass_reginfo_init): ...new function.
3293 * regmove.c (pass_regmove): Convert from a global struct to a subclass
3294 of rtl_opt_pass along with...
3295 (pass_data_regmove): ...new pass_data instance and...
3296 (make_pass_regmove): ...new function.
3297 * regrename.c (pass_regrename): Convert from a global struct to a
3298 subclass of rtl_opt_pass along with...
3299 (pass_data_regrename): ...new pass_data instance and...
3300 (make_pass_regrename): ...new function.
3301 * reorg.c (pass_delay_slots): Convert from a global struct to a
3302 subclass of rtl_opt_pass along with...
3303 (pass_data_delay_slots): ...new pass_data instance and...
3304 (make_pass_delay_slots): ...new function.
3305 (pass_machine_reorg): Convert from a global struct to a subclass of
3306 rtl_opt_pass along with...
3307 (pass_data_machine_reorg): ...new pass_data instance and...
3308 (make_pass_machine_reorg): ...new function.
3309 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
3310 of rtl_opt_pass along with...
3311 (pass_data_sched): ...new pass_data instance and...
3312 (make_pass_sched): ...new function.
3313 (pass_sched2): Convert from a global struct to a subclass of
3314 rtl_opt_pass along with...
3315 (pass_data_sched2): ...new pass_data instance and...
3316 (make_pass_sched2): ...new function.
3317 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
3318 to a subclass of rtl_opt_pass along with...
3319 (pass_data_stack_ptr_mod): ...new pass_data instance and...
3320 (make_pass_stack_ptr_mod): ...new function.
3321 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
3322 to a subclass of rtl_opt_pass along with...
3323 (pass_data_rtl_store_motion): ...new pass_data instance and...
3324 (make_pass_rtl_store_motion): ...new function.
3325 * tracer.c (pass_tracer): Convert from a global struct to a subclass
3326 of gimple_opt_pass along with...
3327 (pass_data_tracer): ...new pass_data instance and...
3328 (make_pass_tracer): ...new function.
3329 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
3330 to a subclass of gimple_opt_pass along with...
3331 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
3332 (make_pass_diagnose_tm_blocks): ...new function.
3333 (pass_lower_tm): Convert from a global struct to a subclass of
3334 gimple_opt_pass along with...
3335 (pass_data_lower_tm): ...new pass_data instance and...
3336 (make_pass_lower_tm): ...new function.
3337 (pass_tm_init): Convert from a global struct to a subclass of
3338 gimple_opt_pass along with...
3339 (pass_data_tm_init): ...new pass_data instance and...
3340 (make_pass_tm_init): ...new function.
3341 (pass_tm_mark): Convert from a global struct to a subclass of
3342 gimple_opt_pass along with...
3343 (pass_data_tm_mark): ...new pass_data instance and...
3344 (make_pass_tm_mark): ...new function.
3345 (pass_tm_edges): Convert from a global struct to a subclass of
3346 gimple_opt_pass along with...
3347 (pass_data_tm_edges): ...new pass_data instance and...
3348 (make_pass_tm_edges): ...new function.
3349 (pass_tm_memopt): Convert from a global struct to a subclass of
3350 gimple_opt_pass along with...
3351 (pass_data_tm_memopt): ...new pass_data instance and...
3352 (make_pass_tm_memopt): ...new function.
3353 (pass_ipa_tm): Convert from a global struct to a subclass of
3354 simple_ipa_opt_pass along with...
3355 (pass_data_ipa_tm): ...new pass_data instance and...
3356 (make_pass_ipa_tm): ...new function.
3357 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
3358 subclass of gimple_opt_pass along with...
3359 (pass_data_call_cdce): ...new pass_data instance and...
3360 (make_pass_call_cdce): ...new function.
3361 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
3362 subclass of gimple_opt_pass along with...
3363 (pass_data_build_cfg): ...new pass_data instance and...
3364 (make_pass_build_cfg): ...new function.
3365 (pass_split_crit_edges): Convert from a global struct to a subclass of
3366 gimple_opt_pass along with...
3367 (pass_data_split_crit_edges): ...new pass_data instance and...
3368 (make_pass_split_crit_edges): ...new function.
3369 (pass_warn_function_return): Convert from a global struct to a
3370 subclass of gimple_opt_pass along with...
3371 (pass_data_warn_function_return): ...new pass_data instance and...
3372 (make_pass_warn_function_return): ...new function.
3373 (pass_warn_function_noreturn): Convert from a global struct to a
3374 subclass of gimple_opt_pass along with...
3375 (pass_data_warn_function_noreturn): ...new pass_data instance and...
3376 (make_pass_warn_function_noreturn): ...new function.
3377 (pass_warn_unused_result): Convert from a global struct to a subclass
3378 of gimple_opt_pass along with...
3379 (pass_data_warn_unused_result): ...new pass_data instance and...
3380 (make_pass_warn_unused_result): ...new function.
3381 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
3382 a subclass of gimple_opt_pass along with...
3383 (pass_data_merge_phi): ...new pass_data instance and...
3384 (make_pass_merge_phi): ...new function.
3385 * tree-complex.c (pass_lower_complex): Convert from a global struct to
3386 a subclass of gimple_opt_pass along with...
3387 (pass_data_lower_complex): ...new pass_data instance and...
3388 (make_pass_lower_complex): ...new function.
3389 (pass_lower_complex_O0): Convert from a global struct to a subclass of
3390 gimple_opt_pass along with...
3391 (pass_data_lower_complex_O0): ...new pass_data instance and...
3392 (make_pass_lower_complex_O0): ...new function.
3393 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
3394 subclass of gimple_opt_pass along with...
3395 (pass_data_lower_eh): ...new pass_data instance and...
3396 (make_pass_lower_eh): ...new function.
3397 (pass_refactor_eh): Convert from a global struct to a subclass of
3398 gimple_opt_pass along with...
3399 (pass_data_refactor_eh): ...new pass_data instance and...
3400 (make_pass_refactor_eh): ...new function.
3401 (pass_lower_resx): Convert from a global struct to a subclass of
3402 gimple_opt_pass along with...
3403 (pass_data_lower_resx): ...new pass_data instance and...
3404 (make_pass_lower_resx): ...new function.
3405 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
3406 of gimple_opt_pass along with...
3407 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
3408 (make_pass_lower_eh_dispatch): ...new function.
3409 (pass_cleanup_eh): Convert from a global struct to a subclass of
3410 gimple_opt_pass along with...
3411 (pass_data_cleanup_eh): ...new pass_data instance and...
3412 (make_pass_cleanup_eh): ...new function.
3413 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
3414 to a subclass of simple_ipa_opt_pass along with...
3415 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
3416 (make_pass_ipa_lower_emutls): ...new function.
3417 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
3418 a subclass of gimple_opt_pass along with...
3419 (pass_data_if_conversion): ...new pass_data instance and...
3420 (make_pass_if_conversion): ...new function.
3421 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
3422 subclass of gimple_opt_pass along with...
3423 (pass_data_build_ssa): ...new pass_data instance and...
3424 (make_pass_build_ssa): ...new function.
3425 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
3426 global struct to a subclass of gimple_opt_pass along with...
3427 (pass_data_loop_distribution): ...new pass_data instance and...
3428 (make_pass_loop_distribution): ...new function.
3429 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
3430 subclass of gimple_opt_pass along with...
3431 (pass_data_mudflap_1): ...new pass_data instance and...
3432 (make_pass_mudflap_1): ...new function.
3433 (pass_mudflap_2): Convert from a global struct to a subclass of
3434 gimple_opt_pass along with...
3435 (pass_data_mudflap_2): ...new pass_data instance and...
3436 (make_pass_mudflap_2): ...new function.
3437 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
3438 subclass of gimple_opt_pass along with...
3439 (pass_data_mudflap_1): ...new pass_data instance and...
3440 (make_pass_mudflap_1): ...new function.
3441 (pass_mudflap_2): Convert from a global struct to a subclass of
3442 gimple_opt_pass along with...
3443 (pass_data_mudflap_2): ...new pass_data instance and...
3444 (make_pass_mudflap_2): ...new function.
3445 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
3446 gimple_opt_pass along with...
3447 (pass_data_nrv): ...new pass_data instance and...
3448 (make_pass_nrv): ...new function.
3449 (pass_return_slot): Convert from a global struct to a subclass of
3450 gimple_opt_pass along with...
3451 (pass_data_return_slot): ...new pass_data instance and...
3452 (make_pass_return_slot): ...new function.
3453 * tree-object-size.c (pass_object_sizes): Convert from a global struct
3454 to a subclass of gimple_opt_pass along with...
3455 (pass_data_object_sizes): ...new pass_data instance and...
3456 (make_pass_object_sizes): ...new function.
3457 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
3458 global struct to a subclass of gimple_opt_pass along with...
3459 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
3461 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3462 (pass_fixup_cfg): Convert from a global struct to a subclass of
3463 gimple_opt_pass along with...
3464 (pass_data_fixup_cfg): ...new pass_data instance and...
3465 (make_pass_fixup_cfg): ...new function.
3466 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
3467 (make_pass_mudflap_1): ...new function.
3468 (pass_mudflap_2): Replace declaration with that of...
3469 (make_pass_mudflap_2): ...new function.
3470 (pass_asan): Replace declaration with that of...
3471 (make_pass_asan): ...new function.
3472 (pass_asan_O0): Replace declaration with that of...
3473 (make_pass_asan_O0): ...new function.
3474 (pass_tsan): Replace declaration with that of...
3475 (make_pass_tsan): ...new function.
3476 (pass_tsan_O0): Replace declaration with that of...
3477 (make_pass_tsan_O0): ...new function.
3478 (pass_lower_cf): Replace declaration with that of...
3479 (make_pass_lower_cf): ...new function.
3480 (pass_refactor_eh): Replace declaration with that of...
3481 (make_pass_refactor_eh): ...new function.
3482 (pass_lower_eh): Replace declaration with that of...
3483 (make_pass_lower_eh): ...new function.
3484 (pass_lower_eh_dispatch): Replace declaration with that of...
3485 (make_pass_lower_eh_dispatch): ...new function.
3486 (pass_lower_resx): Replace declaration with that of...
3487 (make_pass_lower_resx): ...new function.
3488 (pass_build_cfg): Replace declaration with that of...
3489 (make_pass_build_cfg): ...new function.
3490 (pass_early_tree_profile): Replace declaration with that of...
3491 (make_pass_early_tree_profile): ...new function.
3492 (pass_cleanup_eh): Replace declaration with that of...
3493 (make_pass_cleanup_eh): ...new function.
3494 (pass_sra): Replace declaration with that of...
3495 (make_pass_sra): ...new function.
3496 (pass_sra_early): Replace declaration with that of...
3497 (make_pass_sra_early): ...new function.
3498 (pass_early_ipa_sra): Replace declaration with that of...
3499 (make_pass_early_ipa_sra): ...new function.
3500 (pass_tail_recursion): Replace declaration with that of...
3501 (make_pass_tail_recursion): ...new function.
3502 (pass_tail_calls): Replace declaration with that of...
3503 (make_pass_tail_calls): ...new function.
3504 (pass_tree_loop): Replace declaration with that of...
3505 (make_pass_tree_loop): ...new function.
3506 (pass_tree_loop_init): Replace declaration with that of...
3507 (make_pass_tree_loop_init): ...new function.
3508 (pass_lim): Replace declaration with that of...
3509 (make_pass_lim): ...new function.
3510 (pass_tree_unswitch): Replace declaration with that of...
3511 (make_pass_tree_unswitch): ...new function.
3512 (pass_predcom): Replace declaration with that of...
3513 (make_pass_predcom): ...new function.
3514 (pass_iv_canon): Replace declaration with that of...
3515 (make_pass_iv_canon): ...new function.
3516 (pass_scev_cprop): Replace declaration with that of...
3517 (make_pass_scev_cprop): ...new function.
3518 (pass_empty_loop): Replace declaration with that of...
3519 (make_pass_empty_loop): ...new function.
3520 (pass_record_bounds): Replace declaration with that of...
3521 (make_pass_record_bounds): ...new function.
3522 (pass_graphite): Replace declaration with that of...
3523 (make_pass_graphite): ...new function.
3524 (pass_graphite_transforms): Replace declaration with that of...
3525 (make_pass_graphite_transforms): ...new function.
3526 (pass_if_conversion): Replace declaration with that of...
3527 (make_pass_if_conversion): ...new function.
3528 (pass_loop_distribution): Replace declaration with that of...
3529 (make_pass_loop_distribution): ...new function.
3530 (pass_vectorize): Replace declaration with that of...
3531 (make_pass_vectorize): ...new function.
3532 (pass_slp_vectorize): Replace declaration with that of...
3533 (make_pass_slp_vectorize): ...new function.
3534 (pass_complete_unroll): Replace declaration with that of...
3535 (make_pass_complete_unroll): ...new function.
3536 (pass_complete_unrolli): Replace declaration with that of...
3537 (make_pass_complete_unrolli): ...new function.
3538 (pass_parallelize_loops): Replace declaration with that of...
3539 (make_pass_parallelize_loops): ...new function.
3540 (pass_loop_prefetch): Replace declaration with that of...
3541 (make_pass_loop_prefetch): ...new function.
3542 (pass_iv_optimize): Replace declaration with that of...
3543 (make_pass_iv_optimize): ...new function.
3544 (pass_tree_loop_done): Replace declaration with that of...
3545 (make_pass_tree_loop_done): ...new function.
3546 (pass_ch): Replace declaration with that of...
3547 (make_pass_ch): ...new function.
3548 (pass_ccp): Replace declaration with that of...
3549 (make_pass_ccp): ...new function.
3550 (pass_phi_only_cprop): Replace declaration with that of...
3551 (make_pass_phi_only_cprop): ...new function.
3552 (pass_build_ssa): Replace declaration with that of...
3553 (make_pass_build_ssa): ...new function.
3554 (pass_build_alias): Replace declaration with that of...
3555 (make_pass_build_alias): ...new function.
3556 (pass_build_ealias): Replace declaration with that of...
3557 (make_pass_build_ealias): ...new function.
3558 (pass_dominator): Replace declaration with that of...
3559 (make_pass_dominator): ...new function.
3560 (pass_dce): Replace declaration with that of...
3561 (make_pass_dce): ...new function.
3562 (pass_dce_loop): Replace declaration with that of...
3563 (make_pass_dce_loop): ...new function.
3564 (pass_cd_dce): Replace declaration with that of...
3565 (make_pass_cd_dce): ...new function.
3566 (pass_call_cdce): Replace declaration with that of...
3567 (make_pass_call_cdce): ...new function.
3568 (pass_merge_phi): Replace declaration with that of...
3569 (make_pass_merge_phi): ...new function.
3570 (pass_split_crit_edges): Replace declaration with that of...
3571 (make_pass_split_crit_edges): ...new function.
3572 (pass_pre): Replace declaration with that of...
3573 (make_pass_pre): ...new function.
3574 (pass_profile): Replace declaration with that of...
3575 (make_pass_profile): ...new function.
3576 (pass_strip_predict_hints): Replace declaration with that of...
3577 (make_pass_strip_predict_hints): ...new function.
3578 (pass_lower_complex_O0): Replace declaration with that of...
3579 (make_pass_lower_complex_O0): ...new function.
3580 (pass_lower_complex): Replace declaration with that of...
3581 (make_pass_lower_complex): ...new function.
3582 (pass_lower_vector): Replace declaration with that of...
3583 (make_pass_lower_vector): ...new function.
3584 (pass_lower_vector_ssa): Replace declaration with that of...
3585 (make_pass_lower_vector_ssa): ...new function.
3586 (pass_lower_omp): Replace declaration with that of...
3587 (make_pass_lower_omp): ...new function.
3588 (pass_diagnose_omp_blocks): Replace declaration with that of...
3589 (make_pass_diagnose_omp_blocks): ...new function.
3590 (pass_expand_omp): Replace declaration with that of...
3591 (make_pass_expand_omp): ...new function.
3592 (pass_expand_omp_ssa): Replace declaration with that of...
3593 (make_pass_expand_omp_ssa): ...new function.
3594 (pass_object_sizes): Replace declaration with that of...
3595 (make_pass_object_sizes): ...new function.
3596 (pass_strlen): Replace declaration with that of...
3597 (make_pass_strlen): ...new function.
3598 (pass_fold_builtins): Replace declaration with that of...
3599 (make_pass_fold_builtins): ...new function.
3600 (pass_stdarg): Replace declaration with that of...
3601 (make_pass_stdarg): ...new function.
3602 (pass_early_warn_uninitialized): Replace declaration with that of...
3603 (make_pass_early_warn_uninitialized): ...new function.
3604 (pass_late_warn_uninitialized): Replace declaration with that of...
3605 (make_pass_late_warn_uninitialized): ...new function.
3606 (pass_cse_reciprocals): Replace declaration with that of...
3607 (make_pass_cse_reciprocals): ...new function.
3608 (pass_cse_sincos): Replace declaration with that of...
3609 (make_pass_cse_sincos): ...new function.
3610 (pass_optimize_bswap): Replace declaration with that of...
3611 (make_pass_optimize_bswap): ...new function.
3612 (pass_optimize_widening_mul): Replace declaration with that of...
3613 (make_pass_optimize_widening_mul): ...new function.
3614 (pass_warn_function_return): Replace declaration with that of...
3615 (make_pass_warn_function_return): ...new function.
3616 (pass_warn_function_noreturn): Replace declaration with that of...
3617 (make_pass_warn_function_noreturn): ...new function.
3618 (pass_cselim): Replace declaration with that of...
3619 (make_pass_cselim): ...new function.
3620 (pass_phiopt): Replace declaration with that of...
3621 (make_pass_phiopt): ...new function.
3622 (pass_forwprop): Replace declaration with that of...
3623 (make_pass_forwprop): ...new function.
3624 (pass_phiprop): Replace declaration with that of...
3625 (make_pass_phiprop): ...new function.
3626 (pass_tree_ifcombine): Replace declaration with that of...
3627 (make_pass_tree_ifcombine): ...new function.
3628 (pass_dse): Replace declaration with that of...
3629 (make_pass_dse): ...new function.
3630 (pass_nrv): Replace declaration with that of...
3631 (make_pass_nrv): ...new function.
3632 (pass_rename_ssa_copies): Replace declaration with that of...
3633 (make_pass_rename_ssa_copies): ...new function.
3634 (pass_sink_code): Replace declaration with that of...
3635 (make_pass_sink_code): ...new function.
3636 (pass_fre): Replace declaration with that of...
3637 (make_pass_fre): ...new function.
3638 (pass_check_data_deps): Replace declaration with that of...
3639 (make_pass_check_data_deps): ...new function.
3640 (pass_copy_prop): Replace declaration with that of...
3641 (make_pass_copy_prop): ...new function.
3642 (pass_vrp): Replace declaration with that of...
3643 (make_pass_vrp): ...new function.
3644 (pass_uncprop): Replace declaration with that of...
3645 (make_pass_uncprop): ...new function.
3646 (pass_return_slot): Replace declaration with that of...
3647 (make_pass_return_slot): ...new function.
3648 (pass_reassoc): Replace declaration with that of...
3649 (make_pass_reassoc): ...new function.
3650 (pass_rebuild_cgraph_edges): Replace declaration with that of...
3651 (make_pass_rebuild_cgraph_edges): ...new function.
3652 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
3653 (make_pass_remove_cgraph_callee_edges): ...new function.
3654 (pass_build_cgraph_edges): Replace declaration with that of...
3655 (make_pass_build_cgraph_edges): ...new function.
3656 (pass_local_pure_const): Replace declaration with that of...
3657 (make_pass_local_pure_const): ...new function.
3658 (pass_tracer): Replace declaration with that of...
3659 (make_pass_tracer): ...new function.
3660 (pass_warn_unused_result): Replace declaration with that of...
3661 (make_pass_warn_unused_result): ...new function.
3662 (pass_diagnose_tm_blocks): Replace declaration with that of...
3663 (make_pass_diagnose_tm_blocks): ...new function.
3664 (pass_lower_tm): Replace declaration with that of...
3665 (make_pass_lower_tm): ...new function.
3666 (pass_tm_init): Replace declaration with that of...
3667 (make_pass_tm_init): ...new function.
3668 (pass_tm_mark): Replace declaration with that of...
3669 (make_pass_tm_mark): ...new function.
3670 (pass_tm_memopt): Replace declaration with that of...
3671 (make_pass_tm_memopt): ...new function.
3672 (pass_tm_edges): Replace declaration with that of...
3673 (make_pass_tm_edges): ...new function.
3674 (pass_split_functions): Replace declaration with that of...
3675 (make_pass_split_functions): ...new function.
3676 (pass_feedback_split_functions): Replace declaration with that of...
3677 (make_pass_feedback_split_functions): ...new function.
3678 (pass_strength_reduction): Replace declaration with that of...
3679 (make_pass_strength_reduction): ...new function.
3680 (pass_ipa_lower_emutls): Replace declaration with that of...
3681 (make_pass_ipa_lower_emutls): ...new function.
3682 (pass_ipa_function_and_variable_visibility): Replace declaration with
3684 (make_pass_ipa_function_and_variable_visibility): ...new function.
3685 (pass_ipa_tree_profile): Replace declaration with that of...
3686 (make_pass_ipa_tree_profile): ...new function.
3687 (pass_early_local_passes): Replace declaration with that of...
3688 (make_pass_early_local_passes): ...new function.
3689 (pass_ipa_whole_program_visibility): Replace declaration with that
3691 (make_pass_ipa_whole_program_visibility): ...new function.
3692 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3693 (make_pass_ipa_lto_gimple_out): ...new function.
3694 (pass_ipa_increase_alignment): Replace declaration with that of...
3695 (make_pass_ipa_increase_alignment): ...new function.
3696 (pass_ipa_inline): Replace declaration with that of...
3697 (make_pass_ipa_inline): ...new function.
3698 (pass_ipa_free_lang_data): Replace declaration with that of...
3699 (make_pass_ipa_free_lang_data): ...new function.
3700 (pass_ipa_free_inline_summary): Replace declaration with that of...
3701 (make_pass_ipa_free_inline_summary): ...new function.
3702 (pass_ipa_cp): Replace declaration with that of...
3703 (make_pass_ipa_cp): ...new function.
3704 (pass_ipa_reference): Replace declaration with that of...
3705 (make_pass_ipa_reference): ...new function.
3706 (pass_ipa_pure_const): Replace declaration with that of...
3707 (make_pass_ipa_pure_const): ...new function.
3708 (pass_ipa_pta): Replace declaration with that of...
3709 (make_pass_ipa_pta): ...new function.
3710 (pass_ipa_lto_finish_out): Replace declaration with that of...
3711 (make_pass_ipa_lto_finish_out): ...new function.
3712 (pass_ipa_tm): Replace declaration with that of...
3713 (make_pass_ipa_tm): ...new function.
3714 (pass_ipa_profile): Replace declaration with that of...
3715 (make_pass_ipa_profile): ...new function.
3716 (pass_ipa_cdtor_merge): Replace declaration with that of...
3717 (make_pass_ipa_cdtor_merge): ...new function.
3718 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3720 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3721 (pass_init_datastructures): Replace declaration with that of...
3722 (make_pass_init_datastructures): ...new function.
3723 (pass_fixup_cfg): Replace declaration with that of...
3724 (make_pass_fixup_cfg): ...new function.
3725 (pass_expand): Replace declaration with that of...
3726 (make_pass_expand): ...new function.
3727 (pass_instantiate_virtual_regs): Replace declaration with that of...
3728 (make_pass_instantiate_virtual_regs): ...new function.
3729 (pass_rtl_fwprop): Replace declaration with that of...
3730 (make_pass_rtl_fwprop): ...new function.
3731 (pass_rtl_fwprop_addr): Replace declaration with that of...
3732 (make_pass_rtl_fwprop_addr): ...new function.
3733 (pass_jump): Replace declaration with that of...
3734 (make_pass_jump): ...new function.
3735 (pass_jump2): Replace declaration with that of...
3736 (make_pass_jump2): ...new function.
3737 (pass_lower_subreg): Replace declaration with that of...
3738 (make_pass_lower_subreg): ...new function.
3739 (pass_cse): Replace declaration with that of...
3740 (make_pass_cse): ...new function.
3741 (pass_fast_rtl_dce): Replace declaration with that of...
3742 (make_pass_fast_rtl_dce): ...new function.
3743 (pass_ud_rtl_dce): Replace declaration with that of...
3744 (make_pass_ud_rtl_dce): ...new function.
3745 (pass_rtl_dce): Replace declaration with that of...
3746 (make_pass_rtl_dce): ...new function.
3747 (pass_rtl_dse1): Replace declaration with that of...
3748 (make_pass_rtl_dse1): ...new function.
3749 (pass_rtl_dse2): Replace declaration with that of...
3750 (make_pass_rtl_dse2): ...new function.
3751 (pass_rtl_dse3): Replace declaration with that of...
3752 (make_pass_rtl_dse3): ...new function.
3753 (pass_rtl_cprop): Replace declaration with that of...
3754 (make_pass_rtl_cprop): ...new function.
3755 (pass_rtl_pre): Replace declaration with that of...
3756 (make_pass_rtl_pre): ...new function.
3757 (pass_rtl_hoist): Replace declaration with that of...
3758 (make_pass_rtl_hoist): ...new function.
3759 (pass_rtl_store_motion): Replace declaration with that of...
3760 (make_pass_rtl_store_motion): ...new function.
3761 (pass_cse_after_global_opts): Replace declaration with that of...
3762 (make_pass_cse_after_global_opts): ...new function.
3763 (pass_rtl_ifcvt): Replace declaration with that of...
3764 (make_pass_rtl_ifcvt): ...new function.
3765 (pass_into_cfg_layout_mode): Replace declaration with that of...
3766 (make_pass_into_cfg_layout_mode): ...new function.
3767 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3768 (make_pass_outof_cfg_layout_mode): ...new function.
3769 (pass_loop2): Replace declaration with that of...
3770 (make_pass_loop2): ...new function.
3771 (pass_rtl_loop_init): Replace declaration with that of...
3772 (make_pass_rtl_loop_init): ...new function.
3773 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3774 (make_pass_rtl_move_loop_invariants): ...new function.
3775 (pass_rtl_unswitch): Replace declaration with that of...
3776 (make_pass_rtl_unswitch): ...new function.
3777 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3778 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3779 (pass_rtl_doloop): Replace declaration with that of...
3780 (make_pass_rtl_doloop): ...new function.
3781 (pass_rtl_loop_done): Replace declaration with that of...
3782 (make_pass_rtl_loop_done): ...new function.
3783 (pass_web): Replace declaration with that of...
3784 (make_pass_web): ...new function.
3785 (pass_cse2): Replace declaration with that of...
3786 (make_pass_cse2): ...new function.
3787 (pass_df_initialize_opt): Replace declaration with that of...
3788 (make_pass_df_initialize_opt): ...new function.
3789 (pass_df_initialize_no_opt): Replace declaration with that of...
3790 (make_pass_df_initialize_no_opt): ...new function.
3791 (pass_reginfo_init): Replace declaration with that of...
3792 (make_pass_reginfo_init): ...new function.
3793 (pass_inc_dec): Replace declaration with that of...
3794 (make_pass_inc_dec): ...new function.
3795 (pass_stack_ptr_mod): Replace declaration with that of...
3796 (make_pass_stack_ptr_mod): ...new function.
3797 (pass_initialize_regs): Replace declaration with that of...
3798 (make_pass_initialize_regs): ...new function.
3799 (pass_combine): Replace declaration with that of...
3800 (make_pass_combine): ...new function.
3801 (pass_if_after_combine): Replace declaration with that of...
3802 (make_pass_if_after_combine): ...new function.
3803 (pass_ree): Replace declaration with that of...
3804 (make_pass_ree): ...new function.
3805 (pass_partition_blocks): Replace declaration with that of...
3806 (make_pass_partition_blocks): ...new function.
3807 (pass_match_asm_constraints): Replace declaration with that of...
3808 (make_pass_match_asm_constraints): ...new function.
3809 (pass_regmove): Replace declaration with that of...
3810 (make_pass_regmove): ...new function.
3811 (pass_split_all_insns): Replace declaration with that of...
3812 (make_pass_split_all_insns): ...new function.
3813 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3814 (make_pass_fast_rtl_byte_dce): ...new function.
3815 (pass_lower_subreg2): Replace declaration with that of...
3816 (make_pass_lower_subreg2): ...new function.
3817 (pass_mode_switching): Replace declaration with that of...
3818 (make_pass_mode_switching): ...new function.
3819 (pass_sms): Replace declaration with that of...
3820 (make_pass_sms): ...new function.
3821 (pass_sched): Replace declaration with that of...
3822 (make_pass_sched): ...new function.
3823 (pass_ira): Replace declaration with that of...
3824 (make_pass_ira): ...new function.
3825 (pass_reload): Replace declaration with that of...
3826 (make_pass_reload): ...new function.
3827 (pass_clean_state): Replace declaration with that of...
3828 (make_pass_clean_state): ...new function.
3829 (pass_branch_prob): Replace declaration with that of...
3830 (make_pass_branch_prob): ...new function.
3831 (pass_value_profile_transformations): Replace declaration with that
3833 (make_pass_value_profile_transformations): ...new function.
3834 (pass_postreload_cse): Replace declaration with that of...
3835 (make_pass_postreload_cse): ...new function.
3836 (pass_gcse2): Replace declaration with that of...
3837 (make_pass_gcse2): ...new function.
3838 (pass_split_after_reload): Replace declaration with that of...
3839 (make_pass_split_after_reload): ...new function.
3840 (pass_branch_target_load_optimize1): Replace declaration with that
3842 (make_pass_branch_target_load_optimize1): ...new function.
3843 (pass_thread_prologue_and_epilogue): Replace declaration with that
3845 (make_pass_thread_prologue_and_epilogue): ...new function.
3846 (pass_stack_adjustments): Replace declaration with that of...
3847 (make_pass_stack_adjustments): ...new function.
3848 (pass_peephole2): Replace declaration with that of...
3849 (make_pass_peephole2): ...new function.
3850 (pass_if_after_reload): Replace declaration with that of...
3851 (make_pass_if_after_reload): ...new function.
3852 (pass_regrename): Replace declaration with that of...
3853 (make_pass_regrename): ...new function.
3854 (pass_cprop_hardreg): Replace declaration with that of...
3855 (make_pass_cprop_hardreg): ...new function.
3856 (pass_reorder_blocks): Replace declaration with that of...
3857 (make_pass_reorder_blocks): ...new function.
3858 (pass_branch_target_load_optimize2): Replace declaration with that
3860 (make_pass_branch_target_load_optimize2): ...new function.
3861 (pass_leaf_regs): Replace declaration with that of...
3862 (make_pass_leaf_regs): ...new function.
3863 (pass_split_before_sched2): Replace declaration with that of...
3864 (make_pass_split_before_sched2): ...new function.
3865 (pass_compare_elim_after_reload): Replace declaration with that of...
3866 (make_pass_compare_elim_after_reload): ...new function.
3867 (pass_sched2): Replace declaration with that of...
3868 (make_pass_sched2): ...new function.
3869 (pass_stack_regs): Replace declaration with that of...
3870 (make_pass_stack_regs): ...new function.
3871 (pass_stack_regs_run): Replace declaration with that of...
3872 (make_pass_stack_regs_run): ...new function.
3873 (pass_df_finish): Replace declaration with that of...
3874 (make_pass_df_finish): ...new function.
3875 (pass_compute_alignments): Replace declaration with that of...
3876 (make_pass_compute_alignments): ...new function.
3877 (pass_duplicate_computed_gotos): Replace declaration with that of...
3878 (make_pass_duplicate_computed_gotos): ...new function.
3879 (pass_variable_tracking): Replace declaration with that of...
3880 (make_pass_variable_tracking): ...new function.
3881 (pass_free_cfg): Replace declaration with that of...
3882 (make_pass_free_cfg): ...new function.
3883 (pass_machine_reorg): Replace declaration with that of...
3884 (make_pass_machine_reorg): ...new function.
3885 (pass_cleanup_barriers): Replace declaration with that of...
3886 (make_pass_cleanup_barriers): ...new function.
3887 (pass_delay_slots): Replace declaration with that of...
3888 (make_pass_delay_slots): ...new function.
3889 (pass_split_for_shorten_branches): Replace declaration with that of...
3890 (make_pass_split_for_shorten_branches): ...new function.
3891 (pass_split_before_regstack): Replace declaration with that of...
3892 (make_pass_split_before_regstack): ...new function.
3893 (pass_convert_to_eh_region_ranges): Replace declaration with that
3895 (make_pass_convert_to_eh_region_ranges): ...new function.
3896 (pass_shorten_branches): Replace declaration with that of...
3897 (make_pass_shorten_branches): ...new function.
3898 (pass_set_nothrow_function_flags): Replace declaration with that of...
3899 (make_pass_set_nothrow_function_flags): ...new function.
3900 (pass_dwarf2_frame): Replace declaration with that of...
3901 (make_pass_dwarf2_frame): ...new function.
3902 (pass_final): Replace declaration with that of...
3903 (make_pass_final): ...new function.
3904 (pass_rtl_seqabstr): Replace declaration with that of...
3905 (make_pass_rtl_seqabstr): ...new function.
3906 (pass_release_ssa_names): Replace declaration with that of...
3907 (make_pass_release_ssa_names): ...new function.
3908 (pass_early_inline): Replace declaration with that of...
3909 (make_pass_early_inline): ...new function.
3910 (pass_inline_parameters): Replace declaration with that of...
3911 (make_pass_inline_parameters): ...new function.
3912 (pass_update_address_taken): Replace declaration with that of...
3913 (make_pass_update_address_taken): ...new function.
3914 (pass_convert_switch): Replace declaration with that of...
3915 (make_pass_convert_switch): ...new function.
3916 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
3917 to a subclass of simple_ipa_opt_pass along with...
3918 (pass_data_ipa_tree_profile): ...new pass_data instance and...
3919 (make_pass_ipa_tree_profile): ...new function.
3920 * tree-sra.c (pass_sra_early): Convert from a global struct to a
3921 subclass of gimple_opt_pass along with...
3922 (pass_data_sra_early): ...new pass_data instance and...
3923 (make_pass_sra_early): ...new function.
3924 (pass_sra): Convert from a global struct to a subclass of
3925 gimple_opt_pass along with...
3926 (pass_data_sra): ...new pass_data instance and...
3927 (make_pass_sra): ...new function.
3928 (pass_early_ipa_sra): Convert from a global struct to a subclass of
3929 gimple_opt_pass along with...
3930 (pass_data_early_ipa_sra): ...new pass_data instance and...
3931 (make_pass_early_ipa_sra): ...new function.
3932 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
3933 subclass of gimple_opt_pass along with...
3934 (pass_data_ccp): ...new pass_data instance and...
3935 (make_pass_ccp): ...new function.
3936 (pass_fold_builtins): Convert from a global struct to a subclass of
3937 gimple_opt_pass along with...
3938 (pass_data_fold_builtins): ...new pass_data instance and...
3939 (make_pass_fold_builtins): ...new function.
3940 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
3941 subclass of gimple_opt_pass along with...
3942 (pass_data_copy_prop): ...new pass_data instance and...
3943 (make_pass_copy_prop): ...new function.
3944 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
3945 global struct to a subclass of gimple_opt_pass along with...
3946 (pass_data_rename_ssa_copies): ...new pass_data instance and...
3947 (make_pass_rename_ssa_copies): ...new function.
3948 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
3949 subclass of gimple_opt_pass along with...
3950 (pass_data_dce): ...new pass_data instance and...
3951 (make_pass_dce): ...new function.
3952 (pass_dce_loop): Convert from a global struct to a subclass of
3953 gimple_opt_pass along with...
3954 (pass_data_dce_loop): ...new pass_data instance and...
3955 (make_pass_dce_loop): ...new function.
3956 (pass_cd_dce): Convert from a global struct to a subclass of
3957 gimple_opt_pass along with...
3958 (pass_data_cd_dce): ...new pass_data instance and...
3959 (make_pass_cd_dce): ...new function.
3960 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
3961 subclass of gimple_opt_pass along with...
3962 (pass_data_dominator): ...new pass_data instance and...
3963 (make_pass_dominator): ...new function.
3964 (pass_phi_only_cprop): Convert from a global struct to a subclass of
3965 gimple_opt_pass along with...
3966 (pass_data_phi_only_cprop): ...new pass_data instance and...
3967 (make_pass_phi_only_cprop): ...new function.
3968 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
3969 subclass of gimple_opt_pass along with...
3970 (pass_data_dse): ...new pass_data instance and...
3971 (make_pass_dse): ...new function.
3972 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
3973 a subclass of gimple_opt_pass along with...
3974 (pass_data_forwprop): ...new pass_data instance and...
3975 (make_pass_forwprop): ...new function.
3976 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
3977 struct to a subclass of gimple_opt_pass along with...
3978 (pass_data_tree_ifcombine): ...new pass_data instance and...
3979 (make_pass_tree_ifcombine): ...new function.
3980 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
3981 subclass of gimple_opt_pass along with...
3982 (pass_data_ch): ...new pass_data instance and...
3983 (make_pass_ch): ...new function.
3984 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
3985 subclass of gimple_opt_pass along with...
3986 (pass_data_tree_loop): ...new pass_data instance and...
3987 (make_pass_tree_loop): ...new function.
3988 (pass_tree_loop_init): Convert from a global struct to a subclass of
3989 gimple_opt_pass along with...
3990 (pass_data_tree_loop_init): ...new pass_data instance and...
3991 (make_pass_tree_loop_init): ...new function.
3992 (pass_lim): Convert from a global struct to a subclass of
3993 gimple_opt_pass along with...
3994 (pass_data_lim): ...new pass_data instance and...
3995 (make_pass_lim): ...new function.
3996 (pass_tree_unswitch): Convert from a global struct to a subclass of
3997 gimple_opt_pass along with...
3998 (pass_data_tree_unswitch): ...new pass_data instance and...
3999 (make_pass_tree_unswitch): ...new function.
4000 (pass_predcom): Convert from a global struct to a subclass of
4001 gimple_opt_pass along with...
4002 (pass_data_predcom): ...new pass_data instance and...
4003 (make_pass_predcom): ...new function.
4004 (pass_vectorize): Convert from a global struct to a subclass of
4005 gimple_opt_pass along with...
4006 (pass_data_vectorize): ...new pass_data instance and...
4007 (make_pass_vectorize): ...new function.
4008 (pass_graphite): Convert from a global struct to a subclass of
4009 gimple_opt_pass along with...
4010 (pass_data_graphite): ...new pass_data instance and...
4011 (make_pass_graphite): ...new function.
4012 (pass_graphite_transforms): Convert from a global struct to a subclass
4013 of gimple_opt_pass along with...
4014 (pass_data_graphite_transforms): ...new pass_data instance and...
4015 (make_pass_graphite_transforms): ...new function.
4016 (pass_check_data_deps): Convert from a global struct to a subclass of
4017 gimple_opt_pass along with...
4018 (pass_data_check_data_deps): ...new pass_data instance and...
4019 (make_pass_check_data_deps): ...new function.
4020 (pass_iv_canon): Convert from a global struct to a subclass of
4021 gimple_opt_pass along with...
4022 (pass_data_iv_canon): ...new pass_data instance and...
4023 (make_pass_iv_canon): ...new function.
4024 (pass_scev_cprop): Convert from a global struct to a subclass of
4025 gimple_opt_pass along with...
4026 (pass_data_scev_cprop): ...new pass_data instance and...
4027 (make_pass_scev_cprop): ...new function.
4028 (pass_record_bounds): Convert from a global struct to a subclass of
4029 gimple_opt_pass along with...
4030 (pass_data_record_bounds): ...new pass_data instance and...
4031 (make_pass_record_bounds): ...new function.
4032 (pass_complete_unroll): Convert from a global struct to a subclass of
4033 gimple_opt_pass along with...
4034 (pass_data_complete_unroll): ...new pass_data instance and...
4035 (make_pass_complete_unroll): ...new function.
4036 (pass_complete_unrolli): Convert from a global struct to a subclass of
4037 gimple_opt_pass along with...
4038 (pass_data_complete_unrolli): ...new pass_data instance and...
4039 (make_pass_complete_unrolli): ...new function.
4040 (pass_parallelize_loops): Convert from a global struct to a subclass
4041 of gimple_opt_pass along with...
4042 (pass_data_parallelize_loops): ...new pass_data instance and...
4043 (make_pass_parallelize_loops): ...new function.
4044 (pass_loop_prefetch): Convert from a global struct to a subclass of
4045 gimple_opt_pass along with...
4046 (pass_data_loop_prefetch): ...new pass_data instance and...
4047 (make_pass_loop_prefetch): ...new function.
4048 (pass_iv_optimize): Convert from a global struct to a subclass of
4049 gimple_opt_pass along with...
4050 (pass_data_iv_optimize): ...new pass_data instance and...
4051 (make_pass_iv_optimize): ...new function.
4052 (pass_tree_loop_done): Convert from a global struct to a subclass of
4053 gimple_opt_pass along with...
4054 (pass_data_tree_loop_done): ...new pass_data instance and...
4055 (make_pass_tree_loop_done): ...new function.
4056 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
4057 struct to a subclass of gimple_opt_pass along with...
4058 (pass_data_cse_reciprocals): ...new pass_data instance and...
4059 (make_pass_cse_reciprocals): ...new function.
4060 (pass_cse_sincos): Convert from a global struct to a subclass of
4061 gimple_opt_pass along with...
4062 (pass_data_cse_sincos): ...new pass_data instance and...
4063 (make_pass_cse_sincos): ...new function.
4064 (pass_optimize_bswap): Convert from a global struct to a subclass of
4065 gimple_opt_pass along with...
4066 (pass_data_optimize_bswap): ...new pass_data instance and...
4067 (make_pass_optimize_bswap): ...new function.
4068 (pass_optimize_widening_mul): Convert from a global struct to a
4069 subclass of gimple_opt_pass along with...
4070 (pass_data_optimize_widening_mul): ...new pass_data instance and...
4071 (make_pass_optimize_widening_mul): ...new function.
4072 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
4073 subclass of gimple_opt_pass along with...
4074 (pass_data_phiopt): ...new pass_data instance and...
4075 (make_pass_phiopt): ...new function.
4076 (pass_cselim): Convert from a global struct to a subclass of
4077 gimple_opt_pass along with...
4078 (pass_data_cselim): ...new pass_data instance and...
4079 (make_pass_cselim): ...new function.
4080 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
4081 subclass of gimple_opt_pass along with...
4082 (pass_data_phiprop): ...new pass_data instance and...
4083 (make_pass_phiprop): ...new function.
4084 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
4085 subclass of gimple_opt_pass along with...
4086 (pass_data_pre): ...new pass_data instance and...
4087 (make_pass_pre): ...new function.
4088 (pass_fre): Convert from a global struct to a subclass of
4089 gimple_opt_pass along with...
4090 (pass_data_fre): ...new pass_data instance and...
4091 (make_pass_fre): ...new function.
4092 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
4093 subclass of gimple_opt_pass along with...
4094 (pass_data_reassoc): ...new pass_data instance and...
4095 (make_pass_reassoc): ...new function.
4096 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
4097 subclass of gimple_opt_pass along with...
4098 (pass_data_sink_code): ...new pass_data instance and...
4099 (make_pass_sink_code): ...new function.
4100 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
4101 subclass of gimple_opt_pass along with...
4102 (pass_data_strlen): ...new pass_data instance and...
4103 (make_pass_strlen): ...new function.
4104 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
4105 struct to a subclass of gimple_opt_pass along with...
4106 (pass_data_build_alias): ...new pass_data instance and...
4107 (make_pass_build_alias): ...new function.
4108 (pass_build_ealias): Convert from a global struct to a subclass of
4109 gimple_opt_pass along with...
4110 (pass_data_build_ealias): ...new pass_data instance and...
4111 (make_pass_build_ealias): ...new function.
4112 (pass_ipa_pta): Convert from a global struct to a subclass of
4113 simple_ipa_opt_pass along with...
4114 (pass_data_ipa_pta): ...new pass_data instance and...
4115 (make_pass_ipa_pta): ...new function.
4116 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
4117 subclass of gimple_opt_pass along with...
4118 (pass_data_uncprop): ...new pass_data instance and...
4119 (make_pass_uncprop): ...new function.
4120 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
4121 global struct to a subclass of gimple_opt_pass along with...
4122 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
4123 (make_pass_late_warn_uninitialized): ...new function.
4124 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
4125 to a subclass of gimple_opt_pass along with...
4126 (pass_data_init_datastructures): ...new pass_data instance and...
4127 (make_pass_init_datastructures): ...new function.
4128 (pass_early_warn_uninitialized): Convert from a global struct to a
4129 subclass of gimple_opt_pass along with...
4130 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
4131 (make_pass_early_warn_uninitialized): ...new function.
4132 (pass_update_address_taken): Convert from a global struct to a
4133 subclass of gimple_opt_pass along with...
4134 (pass_data_update_address_taken): ...new pass_data instance and...
4135 (make_pass_update_address_taken): ...new function.
4136 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
4137 struct to a subclass of gimple_opt_pass along with...
4138 (pass_data_release_ssa_names): ...new pass_data instance and...
4139 (make_pass_release_ssa_names): ...new function.
4140 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
4141 subclass of gimple_opt_pass along with...
4142 (pass_data_stdarg): ...new pass_data instance and...
4143 (make_pass_stdarg): ...new function.
4144 * tree-switch-conversion.c (pass_convert_switch): Convert from a
4145 global struct to a subclass of gimple_opt_pass along with...
4146 (pass_data_convert_switch): ...new pass_data instance and...
4147 (make_pass_convert_switch): ...new function.
4148 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
4149 to a subclass of gimple_opt_pass along with...
4150 (pass_data_tail_recursion): ...new pass_data instance and...
4151 (make_pass_tail_recursion): ...new function.
4152 (pass_tail_calls): Convert from a global struct to a subclass of
4153 gimple_opt_pass along with...
4154 (pass_data_tail_calls): ...new pass_data instance and...
4155 (make_pass_tail_calls): ...new function.
4156 * tree-vect-generic.c (pass_lower_vector): Convert from a global
4157 struct to a subclass of gimple_opt_pass along with...
4158 (pass_data_lower_vector): ...new pass_data instance and...
4159 (make_pass_lower_vector): ...new function.
4160 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
4161 gimple_opt_pass along with...
4162 (pass_data_lower_vector_ssa): ...new pass_data instance and...
4163 (make_pass_lower_vector_ssa): ...new function.
4164 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
4165 to a subclass of gimple_opt_pass along with...
4166 (pass_data_slp_vectorize): ...new pass_data instance and...
4167 (make_pass_slp_vectorize): ...new function.
4168 (pass_ipa_increase_alignment): Convert from a global struct to a
4169 subclass of simple_ipa_opt_pass along with...
4170 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
4171 (make_pass_ipa_increase_alignment): ...new function.
4172 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
4173 gimple_opt_pass along with...
4174 (pass_data_vrp): ...new pass_data instance and...
4175 (make_pass_vrp): ...new function.
4176 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
4177 subclass of simple_ipa_opt_pass along with...
4178 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
4179 (make_pass_ipa_free_lang_data): ...new function.
4180 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
4181 gimple_opt_pass along with...
4182 (pass_data_tsan): ...new pass_data instance and...
4183 (make_pass_tsan): ...new function.
4184 (pass_tsan_O0): Convert from a global struct to a subclass of
4185 gimple_opt_pass along with...
4186 (pass_data_tsan_O0): ...new pass_data instance and...
4187 (make_pass_tsan_O0): ...new function.
4188 * var-tracking.c (pass_variable_tracking): Convert from a global
4189 struct to a subclass of rtl_opt_pass along with...
4190 (pass_data_variable_tracking): ...new pass_data instance and...
4191 (make_pass_variable_tracking): ...new function.
4192 * web.c (pass_web): Convert from a global struct to a subclass of
4193 rtl_opt_pass along with...
4194 (pass_data_web): ...new pass_data instance and...
4195 (make_pass_web): ...new function.
4196 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
4197 declaration with that of...
4198 (make_pass_mode_switch_use): ...new function.
4199 (pass_resolve_sw_modes): Replace declaration with that of...
4200 (make_pass_resolve_sw_modes): ...new function.
4201 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
4202 from a global struct to a subclass of rtl_opt_pass along with...
4203 (pass_data_mode_switch_use): ...new pass_data instance and...
4204 (make_pass_mode_switch_use): ...new function.
4205 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
4206 from a global struct to a subclass of rtl_opt_pass along with...
4207 (pass_data_resolve_sw_modes): ...new pass_data instance and...
4208 (make_pass_resolve_sw_modes): ...new function.
4209 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
4210 struct to a subclass of rtl_opt_pass along with...
4211 (pass_data_insert_vzeroupper): ...new pass_data instance and...
4212 (make_pass_insert_vzeroupper): ...new function.
4213 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
4214 global struct to a subclass of rtl_opt_pass along with...
4215 (pass_data_work_around_errata): ...new pass_data instance and...
4216 (make_pass_work_around_errata): ...new function.
4217 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
4218 struct to a subclass of rtl_opt_pass along with...
4219 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
4220 (make_pass_mips_machine_reorg2): ...new function.
4222 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4224 * passes.c (pass_manager::operator new): New.
4226 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4228 Handwritten part of conversion of passes to C++ classes.
4230 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
4231 (toplev.o): Add dep on PASS_MANAGER_H.
4232 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
4233 of early local pases to reflect this moving from a global to a
4234 member of gcc::pass_manager.
4235 (cgraph_add_new_function): Likewise.
4236 * lto-cgraph.c (lto_output_node): Update for conversion of
4237 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
4238 * passes.c (opt_pass::clone): New.
4239 (opt_pass::gate): New.
4240 (opt_pass::execute): New.
4241 (opt_pass::opt_pass): New.
4242 (pass_manager::execute_early_local_passes): New.
4243 (pass_manager::execute_pass_mode_switching): new.
4244 (finish_optimization_passes): Convert to...
4245 (pass_manager::finish_optimization_passes): ...this.
4246 (finish_optimization_passes): Update for conversion of passes to
4248 (register_dump_files_1): Use has_gate since we cannot portably
4249 check a vtable entry against NULL.
4250 (dump_one_pass): Likewise.
4251 (ipa_write_summaries_2): Likewise.
4252 (ipa_write_optimization_summaries_1): Likewise.
4253 (ipa_read_summaries_1): Likewise.
4254 (ipa_read_optimization_summaries_1): Likewise.
4255 (execute_ipa_stmt_fixups): Likewise.
4256 (pass_manager::pass_manager): Rewrite pass-creation, invoking
4257 pass-creation functions rather than wiring up globals, and
4258 storing the results in fields of pass_manager generated using
4260 (pass_manager::dump_profile_report): Update for conversion of
4261 passes to C++ classes.
4262 (pass_manager::execute_ipa_summary_passes): Likewise.
4263 (execute_one_ipa_transform_pass): Likewise.
4264 (execute_one_pass): Use has_gate and has_execute since we cannot
4265 portably check a vtable entry against NULL.
4266 * pass_manager.h (pass_manager::finish_optimization_passes): New.
4267 (pass_manager): Use pass-instances.def to add fields for the
4268 various pass instances.
4269 * toplev.c (finalize): Update for move of
4270 finish_optimization_passes to a method of gcc::pass_manager.
4271 * toplev.h (finish_optimization_passes): Move to method of class
4273 * tree-pass.h (struct pass_data): New.
4274 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
4275 (opt_pass::gate): Convert to virtual function.
4276 (opt_pass::~opt_pass): New.
4277 (opt_pass::clone): New.
4278 (opt_pass::execute): Convert to virtual function.
4279 (opt_pass::opt_pass): New.
4280 (opt_pass::ctxt_): new.
4281 (gimple_opt_pass): Convert to subclass of opt_pass.
4282 (gimple_opt_pass::gimple_opt_pass): New.
4283 (rtl_opt_pass): Convert to subclass of opt_pass.
4284 (rtl_opt_pass::rtl_opt_pass): New.
4285 (ipa_opt_pass_d): Convert to subclass of opt_pass.
4286 (ipa_opt_pass_d::ipa_opt_pass_d): New.
4287 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
4288 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
4289 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
4290 invocation of pass_mode_switching to reflect this moving from a
4291 global to a member of gcc::pass_manager.
4292 (ix86_option_override): Rework how pass_insert_vzeroupper is
4293 added to the pass_manager to reflect autogenerated changes.
4294 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
4296 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
4298 PR rtl-optimization/57708
4299 * recog.c (peep2_find_free_register): Validate all regs in a
4302 2013-08-05 Jan Hubicka <jh@suse.cz>
4305 * cgraph.c (verify_cgraph_node): Accept local flags from other
4307 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
4308 (function_and_variable_visibility): Likewise.
4309 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
4311 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4313 * graph.c (init_graph_slim_pretty_print): Remove.
4314 (print_graph_cfg): Do not call it. Use local pretty printer.
4315 (start_graph_dump): Likewise.
4317 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4319 * gimple-pretty-print.c (buffer): Remove.
4320 (initialized): Likewise.
4321 (maybe_init_pretty_print): Likewise.
4322 (print_gimple_stmt): Do not call it. Use non-static local
4323 pretty_printer variable.
4324 (print_gimple_expr): Likewise.
4325 (print_gimple_seq): Likewise.
4326 (gimple_dump_bb): Likewise.
4328 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4330 * asan.c (asan_pp): Remove.
4331 (asan_pp_initialized): Likewise.
4332 (asan_pp_initialize): Likewise.
4333 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
4334 (asan_emit_stack_protection): Tidy. Use local pretty printer.
4335 (asan_add_global): Likewise.
4337 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
4339 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
4340 * diagnostic.h (diagnostic_flush_buffer): Adjust.
4341 * pretty-print.c (pp_formatted_text_data): Likewise.
4342 (pp_indent): Rename from pp_base_indent.
4343 (pp_format): Rename from pp_base_format.
4344 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
4345 (pp_format_verbatim): Rename from pp_base_format_verbatim.
4346 (pp_flush): Rename from pp_base_flush.
4347 (pp_set_line_maximum_length): Rename from
4348 pp_base_set_line_maximum_length.
4349 (pp_clear_output_area): Rename from pp_base_clear_output_area.
4350 (pp_set_prefix): Rename from pp_base_set_prefix.
4351 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
4352 (pp_emit_prefix): Rename from pp_base_emit_prefix.
4353 (pp_append_text): Rename from pp_base_append_text.
4354 (pp_formatted_text): Rename from pp_base_formatted_text.
4355 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
4356 (pp_remaining_character_count_for_line): Rename from
4357 pp_base_remaining_character_count_for_line.
4358 (pp_newline): Rename from pp_base_newline.
4359 (pp_character): Rename from pp_base_character.
4360 (pp_string): Rename from pp_base_string.
4361 (pp_maybe_space): Rename from pp_base_maybe_space.
4362 * asan.c (asan_pp_string): Adjust.
4363 (asan_emit_stack_protection): Likewise.
4364 (asan_add_global): Likewise.
4365 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
4366 * tree-mudflap.c (mf_varname_tree): Likewise.
4367 * tree-pretty-print.c (pp_tree_identifier): Rename from
4368 pp_base_tree_identifier.
4369 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
4370 Declare as function.
4372 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4374 * pretty-print.h (pp_bar_bar): New.
4375 (pp_ampersand_ampersand): Likewise.
4376 (pp_less_equal): Likewise.
4377 (pp_greater_equal): Likewise.
4378 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
4379 printer functions instead of pp_string or operators and punctuators.
4380 (dump_gimple_call): Likewise.
4381 (dump_gimple_omp_for): Likewise.
4382 (dump_gimple_transaction): Likewise.
4383 (dump_gimple_phi): Likewise.
4384 (pp_gimple_stmt_1): Likewise.
4385 * sched-vis.c (print_insn): Likewise.
4386 * tree-mudflap.c (mf_varname_tree): Likewise.
4387 * tree-pretty-print.c (dump_block_node): Likewise.
4388 (dump_generic_node): Likewise.
4390 2013-08-02 Jan Hubicka <jh@suse.cz>
4392 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
4394 * lto-streamer-out.c (tree_is_indexable): Results decls and
4395 parm decls are not indexable.
4396 (DFS_write_tree_body): Do not follow args and results.
4397 (hash_tree): Likewise.
4398 (output_functions): Rearrange so struct function is needed
4399 only when real body is output; be able to also ouptut abstract
4400 functions; output DECL_ARGUMENTS and DECL_RESULT.
4401 (lto_output): When not in WPA, ale store abstract functions.
4402 (write_symbol): Do not care about RESULT_DECL.
4403 (output_symbol_p): Handle correctly sbtract decls.
4404 * lto-streamer-in.c (input_function): Rearrange so struct
4405 function can be NULL at entry; allow streaming of
4406 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
4407 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
4408 sanity check during LTO.
4409 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
4410 RESULT_DECl and DECL_ARGUMENTS.
4411 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
4414 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4416 * pretty-print.h (pp_underscore): New.
4418 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
4419 printer functions instead of pp_character.
4420 (dump_binary_rhs): Likewise.
4421 (dump_ternary_rhs): Likewise.
4422 (dump_gimple_call_args): Likewise.
4423 (pp_points_to_solution): Likewise.
4424 (dump_gimple_call): Likewise.
4425 (dump_gimple_switch): Likewise.
4426 (dump_gimple_cond): Likewise.
4427 (dump_gimple_bind): Likewise.
4428 (dump_gimple_try): Likewise.
4429 (dump_gimple_omp_for): Likewise.
4430 (dump_gimple_omp_continue): Likewise.
4431 (dump_gimple_omp_single): Likewise.
4432 (dump_gimple_omp_sections): Likewise.
4433 (dump_gimple_omp_block): Likewise.
4434 (dump_gimple_omp_critical): Likewise.
4435 (dump_gimple_transaction): Likewise.
4436 (dump_gimple_asm): Likewise.
4437 (dump_gimple_phi): Likewise.
4438 (dump_gimple_omp_parallel): Likewise.
4439 (dump_gimple_omp_task): Likewise.
4440 (dump_gimple_omp_atomic_load): Likewise.
4441 (dump_gimple_omp_atomic_store): Likewise.
4442 (dump_gimple_mem_ops): Likewise.
4443 (pp_gimple_stmt_1): Likewise.
4444 (pp_cfg_jump): Likewise.
4445 (dump_implicit_edges): Likewise.
4446 (gimple_dump_bb_for_graph): Likewise.
4447 * graph.c (draw_cfg_node): Likewise.
4448 * langhooks.c (lhd_print_error_function): Likewise.
4449 * sched-vis.c (print_exp): Likewise.
4450 (print_value): Likewise.
4451 (print_pattern): Likewise.
4452 (print_insn): Likewise.
4453 (rtl_dump_bb_for_graph): Likewise.
4454 * tree-pretty-print.c (dump_function_declaration): Likewise.
4455 (dump_array_domain): Likewise.
4456 (dump_omp_clause): Likewise.
4457 (dump_location): Likewise.
4458 (dump_generic_node): Likewise.
4459 (print_struct_decl): Likewise.
4460 * diagnostic.c (diagnostic_show_locus): Use pp_space.
4462 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4464 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
4465 candidate table when replacing a candidate statement.
4466 (replace_rhs_if_not_dup): Likewise.
4467 (replace_one_candidate): Likewise.
4469 2013-08-02 Jan Hubicka <jh@suse.cz>
4470 Martin Liska <marxin.liska@gmail.com>
4472 * cgraphunit.c (add_new_function): Fix logic when adding from
4474 (assemble_thunk): Rename to ...
4475 (expand_thunk); .. this one; export; get it working with
4476 general functions; make produced gimple valid.
4477 * cgraph.h (expand_thunk): Declare.
4479 2013-08-02 Jan Hubicka <jh@suse.cz>
4481 * ipa-cp.c (gather_context_independent_values): Use
4482 ipa_get_param_move_cost.
4483 (get_replacement_map): Remove PARAM; move parameter folding
4485 (create_specialized_node): Update.
4486 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
4487 assert that we have gimple body; update move_cost.
4488 (count_formal_params): Assert that we have gimple body.
4489 (ipa_dump_param): New function.
4490 (ipa_alloc_node_params): Break out from ...
4491 (ipa_initialize_node_params): ... here.
4492 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
4493 (ipa_write_node_info): Stream move costs.
4494 (ipa_read_node_info): Read move costs.
4495 (ipa_update_after_lto_read): Do not recompute node params.
4496 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
4497 (ipa_get_param): Check we are not in WPA.
4498 (ipa_get_param_move_cost): New.
4499 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
4500 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
4501 parm numbers to be present.
4503 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4505 PR rtl-optimization/58048
4506 * lra-constraints.c (process_alt_operands): Don't check asm
4507 operand on register.
4509 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
4511 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
4512 the implied StoreLoad barrier for atomic operations if before.
4514 2013-08-02 Jan Hubicka <jh@suse.cz>
4515 Martin Liska <marxin.liska@gmail.com>
4517 * cgraph.c (cgraph_function_body_availability): Do not check
4519 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
4520 symtab_node_availability): Declare.
4521 * ipa.c (can_replace_by_local_alias): New.
4522 (function_and_variable_visibility): Use it.
4523 * symtab.c (symtab_for_node_and_aliases,
4524 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
4526 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4528 PR rtl-optimization/57963
4529 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
4530 (lra_constraints): Use them.
4532 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
4534 * config/arm/types.md (define_attr "type"): Add "load_acq"
4536 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
4538 (cortex_a53_store1): Likewise.
4540 2013-08-01 Jan Hubicka <jh@suse.cz>
4542 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
4543 partitions are not needed.
4545 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
4547 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
4548 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
4549 MAYBE_NON_Q_CLASS_P where appropriate.
4551 2013-08-01 Jan Hubicka <jh@suse.cz>
4553 * cgraph.h (release_function_body): Declare.
4554 * tree.c (free_lang_data_in_decl): Free, parameters and return values
4555 of unused delcarations.
4557 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4559 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
4560 RTL form when subtracting a constant.
4562 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4564 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
4565 Generate canonical plus rtx with negated immediate instead of minus
4567 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
4569 2013-08-01 Jan Hubicka <jh@suse.cz>
4571 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
4572 (cgraph_release_function_body): Likewise.
4573 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
4574 * cgraph.h (cgrpah_node): Rename abstract_and_needed
4575 to used_as_abstract_origin.
4576 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
4577 symbols used as abstract origins.
4578 * cgraphunit.c (analyze_functions): Update.
4579 * ipa.c (symtab_remove_unreachable_nodes): Recompute
4580 used_as_abstract_origin.
4581 * tree-inline.c (tree_function_versioning): Update
4582 used_as_abstract_origin; be ready for DECL_RESULT and
4583 DECL_ARGUMENTS to be NULL.
4585 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
4586 for abstract functions.
4587 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
4590 2013-08-01 Jan Hubicka <jh@suse.cz>
4592 * profile.c (compute_value_histograms): Fix thinko.
4594 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4596 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
4597 aarch-common-protos.h to extra_headers.
4598 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4599 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
4600 * config/aarch64/t-aarch64 (aarch-common.o): Define.
4602 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4604 * config/aarch64/aarch64.md (define_attr "type"): Delete.
4605 Include "../arm/types.md". Define "type" attribute for all patterns.
4606 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
4609 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
4611 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
4612 to support power8 load fusion.
4613 (fusion_gpr_mem_load): Likewise.
4615 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
4617 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
4618 declarations for power8 load fusion.
4619 (emit_fusion_gpr_load): Likewise.
4621 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
4622 tuning for power8, turn on fusion mode by default. Turn on sign
4623 extending fusion mode if normal fusion mode is on, and we are at
4625 (fusion_gpr_load_p): New function, return true if we can fuse an
4626 addis instruction with a dependent load to a GPR.
4627 (emit_fusion_gpr_load): Emit the instructions for power8 load
4630 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
4631 (VSX load fusion peepholes): New peepholes to fuse together an
4632 addi instruction with a VSX load instruction.
4634 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
4635 peepholes to fuse an addis instruction with a load to a GPR base
4636 register. If we are supporting sign extending fusions, convert
4637 sign extending loads to zero extending loads and add an explicit
4640 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4642 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
4643 aarch-common-protos.h to extra_headers.
4644 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4645 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
4646 (arm_early_store_addr_dep): Likewise.
4647 (arm_no_early_alu_shift_dep): Likewise.
4648 (arm_no_early_alu_shift_value_dep): Likewise.
4649 (arm_no_early_mul_dep): Likewise.
4650 (arm_no_early_store_addr_dep): Likewise.
4651 (arm_mac_accumulator_is_mul_result): Likewise.
4652 (arm_mac_accumulator_is_result): Likewise.
4653 * config/arm/aarch-common.c: ... here. New file.
4654 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
4656 (arm_early_store_addr_dep): Likewise.
4657 (arm_no_early_alu_shift_dep): Likewise.
4658 (arm_no_early_alu_shift_value_dep): Likewise.
4659 (arm_no_early_mul_dep): Likewise.
4660 (arm_no_early_store_addr_dep): Likewise.
4661 (arm_mac_accumulator_is_mul_result): Likewise.
4662 (arm_mac_accumulator_is_result): Likewise.
4663 * config/arm/aarch-common-protos.h: ... here. New file.
4664 * config/arm/t-arm (aarch-common.o): Define.
4666 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4668 * config/arm/arm.md: Include new file "types.md".
4669 (define_attr "type"): Move from here to ...
4670 (define_attr "mul32"): Likewise.
4671 (define_attr "mul64"): Likewise.
4672 * config/arm/types.md: ... here. New file.
4674 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4676 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4677 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4679 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4681 * gen-pass-instances.awk: Fix offset of substr().
4683 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4685 * Makefile.in (pass-instances.def): New.
4686 (passes.o): Replace dependency on passes.def with one on
4689 * gen-pass-instances.awk: New.
4691 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4692 rather than passes.def, updating local definition of NEXT_PASS
4693 macro to add an extra NUM parameter (currently unused).
4695 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4697 * Makefile.in (PASS_MANAGER_H): New.
4698 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4699 (passes.o): Likewise.
4700 (statistics.o): Likewise.
4701 (cgraphunit.o): Likewise.
4702 (context.o): Depend on PASS_MANAGER_H.
4704 * pass_manager.h: New.
4706 * cgraphunit.c (cgraph_add_new_function): Update for moves
4707 of globals to fields of pass_manager.
4708 (analyze_function): Likewise.
4709 (expand_function): Likewise.
4710 (ipa_passes): Likewise.
4711 (compile): Likewise.
4713 * context.c (context::context): New.
4714 * context.h (context::context): New.
4715 (context::get_passes): New.
4716 (context::passes_): New.
4718 * lto-cgraph.c (input_node): Update for moves of globals to
4719 fields of pass_manager.
4721 * passes.c (all_passes): Remove, in favor of a field of the
4722 same name within the new class pass_manager.
4723 (all_small_ipa_passes): Likewise.
4724 (all_lowering_passes): Likewise.
4725 (all_regular_ipa_passes): Likewise.
4726 (all_late_ipa_passes): Likewise.
4727 (all_lto_gen_passes): Likewise.
4728 (passes_by_id): Likewise.
4729 (passes_by_id_size): Likewise.
4730 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4731 the new class pass_manager.
4732 (set_pass_for_id): Convert to...
4733 (pass_manager::set_pass_for_id): ...method.
4734 (get_pass_for_id): Convert to...
4735 (pass_manager::get_pass_for_id): ...method.
4736 (register_one_dump_file): Move body of implementation into...
4737 (pass_manager::register_one_dump_file): ...here.
4738 (register_dump_files_1): Convert to...
4739 (pass_manager::register_dump_files_1): ...method.
4740 (register_dump_files): Convert to...
4741 (pass_manager::register_dump_files): ...method.
4742 (create_pass_tab): Update for moves of globals to fields of
4744 (dump_passes): Move body of implementation into...
4745 (pass_manager::dump_passes): ...here.
4746 (register_pass): Move body of implementation into...
4747 (pass_manager::register_pass): ...here.
4748 (init_optimization_passes): Convert into...
4749 (pass_manager::pass_manager): ...constructor for new
4750 pass_manager class, and initialize the pass_lists array.
4751 (check_profile_consistency): Update for moves of globals to
4752 fields of pass_manager.
4753 (dump_profile_report): Move body of implementation into...
4754 (pass_manager::dump_profile_report): ...here.
4755 (ipa_write_summaries_1): Update for moves of pass lists from
4756 being globals to fields of pass_manager.
4757 (ipa_write_optimization_summaries): Likewise.
4758 (ipa_read_summaries): Likewise.
4759 (ipa_read_optimization_summaries): Likewise.
4760 (execute_all_ipa_stmt_fixups): Likewise.
4762 * statistics.c (statistics_fini): Update for moves of globals to
4763 fields of pass_manager.
4765 * toplev.c (general_init): Replace call to
4766 init_optimization_passes with construction of the pass_manager
4769 * tree-pass.h (all_passes): Remove, in favor of a field of the
4770 same name within the new class pass_manager.
4771 (all_small_ipa_passes): Likewise.
4772 (all_lowering_passes): Likewise.
4773 (all_regular_ipa_passes): Likewise.
4774 (all_lto_gen_passes): Likewise.
4775 (all_late_ipa_passes): Likewise.
4776 (passes_by_id): Likewise.
4777 (passes_by_id_size): Likewise.
4778 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4779 the new class pass_manager.
4780 (get_pass_for_id): Remove.
4782 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4784 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4787 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4789 * arm.md (mulhi3): New expand pattern.
4791 2013-07-30 Jan Hubicka <jh@suse.cz>
4792 Martin Liska <marxin.liska@gmail.com>
4794 * profile.c (compute_value_histograms): Do not ICE when
4795 there is mismatch only on some counters.
4797 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4799 PR rtl-optimization/57637
4800 * function.c (move_insn_for_shrink_wrap): Also check the
4801 GEN set of the LIVE problem for the liveness analysis
4802 if it exists, otherwise give up.
4804 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4806 PR tree-optimization/57993
4807 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4808 replaced statement in the candidate table.
4809 (phi_add_costs): Return infinite cost when the hidden basis does
4810 not dominate all phis on which the candidate is dependent.
4811 (replace_one_candidate): Record replaced statement in the
4814 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4816 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4817 (ashlv2si3): New expander.
4818 (*ashlv2si3_i): New define_insn_and_split.
4819 * predicates.md (float_operation): Allow patterns with three
4822 PR rtl-optimization/58021
4823 * mode-switching.c (create_pre_exit): Always split off preceding
4824 insns if we are not at the basic block head.
4826 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4828 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4829 (UCLIBC_DYNAMIC_LINKER): New macro.
4830 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4832 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4833 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4835 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4836 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4837 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4838 for SF and DF modes. Use ieee_quad_format for TF mode.
4839 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
4840 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
4841 (mips_option_override): Handle `-mnan=legacy'.
4842 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
4843 `-mabs=2008' and `-mnan=2008'.
4844 (OPTION_DEFAULT_SPECS): Add "nan" default.
4845 (ASM_SPEC): Handle `-mnan='.
4846 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
4847 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
4848 comment accordingly.
4849 (neg<mode>2): Likewise.
4850 * config/mips/mips.opt (mabs, mnan): New options.
4851 * doc/install.texi (Configuration): Document `--with-nan=' option.
4852 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
4854 (MIPS Options): Document them.
4855 * config.gcc <mips*-*-*>: Handle `--with-nan='.
4856 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
4857 * configure: Regenerate.
4858 * config.in: Regenerate.
4860 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4862 * config/i386/i386.md (float post-reload splitters): Do not check
4863 for subregs of SSE registers.
4865 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4866 H.J. Lu <hongjiu.lu@intel.com>
4870 * config/i386/i386.md (post-reload splitter
4871 to avoid partial SSE reg dependency stalls): New pattern.
4873 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
4875 * config/s390/s390.md ("movcc"): Swap load and store instructions.
4877 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
4879 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
4880 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
4882 2013-07-26 Cary Coutant <ccoutant@google.com>
4884 * dwarf2out.c (die_checksum_ordered): Don't include template
4885 instantiations in signature.
4886 (is_template_parameter): New function.
4887 (is_template_instantiation): New function.
4888 (generate_skeleton_bottom_up): Don't include template instantiations
4890 (generate_skeleton): Likewise.
4891 (break_out_comdat_types): Move recursive call to break out nested
4893 (prune_unused_types_mark_generic_parms_dies): Call
4894 is_template_parameter.
4896 2013-07-26 Ian Bolton <ian.bolton@arm.com>
4898 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
4899 uses vector registers.
4900 * config/aarch64/iterators.md: Add attributes rtn and vas.
4902 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4903 Richard Earnshaw <richard.earnshaw@arm.com>
4905 * combine.c (simplify_comparison): Re-canonicalize operands
4907 * config/arm/arm.md (movcond_addsi): New splitter.
4909 2013-07-25 Sterling Augustine <saugustine@google.com>
4911 * dwarf2out.c (size_of_pubnames): Move code to...
4912 (include_pubname_in_output): ...here. New.
4913 (want_pubnames): Rearrange.
4914 (output_pubnames): Call include_pubname_in_output. Move assertion.
4916 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4918 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
4920 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4923 * doc/extend.texi: Remove obsolete builtins. Fix
4924 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
4926 2013-07-25 Jan Hubicka <jh@suse.cz>
4928 * cgraph.c (release_function_body): Break out from ...
4929 (cgraph_release_function_body): ... this one; also release DECL_RESULT
4931 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
4932 old_tree in the map.
4933 (create_specialized_node): Update.
4934 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
4936 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
4937 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
4938 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
4939 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
4942 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4944 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
4945 addsi3_carryin_alt2_<optab>): Correct output template.
4947 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4949 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
4950 Adjust for arm_restrict_it.
4951 Remove trailing whitespace.
4953 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
4955 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
4956 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
4958 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
4960 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
4962 PR rtl-optimization/57960
4963 * lra-constraints.c (process_alt_operands): Use the right mode
4964 when checking strict_low.
4966 2013-07-25 Jan Hubicka <jh@suse.cz>
4968 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
4969 * cgraph.c (cgraph_remove_node): Do not release function body
4970 when in cgraph streaming.
4971 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
4972 in other partitions are not considered reachable; fix handling of
4975 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4977 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
4979 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4984 * config/arm/arm.md ("*sibcall_insn): Replace use of
4985 Ss with US. Adjust output for v5 and v4t.
4986 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
4988 * config/arm/constraints.md ("Ss"): Rename to US.
4990 2013-07-25 Terry Guo <terry.guo@arm.com>
4992 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
4993 shift_add/shift_sub0/shift_sub1 RTXs.
4995 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
4996 Anton Blanchard <anton@au1.ibm.com>
4998 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
4999 (altivec_vpks<VI_char>ss): Likewise.
5000 (altivec_vpks<VI_char>us): Likewise.
5001 (altivec_vpku<VI_char>us): Likewise.
5002 (altivec_vpku<VI_char>um): Likewise.
5004 2013-07-24 David Malcolm <dmalcolm@redhat.com>
5006 Introduce context class.
5008 * Makefile.in (CONTEXT_H): New.
5009 (OBJS): Add context.o.
5010 (toplev.o): Add CONTEXT_H to dependencies.
5013 * toplev.c (general_init): Create the singleton gcc::context instance.
5019 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
5021 PR rtl-optimization/57968
5022 * mode-switching.c (create_pre_exit): Allow instructions that
5023 don't set a return register to need a non-exit mode.
5025 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5026 Anton Blanchard <anton@au1.ibm.com>
5028 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
5029 operands to vperm for little endian.
5030 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
5031 of lvsl to create the control mask for a vperm for little endian.
5033 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5034 Anton Blanchard <anton@au1.ibm.com>
5036 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5037 two operands for little-endian.
5039 2013-07-23 Steve Ellcey <sellcey@mips.com>
5041 * config/mips/mips.c (mips_case_values_threshold): New.
5042 (TARGET_CASE_VALUES_THRESHOLD): Define.
5044 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5045 Anton Blanchard <anton@au1.ibm.com>
5047 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
5048 selection of field for vector splat in little endian mode.
5050 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5052 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
5053 expanders to rs6000.md.
5054 (ior<mode>3): Likewise.
5055 (and<mode>3): Likewise.
5056 (one_cmpl<mode>2): Likewise.
5057 (nor<mode>3): Likewise.
5058 (andc<mode>3): Likewise.
5059 (eqv<mode>3): Likewise.
5060 (nand<mode>3): Likewise.
5061 (orc<mode>3): Likewise.
5063 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
5066 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
5067 to split multi-word logical operations.
5068 (rs6000_split_logical_di): Likewise.
5069 (rs6000_split_logical): Likewise.
5071 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
5072 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
5073 and allow TImode operations in 32-bit.
5074 (vsx_and<mode>3_64bit): Likewise.
5075 (vsx_ior<mode>3_32bit): Likewise.
5076 (vsx_ior<mode>3_64bit): Likewise.
5077 (vsx_xor<mode>3_32bit): Likewise.
5078 (vsx_xor<mode>3_64bit): Likewise.
5079 (vsx_one_cmpl<mode>2_32bit): Likewise.
5080 (vsx_one_cmpl<mode>2_64bit): Likewise.
5081 (vsx_nor<mode>3_32bit): Likewise.
5082 (vsx_nor<mode>3_64bit): Likewise.
5083 (vsx_andc<mode>3_32bit): Likewise.
5084 (vsx_andc<mode>3_64bit): Likewise.
5085 (vsx_eqv<mode>3_32bit): Likewise.
5086 (vsx_eqv<mode>3_64bit): Likewise.
5087 (vsx_nand<mode>3_32bit): Likewise.
5088 (vsx_nand<mode>3_64bit): Likewise.
5089 (vsx_orc<mode>3_32bit): Likewise.
5090 (vsx_orc<mode>3_64bit): Likewise.
5092 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
5093 logical types in GPRs.
5095 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
5096 logical insns to rs6000.md, and allow TImode operations in
5098 (altivec_ior<mode>3): Likewise.
5099 (altivec_xor<mode>3): Likewise.
5100 (altivec_one_cmpl<mode>2): Likewise.
5101 (altivec_nor<mode>3): Likewise.
5102 (altivec_andc<mode>3): Likewise.
5104 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
5105 attributes for moving the 128-bit logical operations into
5107 (BOOL_REGS_OUTPUT): Likewise.
5108 (BOOL_REGS_OP1): Likewise.
5109 (BOOL_REGS_OP2): Likewise.
5110 (BOOL_REGS_UNARY): Likewise.
5111 (BOOL_REGS_AND_CR0): Likewise.
5112 (one_cmpl<mode>2): Add support for DI logical operations on
5113 32-bit, splitting the operations to 32-bit.
5117 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
5118 changes to combine the 32/64-bit code, allow logical operations on
5119 TI mode in 32-bit, and to use similar match_operator patterns like
5120 scalar mode uses. Combine the Altivec and VSX code for logical
5121 operations, and move it here.
5122 (ior<mode>3, 128-bit types): Likewise.
5123 (xor<mode>3, 128-bit types): Likewise.
5124 (one_cmpl<mode>3, 128-bit types): Likewise.
5125 (nor<mode>3, 128-bit types): Likewise.
5126 (andc<mode>3, 128-bit types): Likewise.
5127 (eqv<mode>3, 128-bit types): Likewise.
5128 (nand<mode>3, 128-bit types): Likewise.
5129 (orc<mode>3, 128-bit types): Likewise.
5130 (and<mode>3_internal): Likewise.
5131 (bool<mode>3_internal): Likewise.
5132 (boolc<mode>3_internal1): Likewise.
5133 (boolc<mode>3_internal2): Likewise.
5134 (boolcc<mode>3_internal1): Likewise.
5135 (boolcc<mode>3_internal2): Likewise.
5136 (eqv<mode>3_internal1): Likewise.
5137 (eqv<mode>3_internal2): Likewise.
5138 (one_cmpl1<mode>3_internal): Likewise.
5140 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5142 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5143 Rename flag_stack_usage to flag_stack_usage_info.
5145 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5147 * config/microblaze/sync.md: New file.
5148 * config/microblaze/microblaze.md: Include sync.md
5149 * config/microblaze/microblaze.c: Add print_operand 'y'.
5150 * config/microblaze/constraints.md: Add memory_contraint
5151 'Q' which is a single register.
5153 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
5155 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
5157 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5159 * reload.c (find_reloads): Exit loop once we find this operand
5160 cannot be reloaded somehow for this alternative.
5162 * reload.c (find_reloads): Exit loop once we find a hard register.
5164 * rtlanal.c (computed_jump_p): Exit loop once we find label
5167 * i386.c (ix86_pad_returns): Exit loop after setting replace.
5169 * cfgloopmanip.c (remove_path): Exit loop after setting
5172 * gensupport.c (subst_dup): Avoid loop if code is not
5173 MATCH_DUP nor MATCH_OP_DUP.
5175 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
5177 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
5179 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5181 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
5182 true for SP_REGNUM if mode == ptr_mode.
5183 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
5184 with value R0_REGNUM + 31.
5186 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5188 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
5189 pad pointer-typed argument downward.
5191 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5193 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
5194 and __ILP32__ when the ILP32 model is in use.
5196 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5198 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
5199 (aarch64_load_symref_appropriately): In the case of
5200 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
5201 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
5202 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
5203 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
5204 if 'mode' doesn't equal to 'ptr_mode'.
5205 (aarch64_output_mi_thunk): Add an assertion on the alignment of
5206 'vcall_offset'; change to call aarch64_emit_move differently depending
5207 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
5208 to calculate the upper bound of 'vcall_offset'.
5209 (aarch64_cannot_force_const_mem): Change to also return true if
5211 (aarch64_legitimize_reload_address): In the case of large
5212 displacements, add new local variable 'xmode' and an assertion
5213 based on it; change to use 'xmode' to generate the new rtx and
5215 (aarch64_asm_trampoline_template): Change to generate the template
5216 differently depending on TARGET_ILP32 or not; change to use
5217 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
5218 (aarch64_trampoline_size): Removed.
5219 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
5220 and replace immediate literals with it. Change to use 'ptr_mode'
5221 instead of 'DImode' and call convert_memory_address if the mode
5222 of 'fnaddr' doesn't equal to 'ptr_mode'.
5223 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
5225 (aarch64_elf_asm_destructor): Likewise.
5226 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
5227 on TARGET_ILP32 instead of aarch64_trampoline_size.
5228 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
5229 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
5230 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
5231 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
5232 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
5233 (storewb_pair<GPI:mode>_<P:mode>): ... this.
5234 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
5235 depending on the value of 'mode'.
5236 (add_losym_<mode>): New.
5237 (ldr_got_small_<mode>): New, based on ldr_got_small.
5238 (ldr_got_small): Remove.
5239 (ldr_got_small_sidi): New.
5240 * config/aarch64/iterators.md (P): New.
5241 (PTR): Change to 'ptr_mode' in the condition.
5243 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5245 * config.gcc (aarch64*-*-*): Support --with-abi.
5246 (aarch64*-*-elf): Support --with-multilib-list.
5247 (aarch64*-*-linux*): Likewise.
5248 (supported_defaults): Add abi to aarch64*-*-*.
5249 * configure.ac: Mention AArch64 for --with-multilib-list.
5250 * configure: Re-generated.
5251 * config/aarch64/biarchilp32.h: New file.
5252 * config/aarch64/biarchlp64.h: New file.
5253 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
5255 (MULTILIB_DEFAULTS): Ditto.
5256 (DRIVER_SELF_SPECS): Ditto.
5257 (ASM_SPEC): Update to also substitute -mabi.
5258 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
5259 file whose name depends on -mabi= and -mbig-endian.
5260 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
5262 (POINTER_SIZE): New define.
5263 (POINTERS_EXTEND_UNSIGNED): Ditto.
5264 (enum aarch64_abi_type): New enumeration tag.
5265 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
5266 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
5267 (TARGET_ILP32): New define.
5268 * config/aarch64/aarch64.opt (mabi): New.
5270 (ilp32, lp64): New values for -mabi.
5271 * config/aarch64/t-aarch64 (comma): New define.
5272 (MULTILIB_OPTIONS): Ditto.
5273 (MULTILIB_DIRNAMES): Ditto.
5274 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
5275 * doc/invoke.texi: Document -mabi for AArch64.
5277 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
5279 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
5281 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5282 Anton Blanchard <anton@au1.ibm.com>
5284 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
5285 endianness when selecting field to splat.
5287 2013-07-22 Eric Christopher <echristo@gmail.com>
5289 * dwarf2out.c (die_odr_checksum): New function to use
5290 CHECKSUM_ macros and ULEB128 for DIE tag.
5291 (generate_type_signature): Use.
5293 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
5295 * config.gcc (sparc*-*-*): Accept leon3 processor.
5296 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
5297 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
5298 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
5299 * config/sparc/sparc.opt (enum processor_type): Add leon3.
5300 (mfix-ut699): Adjust comment.
5301 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
5302 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
5303 (CPP_CPU_SPEC): Likewise.
5304 (ASM_CPU_SPEC): Likewise.
5305 * config/sparc/sparc.c (leon3_cost): New constant.
5306 (sparc_option_override): Add leon3 support.
5307 (mem_ref): New function.
5308 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
5309 (sparc_do_work_around_errata): Look into the instruction in the delay
5310 slot and adjust accordingly. Add fix for the data cache nullify issues
5311 of the UT699. Change insertion position for the NOP.
5312 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
5313 (leon3_load): New reservation.
5314 (leon_store): Bump latency to 2.
5315 (grfpu): New automaton.
5316 (grfpu_alu): New unit.
5317 (grfpu_ds): Likewise.
5318 (leon_fp_alu): Adjust.
5319 (leon_fp_mult): Delete.
5320 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
5321 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
5322 * config/sparc/sparc.md (cpu): Add leon3.
5323 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
5325 (atomic_test_and_set): Likewise.
5328 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
5330 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
5331 default for R5900 targets.
5332 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
5333 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
5334 * config/mips/mips.c (mips_option_override): Report an error for
5335 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
5336 for -march=r5900 -mhard-float.
5338 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5340 * df-problems.c (can_move_insns_across): Exit loop once we
5341 find a non-fixed, non-global register.
5343 * ipa-pure-const.c (propagate_nothrow): Exit loop after
5346 * omega.c (omega_eliminate_red): Break after setting red_found.
5347 (omega_problem_has_red_equations): Similarly after setting found.
5348 (omega_query_variable): Similarly after setting coupled.
5350 2013-07-22 Marek Polacek <polacek@redhat.com>
5352 * gimplify.c: Don't include gimple.h twice.
5354 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5356 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
5357 instead of TARGET_THUMB1.
5358 (Pz): New constraint.
5359 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
5361 (compare_negsi_si): Likewise.
5362 (compare_addsi2_op0): Likewise.
5363 (compare_addsi2_op1): Likewise.
5364 (addsi3_carryin_<optab>): Likewise.
5365 (addsi3_carryin_alt2_<optab>): Likewise.
5366 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
5367 for arm_restrict_it.
5368 (subsi3_carryin): Likewise.
5369 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
5370 (minmax_arithsi): Disable for arm_restrict_it.
5371 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
5372 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
5373 (satsi_<SAT:code>_shift): Likewise.
5374 (arm_shiftsi3): Add alternative for 16-bit encoding.
5375 (arm32_movhf): Disable for arm_restrict_it.
5376 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
5377 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
5379 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
5381 * config/arm/arm.md (attribute "insn"): Delete.
5382 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
5383 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
5384 (not_shiftsi): Update for attribute change.
5385 (not_shiftsi_compare0): Likewise.
5386 (not_shiftsi_compare0_scratch): Likewise.
5387 (arm_one_cmplsi2): Likewise.
5388 (thumb1_one_cmplsi2): Likewise.
5389 (notsi_compare0): Likewise.
5390 (notsi_compare0_scratch): Likewise.
5391 (thumb1_movdi_insn): Likewise.
5392 (arm_movsi_insn): Likewise.
5393 (movhi_insn_arch4): Likewise.
5394 (movhi_bytes): Likewise.
5395 (arm_movqi_insn): Likewise.
5396 (thumb1_movqi_insn): Likewise.
5397 (arm32_movhf): Likewise.
5398 (thumb1_movhf): Likewise.
5399 (arm_movsf_soft_insn): Likewise.
5400 (thumb1_movsf_insn): Likewise.
5401 (thumb_movdf_insn): Likewise.
5402 (movsicc_insn): Likewise.
5403 (movsfcc_soft_insn): Likewise.
5404 (and_scc): Likewise.
5405 (cond_move): Likewise.
5406 (if_move_not): Likewise.
5407 (if_not_move): Likewise.
5408 (if_shift_move): Likewise.
5409 (if_move_shift): Likewise.
5410 (if_shift_shift): Likewise.
5411 (if_not_arith): Likewise.
5412 (if_arith_not): Likewise.
5413 (cond_move_not): Likewise.
5414 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
5415 (neon_mov<mode>): Likewise.
5416 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
5417 (thumb2_movsi_vfp): Likewise.
5418 (movsf_vfp): Likewise.
5419 (thumb2_movsf_vfp): Likewise.
5420 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5422 (cortexa7_older_only): Likewise.
5423 (cortexa7_younger): Likewise.
5424 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
5425 (1020alu_shift_op): Likewise.
5426 (1020alu_shift_reg_op): Likewise.
5427 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
5428 (alu_shift_op): Likewise.
5429 (alu_shift_reg_op): Likewise.
5430 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
5431 (11_alu_shift_op): Likewise.
5432 (11_alu_shift_reg_op): Likewise.
5433 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
5434 (9_alu_shift_reg_op): Likewise.
5435 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5437 (cortex_a15_alu_shift): Likewise.
5438 (cortex_a15_alu_shift_reg): Likewise.
5439 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
5440 (cortex_a5_alu_shift): Likewise.
5441 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
5443 (cortex_a53_alu_shift): Likewise.
5444 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5446 (cortex_a7_alu_reg): Likewise.
5447 (cortex_a7_alu_shift): Likewise.
5448 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5449 (cortex_a8_alu_shift): Likewise.
5450 (cortex_a8_alu_shift_reg): Likewise.
5451 (cortex_a8_mov): Likewise.
5452 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5453 (cortex_a9_dp_shift): Likewise.
5454 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
5455 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
5456 (cortex_r4_mov): Likewise.
5457 (cortex_r4_alu_shift): Likewise.
5458 (cortex_r4_alu_shift_reg): Likewise.
5459 * config/arm/fa526.md (526_alu_op): Update for attribute change.
5460 (526_alu_shift_op): Likewise.
5461 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
5462 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
5463 (626te_alu_shift_op): Likewise.
5464 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
5465 (726te_alu_op): Likewise.
5466 (726te_alu_shift_op): Likewise.
5467 (726te_alu_shift_reg_op): Likewise.
5468 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
5469 (mp626_alu_shift_op): Likewise.
5470 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
5471 (pj4_alu_e1_conds): Likewise.
5472 (pj4_alu): Likewise.
5473 (pj4_alu_conds): Likewise.
5474 (pj4_shift): Likewise.
5475 (pj4_shift_conds): Likewise.
5476 (pj4_alu_shift): Likewise.
5477 (pj4_alu_shift_conds): Likewise.
5479 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5481 * config/arm/predicates.md (shiftable_operator_strict_it):
5483 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
5484 Disable cond_exec version for arm_restrict_it.
5485 (thumb2_smaxsi3): Convert to generate cond_exec.
5486 (thumb2_sminsi3): Likewise.
5487 (thumb32_umaxsi3): Likewise.
5488 (thumb2_uminsi3): Likewise.
5489 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
5490 (thumb2_neg_abssi2): Likewise.
5491 (thumb2_mov_scc): Add alternative for 16-bit encoding.
5492 (thumb2_movsicc_insn): Adjust alternatives.
5493 (thumb2_mov_negscc): Disable for arm_restrict_it.
5494 (thumb2_mov_negscc_strict_it): New pattern.
5495 (thumb2_mov_notscc_strict_it): New pattern.
5496 (thumb2_mov_notscc): Disable for arm_restrict_it.
5497 (thumb2_ior_scc): Likewise.
5498 (thumb2_ior_scc_strict_it): New pattern.
5499 (thumb2_cond_move): Adjust for arm_restrict_it.
5500 (thumb2_cond_arith): Disable for arm_restrict_it.
5501 (thumb2_cond_arith_strict_it): New pattern.
5502 (thumb2_cond_sub): Adjust for arm_restrict_it.
5503 (thumb2_movcond): Likewise.
5504 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
5505 (thumb2_zero_extendhisi2_v6): Likewise.
5506 (thumb2_zero_extendqisi2_v6): Likewise.
5507 (orsi_notsi_si): Likewise.
5508 (orsi_not_shiftsi_si): Likewise.
5510 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
5512 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
5513 instruction sequence is 1 byte shorter.
5515 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
5517 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
5518 it is not needed after split.
5520 2013-07-20 Iain Sandoe <iain@codesourcery.com>
5523 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
5524 second label for nonlocal goto receivers. Don't output pic base labels
5525 unless we're producing PIC; mark that action unreachable().
5526 (ix86_save_reg): If the function contains a nonlocal label, save the
5528 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
5529 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
5530 (update_pic_label_number_if_needed): New.
5531 (machopic_output_function_base_name): Adjust for nonlocal receiver
5533 (machopic_should_output_picbase_label): New.
5534 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
5535 (nonlocal_goto_receiver): New insn and split.
5537 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
5539 * config/aarch64/aarch64-builtins.c
5540 (aarch64_fold_builtin): Fold abs in all modes.
5541 * config/aarch64/aarch64-simd-builtins.def
5542 (abs): Enable for all modes.
5543 * config/aarch64/arm_neon.h
5544 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
5545 (vabs_f64): Add missing intrinsic.
5547 2013-07-19 Ian Bolton <ian.bolton@arm.com>
5549 * config/aarch64/arm_neon.h (vabs_s64): New function
5551 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
5554 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
5555 * config/avr/avr.md (adjust_len): Add `round'.
5556 * config/avr/avr-protos.h (avr_out_round): New prototype.
5557 (avr_out_plus): Add `out_label' argument.
5558 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
5559 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
5560 Handle the case where `insn' is just a pattern.
5561 (avr_out_bitop): Handle the case where `insn' is just a pattern.
5562 (avr_out_round): New function.
5563 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
5565 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
5567 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5568 Add check for flag_stack_usage to handle -fstack-usage support
5570 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
5572 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
5573 interaction for new Power8 flags and VSX.
5575 2013-07-18 Sriraman Tallam <tmsriram@google.com>
5578 * tree-inline.c (expand_call_inline): Emit errors during
5579 early_inlining only if optimization is not turned on.
5581 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5585 * passes.c (init_optimization_passes): Move the construction of
5586 the pass hierarchy into a new passes.def file.
5588 * Makefile.in (passes.o): Add dependency on passes.def.
5590 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5592 * passes.c (init_optimization_passes): Introduce macros for
5593 constructing the tree of passes (INSERT_PASSES_AFTER,
5594 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
5595 TERMINATE_PASS_LIST).
5597 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
5598 Wei Mi <wmi@google.com>
5600 PR rtl-optimization/57878
5601 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
5603 (reload_pseudo_compare_func): Check nregs first for reload
5606 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5608 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
5610 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
5612 * read-rtl.c (validate_const_int): Once an invalid character is
5613 seen, quit the loop.
5615 * gengtype.c (write_roots): Similarly once we find the "deletable"
5616 or "if_marked" option.
5618 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5620 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
5621 "xtab" and "sat". Move value "clz" from here to ...
5622 (attriubte "type"): ... here.
5623 (satsi_<SAT:code>): Delete "insn" attribute.
5624 (satsi_<SAT:code>_shift): Likewise.
5625 (arm_zero_extendqisi2addsi): Likewise.
5626 (arm_extendqisi2addsi): Likewise.
5627 (clzsi2): Update for attribute changes.
5628 (rbitsi2): Likewise.
5629 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
5631 (arm_usatsihi): Likewise.
5632 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5634 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5636 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
5637 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
5638 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
5639 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
5640 in alphabetical order.
5641 (attribute "core_cycles"): Update for attribute changes.
5642 (arm_addsi3): Likewise.
5643 (addsi3_compare0): Likewise.
5644 (addsi3_compare0_scratch): Likewise.
5645 (addsi3_compare_op1): Likewise.
5646 (addsi3_compare_op2): Likewise.
5647 (compare_addsi2_op0): Likewise.
5648 (compare_addsi2_op1): Likewise.
5649 (addsi3_carryin_shift_<optab>): Likewise.
5650 (subsi3_carryin_shift): Likewise.
5651 (rsbsi3_carryin_shift): Likewise.
5652 (arm_subsi3_insn): Likewise.
5653 (subsi3_compare0): Likewise.
5654 (subsi3_compare): Likewise.
5655 (arm_andsi3_insn): Likewise.
5656 (thumb1_andsi3_insn): Likewise.
5657 (andsi3_compare0): Likewise.
5658 (andsi3_compare0_scratch): Likewise.
5659 (zeroextractsi_compare0_scratch
5660 (andsi_not_shiftsi_si): Likewise.
5661 (iorsi3_insn): Likewise.
5662 (iorsi3_compare0): Likewise.
5663 (iorsi3_compare0_scratch): Likewise.
5664 (arm_xorsi3): Likewise.
5665 (thumb1_xorsi3_insn): Likewise.
5666 (xorsi3_compare0): Likewise.
5667 (xorsi3_compare0_scratch): Likewise.
5668 (satsi_<SAT:code>_shift): Likewise.
5670 (arm_shiftsi3): Likewise.
5671 (shiftsi3_compare0): Likewise.
5672 (not_shiftsi): Likewise.
5673 (not_shiftsi_compare0): Likewise.
5674 (not_shiftsi_compare0_scratch): Likewise.
5675 (arm_one_cmplsi2): Likewise.
5676 (thumb_one_complsi2): Likewise.
5677 (notsi_compare0): Likewise.
5678 (notsi_compare0_scratch): Likewise.
5679 (thumb1_zero_extendhisi2): Likewise.
5680 (arm_zero_extendhisi2): Likewise.
5681 (arm_zero_extendhisi2_v6): Likewise.
5682 (arm_zero_extendhisi2addsi): Likewise.
5683 (thumb1_zero_extendqisi2): Likewise.
5684 (thumb1_zero_extendqisi2_v6): Likewise.
5685 (arm_zero_extendqisi2): Likewise.
5686 (arm_zero_extendqisi2_v6): Likewise.
5687 (arm_zero_extendqisi2addsi): Likewise.
5688 (thumb1_extendhisi2): Likewise.
5689 (arm_extendhisi2): Likewise.
5690 (arm_extendhisi2_v6): Likewise.
5691 (arm_extendqisi): Likewise.
5692 (arm_extendqisi_v6): Likewise.
5693 (arm_extendqisi2addsi): Likewise.
5694 (thumb1_extendqisi2): Likewise.
5695 (thumb1_movdi_insn): Likewise.
5696 (arm_movsi_insn): Likewise.
5697 (movsi_compare0): Likewise.
5698 (movhi_insn_arch4): Likewise.
5699 (movhi_bytes): Likewise.
5700 (arm_movqi_insn): Likewise.
5701 (thumb1_movqi_insn): Likewise.
5702 (arm32_movhf): Likewise.
5703 (thumb1_movhf): Likewise.
5704 (arm_movsf_soft_insn): Likewise.
5705 (thumb1_movsf_insn): Likewise.
5706 (movdf_soft_insn): Likewise.
5707 (thumb_movdf_insn): Likewise.
5708 (arm_cmpsi_insn): Likewise.
5709 (cmpsi_shiftsi): Likewise.
5710 (cmpsi_shiftsi_swp): Likewise.
5711 (arm_cmpsi_negshiftsi_si): Likewise.
5712 (movsicc_insn): Likewise.
5713 (movsfcc_soft_insn): Likewise.
5714 (arith_shiftsi): Likewise.
5715 (arith_shiftsi_compare0
5716 (arith_shiftsi_compare0_scratch
5717 (sub_shiftsi): Likewise.
5718 (sub_shiftsi_compare0
5719 (sub_shiftsi_compare0_scratch
5720 (and_scc): Likewise.
5721 (cond_move): Likewise.
5722 (if_plus_move): Likewise.
5723 (if_move_plus): Likewise.
5724 (if_move_not): Likewise.
5725 (if_not_move): Likewise.
5726 (if_shift_move): Likewise.
5727 (if_move_shift): Likewise.
5728 (if_shift_shift): Likewise.
5729 (if_not_arith): Likewise.
5730 (if_arith_not): Likewise.
5731 (cond_move_not): Likewise.
5732 (thumb1_ashlsi3): Set type attribute.
5733 (thumb1_ashrsi3): Likewise.
5734 (thumb1_lshrsi3): Likewise.
5735 (thumb1_rotrsi3): Likewise.
5736 (shiftsi3_compare0_scratch): Likewise.
5737 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5738 (neon_mov<mode>): Likewise.
5739 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5741 (thumb2_movsi_insn): Likewise.
5742 (thumb2_cmpsi_neg_shiftsi): Likewise.
5743 (thumb2_extendqisi_v6): Likewise.
5744 (thumb2_zero_extendhisi2_v6): Likewise.
5745 (thumb2_zero_extendqisi2_v6): Likewise.
5746 (thumb2_shiftsi3_short): Likewise.
5747 (thumb2_addsi3_compare0_scratch): Likewise.
5748 (orsi_not_shiftsi_si): Likewise.
5749 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5750 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5752 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5753 (1020alu_shift_op): Likewise.
5754 (1020alu_shift_reg_op): Likewise.
5755 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5756 (alu_shift_op): Likewise.
5757 (alu_shift_reg_op): Likewise.
5758 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5759 (11_alu_shift_op): Likewise.
5760 (11_alu_shift_reg_op): Likewise.
5761 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5762 (9_alu_shift_reg_op): Likewise.
5763 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5765 (cortex_a15_alu_shift): Likewise.
5766 (cortex_a15_alu_shift_reg): Likewise.
5767 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5769 (cortex_a5_alu_shift): Likewise.
5770 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5772 (cortex_a53_alu_shift): Likewise.
5773 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5775 (cortex_a7_alu_reg): Likewise.
5776 (cortex_a7_alu_shift): Likewise.
5777 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5779 (cortex_a8_alu_shift): Likewise.
5780 (cortex_a8_alu_shift_reg): Likewise.
5781 (cortex_a8_mov): Likewise.
5782 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5783 (cortex_a9_dp_shift): Likewise.
5784 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5786 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5788 (cortex_r4_mov): Likewise.
5789 (cortex_r4_alu_shift): Likewise.
5790 (cortex_r4_alu_shift_reg): Likewise.
5791 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5792 (526_alu_shift_op): Likewise.
5793 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5794 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5795 (626te_alu_shift_op): Likewise.
5796 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5797 (726te_alu_op): Likewise.
5798 (726te_alu_shift_op): Likewise.
5799 (726te_alu_shift_reg_op): Likewise.
5800 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5801 (mp626_alu_shift_op): Likewise.
5802 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5803 (pj4_alu_e1_conds): Likewise.
5804 (pj4_alu): Likewise.
5805 (pj4_alu_conds): Likewise.
5806 (pj4_shift): Likewise.
5807 (pj4_shift_conds): Likewise.
5808 (pj4_alu_shift): Likewise.
5809 (pj4_alu_shift_conds): Likewise.
5810 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5812 (cortexa7_older_only): Likewise.
5813 (cortexa7_younger): Likewise.
5815 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5817 * ipa-pure-const.c (generate_summary): Rename to...
5818 (pure_const_generate_summary): ... this.
5820 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5822 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5824 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5827 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5830 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5832 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5833 enabled without -march=zEC12.
5834 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5837 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5839 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
5840 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
5841 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
5842 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
5843 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
5844 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
5845 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
5847 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
5848 and ISA_HAS_NMADD3_NMSUB3.
5849 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
5850 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
5851 (nmsub4<mode>, nmsub3<mode>): Likewise.
5852 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
5854 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5856 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
5857 TARGET_MIPS5400 checking.
5859 2013-07-16 Jakub Jelinek <jakub@redhat.com>
5860 Peter Bergner <bergner@vnet.ibm.com>
5862 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
5863 registers in the comment.
5864 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
5865 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
5866 rather than FIRST_PSEUDO_REGISTERS.
5868 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
5870 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
5871 enable extra ISA flags with TARGET_HTM.
5873 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5875 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
5878 2013-07-15 Cong Hou <congh@google.com>
5880 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
5881 in compare function for sorting.
5883 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
5885 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
5886 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
5887 * config/rs6000/rs6000.opt: Add -mhtm option.
5888 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
5889 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
5890 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
5891 __HTM__ if the HTM instructions are available.
5892 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
5893 htm_spr_reg_operand): New define_predicates.
5894 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
5895 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
5897 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
5898 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
5899 HTM builtin functions.
5900 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
5901 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
5902 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
5903 (rs6000_builtin_mask_calculate): Likewise.
5904 (rs6000_option_override_internal): Likewise.
5905 (bdesc_htm): Add new HTM builtin support.
5906 (htm_spr_num): New function.
5907 (htm_spr_regno): Likewise.
5908 (rs6000_htm_spr_icode): Likewise.
5909 (htm_expand_builtin): Likewise.
5910 (htm_init_builtins): Likewise.
5911 (rs6000_expand_builtin): Add support for HTM builtin functions.
5912 (rs6000_init_builtins): Likewise.
5913 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
5915 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
5916 (TARGET_HTM, MASK_HTM): Define macros.
5917 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
5918 (FIXED_REGISTERS): Likewise.
5919 (CALL_USED_REGISTERS): Likewise.
5920 (CALL_REALLY_USED_REGISTERS): Likewise.
5921 (REG_ALLOC_ORDER): Likewise.
5922 (enum reg_class): Likewise.
5923 (REG_CLASS_NAMES): Likewise.
5924 (REG_CLASS_CONTENTS): Likewise.
5925 (REGISTER_NAMES): Likewise.
5926 (ADDITIONAL_REGISTER_NAMES): Likewise.
5927 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
5928 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
5929 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
5930 * config/rs6000/htm.md: New file.
5931 * config/rs6000/htmintrin.h: New file.
5932 * config/rs6000/htmxlintrin.h: New file.
5934 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
5936 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5937 Define SYMBOL_TINY_GOT, update comment.
5938 * config/aarch64/aarch64.c
5939 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
5940 (aarch64_expand_mov_immediate): Likewise.
5941 (aarch64_print_operand): Likewise.
5942 (aarch64_classify_symbol): Likewise.
5943 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
5944 (ldr_got_tiny): Define.
5946 2013-07-13 Tobias Grosser <tobias@grosser.es>
5948 PR tree-optimization/54094
5949 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
5950 scheduling dimension for the parallelism check from the polyhedral
5951 information in the AST.
5952 * graphite-dependences.c (carries_deps): Do not assume the schedule is
5955 2013-07-13 Jason Merrill <jason@redhat.com>
5957 * print-tree.c (debug_vec_tree): Use debug_raw.
5958 (debug_raw (vec<tree, va_gc> &)): New.
5959 (debug_raw (vec<tree, va_gc> *)): New.
5960 * tree.h: Declare them.
5962 2013-07-13 Bin Cheng <bin.cheng@arm.com>
5964 * ifcvt.c (ifcvt_after_combine): New static variable.
5965 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
5967 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
5968 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
5969 rest_of_handle_if_after_reload): Pass new argument for if_convert.
5971 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
5973 * config/mips/mips.c (mips_expand_call): Remove empty statement.
5975 2013-07-12 Michael Matz <matz@suse.de>
5978 * convert.c (convert_to_real): Reject non-float inner types.
5980 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
5982 * config/aarch64/aarch64-protos.h
5983 (aarch64_simd_immediate_valid_for_move): Remove.
5984 * config/aarch64/aarch64.c (simd_immediate_info): New member.
5985 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
5987 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
5989 2013-07-11 Steve Ellcey <sellcey@mips.com>
5991 * config/mips/mips.c (mips_conditional_register_usage): Do not
5992 use t[0-7] registers in MIPS16 mode when optimizing for size.
5994 2013-07-11 Sriraman Tallam <tmsriram@google.com>
5996 * config/i386/i386.c (dispatch_function_versions): Fix array
5997 indexing of function_version_info to match actual_versions.
5999 2013-07-11 Teresa Johnson <tejohnson@google.com>
6001 * vec.h (struct va_gc): Move release out-of-line.
6002 (va_gc::release): Call ggc_free on released vec.
6004 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6006 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
6007 Require GOT register as additional operand in UNSPEC.
6008 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
6009 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6010 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6011 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
6012 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
6013 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
6014 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
6016 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
6019 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
6020 name seen by assembler/linker rather if available.
6022 2013-07-11 Andreas Schwab <schwab@suse.de>
6024 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
6026 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
6028 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
6030 2013-07-10 Joseph Myers <joseph@codesourcery.com>
6032 * doc/tm.texi.in: Move hook documentation to ....
6033 * target.def: ... here.
6035 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
6037 * doc/tm.texi: Regenerate.
6039 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
6042 * doc/invoke.texi: Document -Wconditionally-supported.
6044 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6047 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
6050 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6053 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
6054 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
6055 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
6056 Remove duplicate devices.
6057 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
6058 * config/avr/t-multilib: Regenerate.
6059 * config/avr/avr-tables.opt: Regenerate.
6060 * doc/avr-mmcu.texi: Regenerate.
6062 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6065 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
6067 2013-07-10 Graham Stott <graham.stott@btinternet.com>
6069 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
6070 the cost of MULT when optimizing for size.
6072 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6074 * config/cr16/cr16-protos.h: Don't include target.h.
6076 2013-07-09 Joseph Myers <joseph@codesourcery.com>
6078 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
6079 adjust register size for TDmode and TFmode for VSX registers.
6081 2013-07-08 Kai Tietz <ktietz@redhat.com>
6084 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
6085 hook_bool_const_tree_true.
6087 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6089 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
6090 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
6091 * config/s390/s390.md: Define FPR*_REGNUM constants.
6092 Fix FPR2_REGNUM constant (18 -> 17).
6093 ("*trunc<BFP:mode><DFP_ALL:mode>2")
6094 ("*trunc<DFP_ALL:mode><BFP:mode>2")
6095 ("trunc<BFP:mode><DFP_ALL:mode>2")
6096 ("trunc<DFP_ALL:mode><BFP:mode>2")
6097 ("*extend<BFP:mode><DFP_ALL:mode>2")
6098 ("*extend<DFP_ALL:mode><BFP:mode>2")
6099 ("extend<BFP:mode><DFP_ALL:mode>2")
6100 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
6103 2013-07-08 Graham Stott <graham.stott@btinternet.com>
6105 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
6107 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6109 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
6110 and cfun_fpr_bit_p to cfun_fpr_save_p.
6111 (s390_frame_area, s390_register_info, s390_frame_info)
6112 (s390_emit_prologue, s390_emit_epilogue)
6113 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
6115 * config/s390/s390.h: Define *_REGNUM macros for floating point
6118 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
6120 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
6122 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
6124 PR rtl-optimization/57786
6125 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
6126 and break out of the loop when it is set to false.
6128 2013-07-08 Jakub Jelinek <jakub@redhat.com>
6131 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
6132 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
6133 (const_int 63)) 0)).
6134 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
6135 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
6136 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
6138 PR rtl-optimization/57829
6139 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
6140 mask bits outside of mode are just sign-extension from mode to HWI.
6142 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6144 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
6145 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
6146 adjust_address instead of change_address to keep info about alignment.
6147 (emit_strmov): Remove.
6148 (emit_memmov): New function.
6149 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
6150 (expand_movmem_epilogue): Likewise and return updated rtx for
6152 (expand_constant_movmem_prologue): Likewise and return updated rtx for
6153 destination and source.
6154 (decide_alignment): Refactor, handle vector_loop.
6155 (ix86_expand_movmem): Likewise.
6156 (ix86_expand_setmem): Likewise.
6157 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
6159 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
6161 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
6162 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
6164 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6166 * config/i386/sse.md (sse_movlhps): Change alternative 3
6167 of operand 2 to "m".
6169 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6172 * config/i386/sse.md (iptr): New mode attribute.
6173 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
6174 (<sse>_vm<plusminus_insn><mode>3): Ditto.
6175 (<sse>_vmmul<mode>3): Ditto.
6176 (<sse>_vmdiv<mode>3): Ditto.
6177 (sse_vmrcpv4sf2): Ditto.
6178 (<sse>_vmsqrt<mode>2): Ditto.
6179 (sse_vmrsqrtv4sf2): Ditto.
6180 (<sse>_vm<code><mode>3): Ditto.
6181 (avx_vmcmp<mode>3): Ditto.
6182 (<sse>_vmmaskcmp<mode>3): Ditto.
6183 (<sse>_comi): Ditto.
6184 (<sse>_ucomi): Ditto.
6185 (*xop_vmfrcz_<mode>): Ditto.
6186 (*fmai_fmadd_<mode>): Ditto.
6187 (*fmai_fmsub_<mode>): Ditto.
6188 (*fmai_fnmadd_<mode>): Ditto.
6189 (*fmai_fnmsub_<mode>): Ditto.
6190 (*fma4i_vmfmadd_<mode>): Ditto.
6191 (*fma4i_vmfmsub_<mode>): Ditto.
6192 (*fma4i_vmfnmadd_<mode>): Ditto.
6193 (*fma4i_vmfnmsub_<mode>): Ditto.
6194 (*xop_vmfrcz_<mode>): Ditto.
6195 (sse_cvtps2pi): Ditto.
6196 (sse_cvttps2pi): Ditto.
6197 (sse_cvtss2si): Ditto.
6198 (sse_cvtss2si_2): Ditto.
6199 (sse_cvtss2siq_2): Ditto.
6200 (sse_cvttss2si): Ditto.
6201 (sse_cvttss2siq): Ditto.
6202 (sse_cvtsd2si): Ditto.
6203 (sse_cvtsd2si_2): Ditto.
6204 (sse_cvtsd2siq_2): Ditto.
6205 (sse_cvttsd2si): Ditto.
6206 (sse_cvttsd2siq): Ditto.
6207 (sse_cvtsd2ss): Ditto.
6208 (sse_cvtss2sd): Ditto.
6209 (avx2_pbroadcast<mode>): Ditto.
6210 (avx2_pbroadcast<mode>_1): Ditto.
6211 (*avx_vperm_broadcast_v4sf): Ditto.
6213 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
6214 (sse_movlhps): Ditto.
6215 (sse_storehps): Ditto.
6216 (sse_loadhps): Ditto.
6217 (sse_storelps): Ditto.
6218 (sse_loadlps): Ditto.
6219 (*vec_concatv4sf): Ditto.
6220 (*vec_interleave_highv2df): Ditto.
6221 (*vec_interleave_lowv2df): Ditto.
6222 (*vec_extractv2df_1_sse): Ditto.
6223 (*vec_extractv2df_0_sse): Ditto.
6224 (sse2_storelpd): Ditto.
6225 (sse2_loadlpd): Ditto.
6226 (sse2_movsd): Ditto.
6227 (*vec_concatv4si): Ditto.
6228 (vec_concatv2di): Ditto.
6230 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
6231 for Intel asm dialect.
6232 (mmx_punpcklwd): Ditto.
6233 (mmx_punpckldq): Ditto.
6235 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
6236 for intel assembler dialect.
6238 2013-07-06 Jakub Jelinek <jakub@redhat.com>
6241 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
6242 for BUILT_IN_C{LZ,LRSB}*.
6243 * tree.h (CASE_INT_FN): Add FN##IMAX case.
6244 * tree-vrp.c (extract_range_basic): Handle
6245 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
6246 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
6247 fall thru to code calling set_value*.
6248 * builtins.c (expand_builtin): Remove *IMAX cases.
6249 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
6250 if width is bigger than 2*HWI.
6252 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
6254 PR rtl-optimization/55342
6255 * lra-int.h (lra_subreg_reload_pseudos): New.
6256 * lra.c: Add undoing optional reloads to the block diagram.
6257 (lra_subreg_reload_pseudos): New.
6258 (lra_optional_reload_pseudos): Change comments.
6259 (lra): Init and clear lra_subreg_reload_pseudos. Clear
6260 lra_optional_reload_pseudos after undo transformations.
6261 * lra-assigns.c (pseudo_prefix_title): New.
6262 (lra_setup_reg_renumber): Use it.
6263 (spill_for): Ditto. Check subreg reload pseudos too.
6264 (assign_by_spills): Consider subreg reload pseudos too.
6265 * lra-constraints.c (simplify_operand_subreg): Use
6266 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
6267 (curr_insn_transform): Recognize and do optional reloads.
6268 (undo_optional_reloads): New.
6269 (lra_undo_inheritance): Call undo_optional_reloads.
6271 2013-07-05 Thomas Quinot <quinot@adacore.com>
6273 * tree-complex.c (expand_complex_operations_1): Fix typo.
6275 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
6277 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
6278 (tune_params): New member 'const vec_costs'.
6279 * config/aarch64/aarch64.c (generic_vector_cost): New.
6280 (generic_tunings): New member 'generic_vector_cost'.
6281 (aarch64_builtin_vectorization_cost): New.
6282 (aarch64_add_stmt_cost): New.
6283 (TARGET_VECTORIZE_ADD_STMT_COST): New.
6284 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
6286 2013-07-03 Jakub Jelinek <jakub@redhat.com>
6289 * config/i386/predicates.md (vsib_address_operand): Disallow
6290 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
6292 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
6295 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
6296 expand_builtin_setjmp_receiver.
6297 (expand_label): Adjust, call expand_builtin_setjmp_receiver
6298 with NULL for the label parameter.
6299 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
6300 the frame-pointer. Adjust comments.
6301 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
6302 only if LABEL is non-NULL.
6304 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
6306 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
6307 (ARM_ABI_AAPCS64): Ditto.
6309 (ARM_DEFAULT_ABI): Ditto.
6311 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
6313 * config/aarch64/aarch64-builtins.c
6314 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
6315 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
6317 * config/aarch64/aarch64-simd.md
6318 (aarch64_ld1<VALL:mode>): New.
6319 (aarch64_st1<VALL:mode>): Likewise.
6320 * config/aarch64/arm_neon.h
6321 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
6323 2013-07-02 Sriraman Tallam <tmsriram@google.com>
6325 * config/i386/i386.c (gate_insert_vzeroupper): Check if
6327 (ix86_option_override_internal):Turn on all -mavx target flags by
6328 default as they are dependent on AVX anyway.
6330 2013-07-02 Cary Coutant <ccoutant@google.com>
6332 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
6334 (loc_checksum_ordered): Likewise.
6335 (hash_loc_operands): Remove inline keyword.
6337 2013-07-02 Jakub Jelinek <jakub@redhat.com>
6339 PR tree-optimization/57741
6340 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
6341 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
6342 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
6343 Allow REAL_CST step_exprs if flag_associative_math.
6344 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
6346 2013-07-02 Ian Bolton <ian.bolton@arm.com>
6348 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
6350 2013-07-02 Ian Bolton <ian.bolton@arm.com>
6352 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
6354 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6356 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
6358 (iorsi3_insn): Likewise.
6359 (arm_xorsi3): Likewise.
6361 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
6363 * arm.md (attribute "wtype"): Delete. Move attribute values from here
6365 (attribute "type"): ... here, and prefix with "wmmx_".
6366 (attribute "core_cycles"): Update for attribute changes.
6367 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
6368 (tbcstv4hi): Likewise.
6369 (tbcstv2si): Likewise.
6370 (iwmmxt_iordi3): Likewise.
6371 (iwmmxt_xordi3): Likewise.
6372 (iwmmxt_anddi3): Likewise.
6373 (iwmmxt_nanddi3): Likewise.
6374 (iwmmxt_arm_movdi): Likewise.
6375 (iwmmxt_movsi_insn): Likewise.
6376 (mov<mode>_internal): Likewise.
6377 (and<mode>3_iwmmxt): Likewise.
6378 (ior<mode>3_iwmmxt): Likewise.
6379 (xor<mode>3_iwmmxt): Likewise.
6380 (add<mode>3_iwmmxt): Likewise.
6381 (ssaddv8qi3): Likewise.
6382 (ssaddv4hi3): Likewise.
6383 (ssaddv2si3): Likewise.
6384 (usaddv8qi3): Likewise.
6385 (usaddv4hi3): Likewise.
6386 (usaddv2si3): Likewise.
6387 (sub<mode>3_iwmmxt): Likewise.
6388 (sssubv8qi3): Likewise.
6389 (sssubv4hi3): Likewise.
6390 (sssubv2si3): Likewise.
6391 (ussubv8qi3): Likewise.
6392 (ussubv4hi3): Likewise.
6393 (ussubv2si3): Likewise.
6394 (mulv4hi3_iwmmxt): Likewise.
6395 (smulv4hi3_highpart): Likewise.
6396 (umulv4hi3_highpart): Likewise.
6397 (iwmmxt_wmacs): Likewise.
6398 (iwmmxt_wmacsz): Likewise.
6399 (iwmmxt_wmacu): Likewise.
6400 (iwmmxt_wmacuz): Likewise.
6401 (iwmmxt_clrdi): Likewise.
6402 (iwmmxt_clrv8qi): Likewise.
6403 (iwmmxt_clr4hi): Likewise.
6404 (iwmmxt_clr2si): Likewise.
6405 (iwmmxt_uavgrndv8qi3): Likewise.
6406 (iwmmxt_uavgrndv4hi3): Likewise.
6407 (iwmmxt_uavgv8qi3): Likewise.
6408 (iwmmxt_uavgv4hi3): Likewise.
6409 (iwmmxt_tinsrb): Likewise.
6410 (iwmmxt_tinsrh): Likewise.
6411 (iwmmxt_tinsrw): Likewise.
6412 (iwmmxt_textrmub): Likewise.
6413 (iwmmxt_textrmsb): Likewise.
6414 (iwmmxt_textrmuh): Likewise.
6415 (iwmmxt_textrmsh): Likewise.
6416 (iwmmxt_textrmw): Likewise.
6417 (iwmxxt_wshufh): Likewise.
6418 (eqv8qi3): Likewise.
6419 (eqv4hi3): Likewise.
6420 (eqv2si3): Likewise.
6421 (gtuv8qi3): Likewise.
6422 (gtuv4hi3): Likewise.
6423 (gtuv2si3): Likewise.
6424 (gtv8qi3): Likewise.
6425 (gtv4hi3): Likewise.
6426 (gtv2si3): Likewise.
6427 (smax<mode>3_iwmmxt): Likewise.
6428 (umax<mode>3_iwmmxt): Likewise.
6429 (smin<mode>3_iwmmxt): Likewise.
6430 (umin<mode>3_iwmmxt): Likewise.
6431 (iwmmxt_wpackhss): Likewise.
6432 (iwmmxt_wpackwss): Likewise.
6433 (iwmmxt_wpackdss): Likewise.
6434 (iwmmxt_wpackhus): Likewise.
6435 (iwmmxt_wpackwus): Likewise.
6436 (iwmmxt_wpackdus): Likewise.
6437 (iwmmxt_wunpckihb): Likewise.
6438 (iwmmxt_wunpckihh): Likewise.
6439 (iwmmxt_wunpckihw): Likewise.
6440 (iwmmxt_wunpckilb): Likewise.
6441 (iwmmxt_wunpckilh): Likewise.
6442 (iwmmxt_wunpckilw): Likewise.
6443 (iwmmxt_wunpckehub): Likewise.
6444 (iwmmxt_wunpckehuh): Likewise.
6445 (iwmmxt_wunpckehuw): Likewise.
6446 (iwmmxt_wunpckehsb): Likewise.
6447 (iwmmxt_wunpckehsh): Likewise.
6448 (iwmmxt_wunpckehsw): Likewise.
6449 (iwmmxt_wunpckelub): Likewise.
6450 (iwmmxt_wunpckeluh): Likewise.
6451 (iwmmxt_wunpckeluw): Likewise.
6452 (iwmmxt_wunpckelsb): Likewise.
6453 (iwmmxt_wunpckelsh): Likewise.
6454 (iwmmxt_wunpckelsw): Likewise.
6455 (ror<mode>3): Likewise.
6456 (ashr<mode>3_iwmmxt): Likewise.
6457 (lshr<mode>3_iwmmxt): Likewise.
6458 (ashl<mode>3_iwmmxt): Likewise.
6459 (ror<mode>3_di): Likewise.
6460 (ashr<mode>3_di): Likewise.
6461 (lshr<mode>3_di): Likewise.
6462 (ashl<mode>3_di): Likewise.
6463 (iwmmxt_wmadds): Likewise.
6464 (iwmmxt_wmaddu): Likewise.
6465 (iwmmxt_tmia): Likewise.
6466 (iwmmxt_tmiaph): Likewise.
6467 (iwmmxt_tmiabb): Likewise.
6468 (iwmmxt_tmiatb): Likewise.
6469 (iwmmxt_tmiabt): Likewise.
6470 (iwmmxt_tmiatt): Likewise.
6471 (iwmmxt_tmovmskb): Likewise.
6472 (iwmmxt_tmovmskh): Likewise.
6473 (iwmmxt_tmovmskw): Likewise.
6474 (iwmmxt_waccb): Likewise.
6475 (iwmmxt_wacch): Likewise.
6476 (iwmmxt_waccw): Likewise.
6477 (iwmmxt_waligni): Likewise.
6478 (iwmmxt_walignr): Likewise.
6479 (iwmmxt_walignr0): Likewise.
6480 (iwmmxt_walignr1): Likewise.
6481 (iwmmxt_walignr2): Likewise.
6482 (iwmmxt_walignr3): Likewise.
6483 (iwmmxt_wsadb): Likewise.
6484 (iwmmxt_wsadh): Likewise.
6485 (iwmmxt_wsadbz): Likewise.
6486 (iwmmxt_wsadhz): Likewise.
6487 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
6488 (iwmmxt_wabsdiffb): Likewise.
6489 (iwmmxt_wabsdiffh): Likewise.
6490 (iwmmxt_wabsdiffw): Likewise.
6491 (iwmmxt_waddsubhx): Likewise
6492 (iwmmxt_wsubaddhx): Likewise.
6493 (addc<mode>3): Likewise.
6494 (iwmmxt_avg4): Likewise.
6495 (iwmmxt_avg4r): Likewise.
6496 (iwmmxt_wmaddsx): Likewise.
6497 (iwmmxt_wmaddux): Likewise.
6498 (iwmmxt_wmaddsn): Likewise.
6499 (iwmmxt_wmaddun): Likewise.
6500 (iwmmxt_wmulwsm): Likewise.
6501 (iwmmxt_wmulwum): Likewise.
6502 (iwmmxt_wmulsmr): Likewise.
6503 (iwmmxt_wmulumr): Likewise.
6504 (iwmmxt_wmulwsmr): Likewise.
6505 (iwmmxt_wmulwumr): Likewise.
6506 (iwmmxt_wmulwl): Likewise.
6507 (iwmmxt_wqmulm): Likewise.
6508 (iwmmxt_wqmulwm): Likewise.
6509 (iwmmxt_wqmulmr): Likewise.
6510 (iwmmxt_wqmulwmr): Likewise.
6511 (iwmmxt_waddbhusm): Likewise.
6512 (iwmmxt_waddbhusl): Likewise.
6513 (iwmmxt_wqmiabb): Likewise.
6514 (iwmmxt_wqmiabt): Likewise.
6515 (iwmmxt_wqmiatb): Likewise.
6516 (iwmmxt_wqmiatt): Likewise.
6517 (iwmmxt_wqmiabbn): Likewise.
6518 (iwmmxt_wqmiabtn): Likewise.
6519 (iwmmxt_wqmiatbn): Likewise.
6520 (iwmmxt_wqmiattn): Likewise.
6521 (iwmmxt_wmiabb): Likewise.
6522 (iwmmxt_wmiabt): Likewise.
6523 (iwmmxt_wmiatb): Likewise.
6524 (iwmmxt_wmiatt): Likewise.
6525 (iwmmxt_wmiabbn): Likewise.
6526 (iwmmxt_wmiabtn): Likewise.
6527 (iwmmxt_wmiatbn): Likewise.
6528 (iwmmxt_wmiattn): Likewise.
6529 (iwmmxt_wmiawbb): Likewise.
6530 (iwmmxt_wmiawbt): Likewise.
6531 (iwmmxt_wmiawtb): Likewise.
6532 (iwmmxt_wmiawtt): Likewise.
6533 (iwmmxt_wmiawbbn): Likewise.
6534 (iwmmxt_wmiawbtn): Likewise.
6535 (iwmmxt_wmiawtbn): Likewise.
6536 (iwmmxt_wmiawttn): Likewise.
6537 (iwmmxt_wmerge): Likewise.
6538 (iwmmxt_tandc<mode>3): Likewise.
6539 (iwmmxt_torc<mode>3): Likewise.
6540 (iwmmxt_torvsc<mode>3): Likewise.
6541 (iwmmxt_textrc<mode>3): Likewise.
6542 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
6543 (wmmxt_pack): Likewise.
6544 (wmmxt_mult_c1): Likewise.
6545 (wmmxt_mult_c2): Likewise.
6546 (wmmxt_alu_c1): Likewise.
6547 (wmmxt_alu_c2): Likewise.
6548 (wmmxt_alu_c3): Likewise.
6549 (wmmxt_transfer_c1): Likewise.
6550 (wmmxt_transfer_c2): Likewise.
6551 (wmmxt_transfer_c3): Likewise.
6552 (marvell_f_iwmmxt_wstr): Likewise.
6553 (marvell_f_iwmmxt_wldr): Likewise.
6555 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
6557 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
6559 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6562 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6563 * lra-constraints.c (need_for_split_p): Check call used hard regs
6564 living through calls.
6566 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
6567 call used regs for call insn.
6569 2013-06-28 Jakub Jelinek <jakub@redhat.com>
6572 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
6573 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
6575 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6577 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
6579 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6581 * lra-constraints.c (need_for_split_p): Check call used hard regs
6582 living through calls.
6584 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
6587 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
6588 to tie with any other modes. Eliminate Altivec vector mode tests,
6589 since these are a subset of ALTIVEC or VSX vector modes. Simplify
6590 code, to return 0 if testing MODE2 for a condition, if we've
6591 already tested MODE1 for the same condition.
6593 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6595 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
6598 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6600 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6601 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
6603 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6605 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
6607 (aarch64_symbolic_constant_p): Remove.
6608 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
6609 static. Fix line length and white space.
6610 (aarch64_symbolic_constant_p): Remove.
6611 * config/aarch64/predicates.md (aarch64_valid_symref):
6612 Use aarch64_classify_symbol_expression.
6614 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6616 * config/arm/constraints.md (Ts): New constraint.
6617 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
6619 (compare_scc): Use "Ts" constraint for operand 0.
6620 (ior_scc_scc): Likewise.
6621 (and_scc_scc): Likewise.
6622 (and_scc_scc_nodom): Likewise.
6623 (ior_scc_scc_cmp): Likewise for operand 7.
6624 (and_scc_scc_cmp): Likewise.
6625 * config/arm/thumb2.md (thumb2_movsi_insn):
6626 Add alternatives for 16-bit encodings.
6627 (thumb2_movhi_insn): Likewise.
6628 (thumb2_movsicc_insn): Likewise.
6629 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
6630 (thumb2_negscc): Use "Ts" constraint.
6631 Move mvn instruction outside cond_exec block.
6632 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
6633 for 16-bit encodings.
6635 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6637 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
6639 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
6640 (mulsi3subsi): Likewise.
6641 (mulsidi3adddi): Likewise.
6642 (mulsidi3_v6): Likewise.
6643 (umulsidi3_v6): Likewise.
6644 (umulsidi3adddi_v6): Likewise.
6645 (smulsi3_highpart_v6): Likewise.
6646 (umulsi3_highpart_v6): Likewise.
6647 (mulhisi3tb): Likewise.
6648 (mulhisi3bt): Likewise.
6649 (mulhisi3tt): Likewise.
6650 (maddhisi4): Likewise.
6651 (maddhisi4tb): Likewise.
6652 (maddhisi4tt): Likewise.
6653 (maddhidi4): Likewise.
6654 (maddhidi4tb): Likewise.
6655 (maddhidi4tt): Likewise.
6656 (zeroextractsi_compare0_scratch): Likewise.
6657 (insv_zero): Likewise.
6658 (insv_t2): Likewise.
6659 (anddi_notzesidi_di): Likewise.
6660 (anddi_notsesidi_di): Likewise.
6661 (andsi_notsi_si): Likewise.
6662 (iordi_zesidi_di): Likewise.
6663 (xordi_zesidi_di): Likewise.
6664 (andsi_iorsi3_notsi): Likewise.
6666 (smax_m1): Likewise.
6668 (not_shiftsi): Likewise.
6669 (unaligned_loadsi): Likewise.
6670 (unaligned_loadhis): Likewise.
6671 (unaligned_loadhiu): Likewise.
6672 (unaligned_storesi): Likewise.
6673 (unaligned_storehi): Likewise.
6674 (extv_reg): Likewise.
6675 (extzv_t2): Likewise.
6677 (udivsi3): Likewise.
6678 (arm_zero_extendhisi2addsi): Likewise.
6679 (arm_zero_extendqisi2addsi): Likewise.
6680 (compareqi_eq0): Likewise.
6681 (arm_extendhisi2_v6): Likewise.
6682 (arm_extendqisi2addsi): Likewise.
6683 (arm_movt): Likewise.
6684 (thumb2_ldrd): Likewise.
6685 (thumb2_ldrd_base): Likewise.
6686 (thumb2_ldrd_base_neg): Likewise.
6687 (thumb2_strd): Likewise.
6688 (thumb2_strd_base): Likewise.
6689 (thumb2_strd_base_neg): Likewise.
6690 (arm_negsi2): Add alternative for 16-bit encoding.
6691 (arm_one_cmplsi2): Likewise.
6693 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6695 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6696 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6697 (movdfcc): Likewise.
6698 * config/arm/vfp.md (*thumb2_movsf_vfp):
6699 Disable predication for arm_restrict_it.
6700 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6701 (*thumb2_movdfcc_vfp): Likewise.
6702 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6703 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6704 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6705 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6706 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6707 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6708 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6709 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6710 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6711 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6712 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6713 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6714 Disable predication for arm_restrict_it.
6716 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6718 * config/i386/bmiintrin.h (_bextr_u32): New.
6719 (_bextr_u64): Ditto.
6721 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6723 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6724 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6725 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6726 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6727 * config/mips/n32-elf.h: ...this new file.
6729 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6732 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6733 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6735 2013-06-27 Catherine Moore <clm@codesourcery.com>
6737 * config/mips/mips-tables.opt: Regenerate.
6738 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6739 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6740 * doc/invoke.texi: Add -m14kc.
6742 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6745 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6746 constraints of operand 1 and 2.
6749 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6750 to match RTL canonicalization. Swap predicates and
6751 constraints of operand 1 and 2.
6753 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6755 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6756 Process OP_INOUT regs for splitting too.
6758 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6760 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6761 decl before the loop, initialize to NULL.
6762 (vectorizable_load): Initialize ptr_incr to NULL.
6764 2013-06-27 Martin Jambor <mjambor@suse.cz>
6767 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6768 * ipa-ref.c (ipa_maybe_record_reference): New function.
6769 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6770 * ipa-cp.c (create_specialized_node): Record potential references from
6772 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6774 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6776 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6777 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6778 'mode' to force_reg.
6779 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6780 (aarch64_expand_mov_immediate): Likewise.
6782 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6784 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6785 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6787 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6789 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6790 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6791 (struct machine_function): Add tbegin_p.
6792 (s390_canonicalize_comparison): Fold CC mode compares to
6793 conditional jump if possible.
6794 (s390_emit_jump): Return the emitted jump.
6795 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6796 Handle CCRAWmode compares.
6797 (s390_option_override): Default to -mhtm if available.
6798 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6799 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6800 FPRs instead of df_regs_ever_live_p.
6801 (s390_optimize_nonescaping_tx): New function.
6802 (s390_init_frame_layout): Extend clobbered_regs array to cover
6804 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6805 (s390_expand_tbegin): New function.
6806 (enum s390_builtin): New enum definition.
6807 (code_for_builtin): New array definition.
6808 (s390_init_builtins): New function.
6809 (s390_expand_builtin): New function.
6810 (TARGET_INIT_BUILTINS): Define.
6811 (TARGET_EXPAND_BUILTIN): Define.
6812 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6813 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6814 (s390_alc_comparison): Likewise.
6815 * config/s390/s390-modes.def: Add CCRAWmode.
6816 * config/s390/s390.h (processor_flags): Add PF_TX.
6817 (TARGET_CPU_HTM): Define macro.
6818 (TARGET_HTM): Define macro.
6819 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6820 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6821 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6822 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6823 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6824 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6825 ("*cc_to_int"): Move up.
6826 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6827 constants other than 0.
6828 ("*ccraw_to_int"): New insn and splitter definition.
6829 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6830 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6831 ("tx_assist"): New expander.
6832 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6833 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6834 * config/s390/s390.opt: Add -mhtm option.
6835 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6836 * config/s390/htmxlintrin.h: New file.
6837 * config/s390/htmintrin.h: New file.
6838 * config/s390/s390intrin.h: New file.
6839 * doc/extend.texi: Document htm builtins.
6840 * config.gcc: Add the new header files to extra_headers.
6842 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
6844 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
6845 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
6847 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6848 Pat Haugen <pthaugen@us.ibm.com>
6849 Peter Bergner <bergner@vnet.ibm.com>
6851 * config/rs6000/power8.md: New.
6852 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
6853 setting for power8 entry.
6854 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
6855 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
6856 test for Power4/Power5 only.
6857 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
6859 (force_new_group): Adjust comment.
6860 * config/rs6000/rs6000.md: Include power8.md.
6862 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
6864 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
6865 * config/arm/arm-protos.h (arm_max_conditional_execute): New
6867 (tune_params): Update comment.
6868 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
6869 (arm_max_conditional_execute): New function.
6870 (thumb2_final_prescan_insn): Use max_insn_skipped and
6871 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
6873 2013-06-25 Jakub Jelinek <jakub@redhat.com>
6875 PR tree-optimization/57705
6876 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
6877 SSA_NAME step, provided that it is not defined inside the loop.
6878 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
6879 (get_initial_def_for_induction): Handle SSA_NAME IV step.
6881 2013-06-25 Martin Jambor <mjambor@suse.cz>
6884 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
6885 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
6887 (ipa_note_param_call): Initialize member_ptr flag.
6888 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
6889 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
6890 (ipa_write_indirect_edge_info): Stream member_ptr flag.
6891 (ipa_read_indirect_edge_info): Likewise.
6893 2013-06-25 Richard Biener <rguenther@suse.de>
6896 * passes.c (init_optimization_passes): Move pass_fold_builtins
6897 and pass_dce earlier with -Og.
6899 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
6901 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
6902 <BIT_FIELD_REF>: Remove trailing TAB.
6903 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
6906 2013-06-24 Martin Jambor <mjambor@suse.cz>
6908 PR tree-optimization/57358
6909 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
6910 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
6911 (ipa_analyze_params_uses): Generate pessimistic info when true.
6913 2013-06-24 Martin Jambor <mjambor@suse.cz>
6915 PR tree-optimization/57539
6916 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
6917 global.inlined_to of the new node to it. All callers changed.
6918 * ipa-inline-transform.c (clone_inlined_nodes): New variable
6919 inlining_into, pass it to cgraph_clone_node.
6920 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
6921 ipa_free_edge_args_substructures.
6922 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
6923 rdesc linked list. Do not assert rdesc edges have inlined caller.
6924 Assert we have found an rdesc in the rdesc list.
6926 2013-06-24 Richard Biener <rguenther@suse.de>
6928 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
6929 (pointer_set_lookup): Declare.
6930 (class pointer_map): New template class implementing a
6931 generic pointer to T map.
6932 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
6933 pointer_map<T>::contains, pointer_map<T>::insert,
6934 pointer_map<T>::traverse): New functions.
6935 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
6936 (pointer_set_lookup): New function.
6937 (pointer_set_contains): Use pointer_set_lookup.
6938 (pointer_set_insert): Likewise.
6939 (insert_aux): Remove.
6940 (struct pointer_map_t): Embed a pointer_set_t.
6941 (pointer_map_create): Adjust.
6942 (pointer_map_destroy): Likewise.
6943 (pointer_map_contains): Likewise.
6944 (pointer_map_insert): Likewise.
6945 (pointer_map_traverse): Likewise.
6946 * tree-streamer.h (struct streamer_tree_cache_d): Use a
6947 pointer_map<unsigned> instead of a pointer_map_t.
6948 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
6949 (streamer_tree_cache_lookup): Likewise.
6950 (streamer_tree_cache_create): Likewise.
6951 (streamer_tree_cache_delete): Likewise.
6952 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
6953 pointer_map<unsigned> instead of a pointer_map_t.
6954 (lto_init_tree_ref_encoder): Adjust.
6955 (lto_destroy_tree_ref_encoder): Likewise.
6956 * lto-section-out.c (lto_output_decl_index): Likewise.
6957 (lto_record_function_out_decl_state): Likewise.
6958 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
6960 2013-06-24 Richard Biener <rguenther@suse.de>
6962 PR tree-optimization/57488
6963 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
6965 2013-06-24 Alan Modra <amodra@gmail.com>
6967 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
6968 (gen_easy_altivec_constant): Likewise.
6969 * config/rs6000/predicates.md (easy_vector_constant_add_self,
6970 easy_vector_constant_msb): Likewise.
6972 2013-06-23 Jakub Jelinek <jakub@redhat.com>
6975 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
6976 add missing return true.
6978 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
6981 * config/sh/predicates.md (general_extend_operand): Invoke
6982 general_movsrc_operand for memory operands.
6983 (general_movsrc_operand): Allow reg+reg addressing, do not use
6984 general_operand for memory operands.
6986 2013-06-23 Sriraman Tallam <tmsriram@google.com>
6988 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
6989 when current target options does not apply.
6990 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
6991 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
6992 * config/i386/bmiintrin.h: Pass appropriate target
6993 attributes to header.
6994 * config/i386/mmintrin.h: Ditto.
6995 * config/i386/nmmintrin.h: Ditto.
6996 * config/i386/avx2intrin.h: Ditto.
6997 * config/i386/fxsrintrin.h: Ditto.
6998 * config/i386/tbmintrin.h: Ditto.
6999 * config/i386/xsaveintrin.h: Ditto.
7000 * config/i386/f16cintrin.h: Ditto.
7001 * config/i386/xtestintrin.h: Ditto.
7002 * config/i386/xsaveoptintrin.h: Ditto.
7003 * config/i386/bmi2intrin.h: Ditto.
7004 * config/i386/lzcntintrin.h: Ditto.
7005 * config/i386/smmintrin.h: Ditto.
7006 * config/i386/wmmintrin.h: Ditto.
7007 * config/i386/x86intrin.h: Remove all header include guards.
7008 * config/i386/prfchwintrin.h: Ditto.
7009 * config/i386/pmmintrin.h: Ditto.
7010 * config/i386/tmmintrin.h: Ditto.
7011 * config/i386/xmmintrin.h: Ditto.
7012 * config/i386/popcntintrin.h: Ditto.
7013 * config/i386/rdseedintrin.h: Ditto.
7014 * config/i386/ammintrin.h: Ditto.
7015 * config/i386/emmintrin.h: Ditto.
7016 * config/i386/immintrin.h: Remove all header include guards.
7017 * config/i386/fma4intrin.h: Ditto.
7018 * config/i386/lwpintrin.h: Ditto.
7019 * config/i386/xopintrin.h: Ditto.
7020 * config/i386/ia32intrin.h: Ditto.
7021 * config/i386/avxintrin.h: Ditto.
7022 * config/i386/rtmintrin.h: Ditto.
7023 * config/i386/fmaintrin.h: Ditto.
7024 * config/i386/mm3dnow.h: Ditto.
7026 2013-06-22 Sriraman Tallam <tmsriram@google.com>
7028 * common/config/i386/i386-common.c: Handle LZCNT.
7030 2013-06-22 Andi Kleen <ak@linux.intel.com>
7032 * doc/extend.texi: Use __atomic_store_n instead of
7033 __atomic_store in HLE example.
7035 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
7037 * config/sh/sh.c: Remove <cstdlib> workaround.
7039 2013-06-21 Andi Kleen <ak@linux.intel.com>
7041 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
7043 2013-06-21 Andi Kleen <ak@linux.intel.com>
7045 * doc/extend.texi: Document that __atomic_clear and
7046 __atomic_test_and_set should only be used with bool.
7048 2013-06-20 Jan Hubicka <jh@suse.cz>
7050 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
7052 * tree.c (decls_same_for_odr): New function.
7053 (same_for_edr): New function.
7054 (types_same_for_odr): New function.
7055 (get_binfo_at_offset): Use it.
7056 * tree.h (types_same_for_odr): Declare.
7058 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
7059 Jason Merrill <jason@redhat.com>
7061 * system.h: Include <cstdlib> as well as <stdlib.h>.
7063 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
7066 * config/i386/i386.c (construct_container): Report error if
7067 long double is used with disabled x87 float returns.
7069 2013-06-20 Jan Hubicka <jh@suse.cz>
7071 * lto-cgraph.c (input_symtab): Do not set cgraph state.
7073 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
7075 PR rtl-optimization/57425
7076 PR rtl-optimization/57569
7077 * alias.c (write_dependence_p): Remove parameters mem_mode and
7078 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
7079 Changed all callers.
7080 (canon_anti_dependence): Get comments and semantics in sync.
7081 Add parameter mem_canonicalized. Changed all callers.
7082 * rtl.h (canon_anti_dependence): Update prototype.
7084 2013-06-20 Richard Biener <rguenther@suse.de>
7086 * data-streamer-in.c (streamer_read_uhwi): Optimize single
7087 byte case, inline streamer_read_uchar and defer section
7090 2013-06-20 Richard Biener <rguenther@suse.de>
7092 PR tree-optimization/57584
7093 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
7094 SSA names into the expanded expression that take part in
7095 abnormal coalescing.
7097 2013-06-19 Sharad Singhai <singhai@google.com>
7099 * gcov.c (print_usage): Handle new option.
7100 (process_args): Ditto.
7101 (get_gcov_intermediate_filename): New function.
7102 (output_intermediate_file): New function.
7103 (output_gcov_file): New function
7104 (generate_results): Handle new option.
7105 (release_function): Relase demangled name.
7106 (read_graph_file): Handle demangled name.
7107 (output_lines): Ditto.
7108 * doc/gcov.texi: Document gcov intermediate format.
7110 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
7113 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
7114 (lra_emit_add): Use the functions. Add comment about Y as an
7117 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
7120 * collect2.c (collect_atexit): New.
7121 (collect_exit): Delete.
7122 (main): Register collect_atexit with atexit.
7123 (collect_wait): Change collect_exit to exit.
7125 * collect2.h (collect_exit): Delete.
7126 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
7128 2013-06-19 Wei Mi <wmi@google.com>
7130 PR rtl-optimization/57518
7131 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
7132 if regno is used in paradoxical subreg.
7133 (update_equiv_regs): Check pdx_subregs[regno] before
7134 set a reg to be equivalent with a mem.
7136 2013-06-19 Matthias Klose <doko@ubuntu.com>
7139 * file-find.h (find_a_file): Add a mode parameter.
7140 * file-find.c (find_a_file): Likewise.
7141 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
7142 with X_OK for the executables.
7143 * collect2.c (main): Call find_a_file with X_OK.
7145 2013-06-19 Steve Ellcey <sellcey@mips.com>
7148 * config/mips/mips.md (casesi_internal_mips16_<mode>):
7149 Use NEXT_INSN instead of next_real_insn.
7151 2013-06-19 Jan Hubicka <jh@suse.cz>
7153 * cgraph.h (const_value_known_p): Replace by ...
7154 (ctor_for_folding): .. this one.
7155 * cgraphunit.c (process_function_and_variable_attributes): Use it.
7156 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
7157 * expr.c (expand_expr_real_1): Likewise.
7158 (string_constant): Likewise.
7159 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
7160 * ipa.c (process_references): Likewise.
7161 (symtab_remove_unreachable_nodes): Likewise.
7162 * ipa-inline-analysis.c (param_change_prob): Likewise.
7163 * gimple-fold.c (canonicalize_constructor_val): Likewise.
7164 (get_base_constructor): Likwise.
7165 * varpool.c (varpool_remove_node): Likewise.
7166 (varpool_remove_initializer): LIkewise.
7167 (dump_varpool_node): LIkwise.
7168 (const_value_known_p): Rewrite to ...
7169 (ctor_for_folding): ... this one.
7171 2013-06-19 Jakub Jelinek <jakub@redhat.com>
7174 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
7175 PERSONALITY in $PATH derived prefixes.
7177 2013-06-19 Jeff Law <law@redhat.com>
7179 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
7182 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
7183 (simplify_bitwise_binary): Use it to simpify certain binary ops on
7186 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
7188 * config/arm/vfp.md: Move VFP instruction classification documentation
7190 * config/arm/arm.md: ... here. Update instruction classification
7193 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
7195 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
7196 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
7197 pattern. Use more efficient sequences on ARMv5 and Thumb2.
7199 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
7202 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7203 with NEXT_INSN. Use tablejump_p to check for jump table data
7206 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
7209 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
7210 that now in C++ the value is correct per the C++ standards.
7212 2013-06-19 Richard Biener <rguenther@suse.de>
7214 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
7217 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7220 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7223 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7224 with next_active_insn.
7226 2013-06-18 Sriraman Tallam <tmsriram@google.com>
7228 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
7229 functions are inlined during failures to flag an error.
7230 * tree-inline.c (expand_call_inline): Allow the error to be flagged
7231 in early inline pass.
7233 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
7235 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
7238 2013-06-18 Julian Brown <julian@codesourcery.com>
7240 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
7241 Permit virtual register pre-reload if !strict.
7242 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
7244 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
7246 * config/arm/neon.md (movmisalign<mode>): Use
7247 neon_perm_struct_or_reg_operand instead of
7248 neon_struct_or_register_operand.
7249 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
7250 neon_permissive_struct_operand instead of neon_struct_operand.
7251 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
7252 neon_vector_mem_operand.
7253 * config/arm/predicates.md (neon_struct_operand): Adjust call to
7254 neon_vector_mem_operand.
7255 (neon_permissive_struct_operand): New.
7256 (neon_struct_or_register_operand): Rename to...
7257 (neon_perm_struct_or_reg_operand): This. Adjust call to
7258 neon_vector_mem_operand.
7260 2013-06-18 Richard Biener <rguenther@suse.de>
7262 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
7263 * lto-streamer.h: Include pointer-set.h.
7264 (struct lto_decl_slot): Remove.
7265 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
7266 Remove next_index entry.
7267 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7268 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7269 (lto_init_tree_ref_encoder): Adjust.
7270 (lto_destroy_tree_ref_encoder): Likewise.
7271 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7272 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7273 (lto_output_decl_index): Adjust.
7274 (lto_new_out_decl_state): Likewise.
7275 (lto_record_function_out_decl_state): Likewise.
7276 * lto-streamer-out.c (copy_function): Likewise.
7278 2013-06-18 Richard Biener <rguenther@suse.de>
7280 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
7281 * cgraphunit.c: Include cfgloop.h.
7282 (init_lowered_empty_function): Initialize the loop tree.
7283 (assemble_thunk): Insert new BBs into loops.
7285 2013-06-18 Richard Biener <rguenther@suse.de>
7287 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
7288 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
7289 the map from cache entry to cache index optional.
7290 (streamer_tree_cache_replace_tree): Adjust accordingly.
7291 (streamer_tree_cache_append): Likewise.
7292 (streamer_tree_cache_delete): Likewise.
7293 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
7294 streamer cache map from cache entry to cache index.
7295 * lto-streamer-out.c (create_output_block): Adjust.
7297 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
7299 * config/arm/arm.md (attribute "insn"): Move multiplication and
7300 division attributes to...
7301 (attribute "type"): ... here. Remove mult.
7302 (attribute "mul32"): New attribute.
7303 (attribute "mul64"): Add umaal.
7304 (*arm_mulsi3): Update attributes.
7305 (*arm_mulsi3_v6): Likewise.
7306 (*thumb_mulsi3): Likewise.
7307 (*thumb_mulsi3_v6): Likewise.
7308 (*mulsi3_compare0): Likewise.
7309 (*mulsi3_compare0_v6): Likewise.
7310 (*mulsi_compare0_scratch): Likewise.
7311 (*mulsi_compare0_scratch_v6): Likewise.
7312 (*mulsi3addsi): Likewise.
7313 (*mulsi3addsi_v6): Likewise.
7314 (*mulsi3addsi_compare0): Likewise.
7315 (*mulsi3addsi_compare0_v6): Likewise.
7316 (*mulsi3addsi_compare0_scratch): Likewise.
7317 (*mulsi3addsi_compare0_scratch_v6): Likewise.
7318 (*mulsi3subsi): Likewise.
7319 (*mulsidi3adddi): Likewise.
7320 (*mulsi3addsi_v6): Likewise.
7321 (*mulsidi3adddi_v6): Likewise.
7322 (*mulsidi3_nov6): Likewise.
7323 (*mulsidi3_v6): Likewise.
7324 (*umulsidi3_nov6): Likewise.
7325 (*umulsidi3_v6): Likewise.
7326 (*umulsidi3adddi): Likewise.
7327 (*umulsidi3adddi_v6): Likewise.
7328 (*smulsi3_highpart_nov6): Likewise.
7329 (*smulsi3_highpart_v6): Likewise.
7330 (*umulsi3_highpart_nov6): Likewise.
7331 (*umulsi3_highpart_v6): Likewise.
7332 (mulhisi3): Likewise.
7333 (*mulhisi3tb): Likewise.
7334 (*mulhisi3bt): Likewise.
7335 (*mulhisi3tt): Likewise.
7336 (maddhisi4): Likewise.
7337 (*maddhisi4tb): Likewise.
7338 (*maddhisi4tt): Likewise.
7339 (maddhidi4): Likewise.
7340 (*maddhidi4tb): Likewise.
7341 (*maddhidi4tt): Likewise.
7343 (udivsi3): Likewise.
7344 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
7345 (thumb2_mulsi_short_compare0): Likewise.
7346 (thumb2_mulsi_short_compare0_scratch): Likewise.
7347 * config/arm/arm1020e.md (1020mult1): Update attribute change.
7348 (1020mult2): Likewise.
7349 (1020mult3): Likewise.
7350 (1020mult4): Likewise.
7351 (1020mult5): Likewise.
7352 (1020mult6): Likewise.
7353 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
7355 (cortex_a15_mult64): Likewise.
7356 (cortex_a15_sdiv): Likewise.
7357 (cortex_a15_udiv): Likewise.
7358 * config/arm/arm1026ejs.md (mult1): Update attribute change.
7364 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
7365 (pj4_ir_div): Likewise.
7366 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
7367 (11_mult2): Likewise.
7368 (11_mult3): Likewise.
7369 (11_mult4): Likewise.
7370 (11_mult5): Likewise.
7371 (11_mult6): Likewise.
7372 (11_mult7): Likewise.
7373 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
7374 (cortex_a8_mla): Likewise.
7375 (cortex_a8_mull): Likewise.
7376 (cortex_a8_smulwy): Likewise.
7377 (cortex_a8_smlald): Likewise.
7378 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
7379 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
7380 (cortex_r4_mul_3): Likewise.
7381 (cortex_r4_mla_4): Likewise.
7382 (cortex_r4_mla_3): Likewise.
7383 (cortex_r4_smlald): Likewise.
7384 (cortex_r4_mull): Likewise.
7385 (cortex_r4_sdiv): Likewise.
7386 (cortex_r4_udiv): Likewise.
7387 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
7388 (cortex_a7_idiv): Likewise.
7389 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
7390 (9_mult2): Likewise.
7391 (9_mult3): Likewise.
7392 (9_mult4): Likewise.
7393 (9_mult5): Likewise.
7394 (9_mult6): Likewise.
7395 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
7396 (cortex_a53_sdiv): Likewise.
7397 (cortex_a53_udiv): Likewise.
7398 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
7399 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
7400 (mp626_mult2): Likewise.
7401 (mp626_mult3): Likewise.
7402 (mp626_mult4): Likewise.
7403 * config/arm/fa526.md (526_mult1): Update attribute change.
7404 (526_mult2): Likewise.
7405 * config/arm/arm-generic.md (mult): Update attribute change.
7406 (mult_ldsched_strongarm): Likewise.
7407 (mult_ldsched): Likewise.
7408 (multi_cycle): Likewise.
7409 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
7410 * config/arm/fa606te.md (606te_mult1): Update attribute change.
7411 (606te_mult2): Likewise.
7412 (606te_mult3): Likewise.
7413 (606te_mult4): Likewise.
7414 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
7415 (cortex_a9_mac16): Likewise.
7416 (cortex_a9_multiply): Likewise.
7417 (cortex_a9_mac): Likewise.
7418 (cortex_a9_multiply_long): Likewise.
7419 * config/arm/fa626te.md (626te_mult1): Update attribute change.
7420 (626te_mult2): Likewise.
7421 (626te_mult3): Likewise.
7422 (626te_mult4): Likewise.
7424 2013-06-18 Richard Biener <rguenther@suse.de>
7427 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
7429 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7432 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7433 with next_active_insn.
7435 2013-06-18 Alan Modra <amodra@gmail.com>
7437 * config/rs6000/rs6000.h (enum data_align): New.
7438 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
7439 (DATA_ABI_ALIGNMENT): Define.
7440 (CONSTANT_ALIGNMENT): Correct comment.
7441 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
7442 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
7444 2013-06-17 David Malcolm <dmalcolm@redhat.com>
7446 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
7447 ATTRIBUTE_UNUSED marking.
7449 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
7451 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
7452 alternative and update.
7453 (aarch64_dup_lanedi): Delete.
7454 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
7455 * config/aarch64/aarch64-simd-builtins.def: Update.
7457 2013-06-17 Richard Biener <rguenther@suse.de>
7459 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
7460 (lto_input_scc): Declare.
7461 (lto_input_tree_1): Likewise.
7462 (struct lto_stats_d): Add num_tree_bodies_output and
7463 num_pickle_refs_output.
7464 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
7465 (lto_read_tree_1): Split out from ...
7466 (lto_read_tree): ... this.
7467 (lto_input_scc): New function.
7468 (lto_input_tree_1): Split out from ...
7469 (lto_input_tree): ... this. Handle LTO_tree_scc.
7470 (lto_data_in_create): Create the streamer cache without hashes.
7471 * lto-streamer-out.c (create_output_block): Create the streamer
7472 cache with hashes when not doing WPA.
7473 (lto_write_tree_1): Split out from ...
7474 (lto_write_tree): ... this.
7475 (get_symbol_initial_value): New function.
7476 (lto_output_tree_1): Split out from ...
7477 (lto_output_tree): ... this. Write trees as series of SCCs
7478 using a DFS walk via DFS_write_tree.
7479 (struct sccs, struct scc_entry): New types.
7480 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
7481 (DFS_write_tree_body): New function.
7482 (DFS_write_tree): Likewise.
7483 (hash_tree): Likewise.
7484 (scc_entry_compare): Likewise.
7485 (hash_scc): Likewise.
7486 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
7487 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
7488 TREE_CHAIN as regular reference.
7489 (streamer_read_integer_cst): Remove.
7490 (streamer_get_pickled_tree): Adjust.
7491 * tree-streamer-out.c (streamer_write_chain): Disable streaming
7492 of DECL_EXTERNALs in BLOCK_VARS for now.
7493 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
7495 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
7496 Add hash value argument and record that if hashes are recorded
7498 (streamer_tree_cache_insert_1): Adjust.
7499 (streamer_tree_cache_insert): Likewise.
7500 (streamer_tree_cache_insert_at): Rename to ...
7501 (streamer_tree_cache_replace_tree): ... this and adjust.
7502 (streamer_tree_cache_append): Adjust.
7503 (record_common_node): Likewise.
7504 (streamer_tree_cache_create): Add argument whether to
7505 record hash values together with trees.
7506 (streamer_tree_cache_delete): Adjust.
7507 * tree-streamer.h (struct streamer_tree_cache_d): Add
7509 (streamer_read_integer_cst): Remove.
7510 (streamer_tree_cache_insert): Adjust.
7511 (streamer_tree_cache_append): Likewise.
7512 (streamer_tree_cache_insert_at): Rename to ...
7513 (streamer_tree_cache_replace_tree): ... this and adjust.
7514 (streamer_tree_cache_create): Add argument whether to record hashes.
7515 (streamer_tree_cache_get): Rename to ...
7516 (streamer_tree_cache_get_tree): ... this.
7517 (streamer_tree_cache_get_hash): New function.
7518 * tree.c (cache_integer_cst): New function.
7519 * tree.h (cache_integer_cst): Declare.
7520 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
7521 * lto-symtab.c (lto_varpool_replace_node): Only release
7522 DECL_INITIAL of non-prevailing decls.
7523 * varpool.c (varpool_remove_initializer): Do not release
7524 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
7526 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7528 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
7529 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
7530 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
7531 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
7532 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
7533 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
7534 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
7535 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
7536 instead of TARGET_64BIT.
7537 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
7538 Require ISA_HAS_<D>DIV.
7540 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
7542 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
7543 (mips*-*-linux*): Move default with_llsc setting to where other
7545 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
7547 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
7548 Likewise. Remove default with_tune setting. Move default float
7549 setting to its own block. Handle with_llsc in the same block as above.
7551 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
7553 PR rtl-optimization/57425
7554 PR rtl-optimization/57569
7555 * alias.c (write_dependence_p): Add new parameters mem_mode,
7556 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
7557 Changed all callers.
7558 (canon_anti_dependence): New function.
7559 * cse.c (check_dependence): Use canon_anti_dependence.
7560 * cselib.c (cselib_invalidate_mem): Likewise.
7561 * rtl.h (canon_anti_dependence): Declare.
7563 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7565 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
7566 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
7567 ".set mips3" for 64-bit targets.
7569 2013-06-15 Dehao Chen <dehao@google.com>
7571 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
7572 * gimple-low.c (gimple_check_call_matching_types): Likewise.
7573 (gimple_check_call_args): Likewise.
7574 * value-prof.c (check_ic_target): Likewise.
7575 * ipa-inline.c (early_inliner): Likewise.
7576 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
7577 * cgraph.c (cgraph_create_edge_1): Likewise.
7578 (cgraph_make_edge_direct): Likewise.
7580 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
7583 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
7584 rs6000_output_move_128bit to handle emitting quad memory
7585 operations. Set attribute length to 8 bytes.
7587 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
7589 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
7591 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
7592 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
7593 (aarch64_<su>mlsl<mode>): Likewise.
7595 2013-06-14 Mike Stump <mikestump@comcast.net>
7597 * Makefile.in (TARGET_H): Add insn-codes.h.
7599 2013-06-14 Alan Modra <amodra@gmail.com>
7603 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
7604 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
7605 bitfield expansion when EXPAND_MEMORY.
7606 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
7608 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
7610 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
7611 test for clearing quad memory on 32-bit later.
7613 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
7615 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
7616 (fold_negate_expr): Likewise.
7617 (fold_real_zero_addition_p): Handle vectors.
7618 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
7620 2013-06-14 Alan Modra <amodra@gmail.com>
7622 * varasm.c (force_const_mem): Revert 2013-06-07 change.
7624 2013-06-13 Jan Hubicka <jh@suse.cz>
7626 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
7627 Local comdats are not externally visible.
7628 * symtab.c (dump_symtab_base): Dump externally visible.
7629 (verify_symtab_base): Verify back links in the symtab hash.
7631 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7633 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
7634 CONVERT_EXPR as equal nodes.
7636 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7638 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
7640 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
7642 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
7643 Generalize to complex and vector.
7644 * tree.c (build_all_ones_cst): New function.
7645 * tree.h (build_all_ones_cst): Declare it.
7647 2013-06-13 Alan Modra <amodra@gmail.com>
7649 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
7650 * config/rs6000/rs6000.md (signbittf2): New insn.
7651 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
7652 (abstf2_internal, cmptf_internal2): Likewise.
7653 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
7655 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7656 Pat Haugen <pthaugen@us.ibm.com>
7657 Peter Bergner <bergner@vnet.ibm.com>
7659 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7660 power8 byte, half-word, and quad-word atomic instructions.
7661 (emit_store_conditional): Likewise.
7662 (rs6000_expand_atomic_compare_and_swap): Likewise.
7663 (rs6000_expand_atomic_op): Likewise.
7665 * config/rs6000/sync.md (larx): Add new modes for power8.
7667 (AINT): New mode iterator to include TImode as well as normal
7668 integer modes on power8.
7669 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7670 that VSX registers are not considered. Use AINT mode iterator
7671 instead of INT1 to allow inclusion of quad word atomic operations
7673 (load_locked<mode>): Likewise.
7674 (store_conditional<mode>): Likewise.
7675 (atomic_compare_and_swap<mode>): Likewise.
7676 (atomic_exchange<mode>): Likewise.
7677 (atomic_nand<mode>): Likewise.
7678 (atomic_fetch_<fetchop_name><mode>): Likewise.
7679 (atomic_nand_fetch<mode>): Likewise.
7680 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7682 (ATOMIC): On power8, add QImode, HImode modes.
7683 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7684 modes that promote to SImode.
7685 (load_lockedti): Convert TImode arguments to PTImode, so that we
7686 get a guaranteed even/odd register pair.
7687 (load_lockedpti): Likewise.
7688 (store_conditionalti): Likewise.
7689 (store_conditionalpti): Likewise.
7691 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7692 atomic load/store instructions.
7695 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7697 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7699 (insn_count): New attribute, with most cases extracted from...
7700 (length): ...here. Redefine most cases in terms of insn_count.
7701 (single_insn): Delete.
7702 (can_delay): Use insn_count to check for single instructions.
7703 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7704 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7705 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7706 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7707 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7708 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7709 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7710 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7711 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7712 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7713 rather than "length".
7714 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7715 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7716 Use "insn_count" rather than "length".
7717 * config/mips/mips-dsp.md
7718 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7719 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7722 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7724 PR tree-optimization/57361
7725 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7727 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7729 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7731 (aarch64_simd_combine<mode>): New instruction expansion.
7732 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7734 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7735 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7737 2013-06-12 Jan Hubicka <jh@suse.cz>
7739 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7740 decl has when in streaming stage.
7741 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7742 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7744 2013-06-12 Roland Stigge <stigge@antcom.de>
7747 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7749 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7751 PR tree-optimization/57537
7752 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7753 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7755 2013-06-12 Richard Biener <rguenther@suse.de>
7757 * data-streamer.h (streamer_write_char_stream): CSE
7758 obs->current_pointer.
7759 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7760 streamer_write_char_stream manually and optimize the resulting loop.
7761 (streamer_write_hwi_stream): Likewise.
7763 2013-06-12 Jan Hubicka <jh@suse.cz>
7765 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7766 * cgraph.h (varpool_create_empty_node): Declare.
7767 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7769 * symtab.c (symtab_unregister_node): Be lax about missin entries
7771 (symtab_get_node): Update comment.
7772 * varpool.c (varpool_create_empty_node): Break out from ...
7773 (varpool_node_for_decl): ... here.
7774 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7776 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7778 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7779 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7780 part. Use straight-line flow at the end.
7781 <COMPONENT_REF>: Remove superfluous else.
7782 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7784 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7787 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7788 target hook even for !TREE_PUBLIC decls. If no resolution info
7789 is available, return false for common and external decls.
7791 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7793 * config/rl78/constraints.md (U): New constraint.
7794 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7797 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7800 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7801 to allow returning address to AT_PLATFORM name.
7803 2013-06-11 Jan Hubicka <jh@suse.cz>
7805 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7806 * cgraph.h (symtab_node_base): Add weakref flag.
7807 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7808 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7809 (output_weakrefs): Use weakref flag.
7810 * fold-const.c (simple_operand_p): Handle WEAK.
7811 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7812 * ipa.c (varpool_externally_visible_p): Drop weakref.
7813 (function_and_variable_visibility): Update comment; fix weakref
7814 sanity checks; do not clear DECL_WEAK on them.
7815 * lto-cgraph.c (lto_output_node): update.
7816 (lto_output_varpool_node): Update.
7817 (input_overwrite_node): Update.
7818 (input_node): Update.
7819 (input_varpool_node): Update.
7820 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7821 (lto_symtab_merge_symbols): Add sanity check.
7822 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7823 * passes.c (rest_of_decl_compilation): Set static flag, too.
7824 * symtab.c (dump_symtab_base): Dump weakref.
7825 (verify_symtab_base): Sanity check weakrefs.
7826 (symtab_make_decl_local): Remove duplicated code.
7827 (symtab_alias_ultimate_target): Simplify.
7828 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7830 2013-06-11 DJ Delorie <dj@redhat.com>
7832 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7833 (rl78_unwind_word_mode): New.
7835 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7837 * final.c (debug_prefix_maps): Make static.
7839 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7841 * function.c (initial_trampoline): Remove stray copy.
7843 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
7845 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
7847 2013-06-11 Martin Jambor <mjambor@suse.cz>
7849 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
7850 within bounds at the beginning of the function.
7852 2013-06-11 Alan Modra <amodra@gmail.com>
7854 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
7856 (get_named_section): Don't NULL !DECL_P decl.
7858 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
7860 * doc/invoke.texi (core-avx2): Document.
7862 (atom): Updated with MOVBE.
7864 2013-06-11 Richard Biener <rguenther@suse.de>
7866 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
7868 2013-06-11 Anton Blanchard <anton@samba.org>
7870 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
7871 correct shift value in little-endian mode.
7873 2013-06-11 Jakub Jelinek <jakub@redhat.com>
7876 * varasm.c (get_variable_align): Move #endif to the right place.
7878 2013-06-10 Cary Coutant <ccoutant@google.com>
7880 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
7881 for hash so that hash table traversal order is deterministic.
7883 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
7884 Pat Haugen <pthaugen@us.ibm.com>
7885 Peter Bergner <bergner@vnet.ibm.com>
7887 * config/rs6000/vector.md (GPR move splitter): Do not split moves
7888 of vectors in GPRS if they are direct moves or quad word load or
7891 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
7893 (direct_move_p): Likewise.
7894 (quad_load_store_p): Likewise.
7896 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
7897 classes into bins based on the physical register type.
7898 (reg_class_to_reg_type): Likewise.
7899 (IS_STD_REG_TYPE): Likewise.
7900 (IS_FP_VECT_REG_TYPE): Likewise.
7901 (reload_fpr_gpr): Arrays to determine what insn to use if we can
7902 use direct move instructions.
7903 (reload_gpr_vsx): Likewise.
7904 (reload_vsx_gpr): Likewise.
7905 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
7906 information that is a simplification of register classes. Also
7907 precalculate direct move reload helpers.
7908 (direct_move_p): New function to return true if the operation can
7909 be done as a direct move instruciton.
7910 (quad_load_store_p): New function to return true if the operation
7911 is a quad memory operation.
7912 (rs6000_legitimize_address): If quad memory, only allow register
7913 indirect for TImode addresses.
7914 (rs6000_legitimate_address_p): Likewise.
7915 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
7916 (rs6000_reload_register_type): Likewise.
7917 (register_to_reg_type): Return register type.
7918 (rs6000_secondary_reload_simple_move): New helper function for
7919 secondary reload and secondary memory needed to identify anything
7920 that is a simple move, and does not need reloading.
7921 (rs6000_secondary_reload_direct_move): New helper function for
7922 secondary reload to identify cases that can be done with several
7923 instructions via the direct move instructions.
7924 (rs6000_secondary_reload_move): New helper function for secondary
7925 reload to identify moves between register types that can be done.
7926 (rs6000_secondary_reload): Add support for quad memory operations
7927 and for direct move.
7928 (rs6000_secondary_memory_needed): Likewise.
7929 (rs6000_debug_secondary_memory_needed): Change argument names.
7930 (rs6000_output_move_128bit): New function to return the move to
7931 use for 128-bit moves, including knowing about the various
7932 limitations of quad memory operations.
7934 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
7935 memory operations. call rs6000_output_move_128bit for the actual
7936 instruciton(s) to generate.
7937 (vsx_movti_64bit): Likewise.
7939 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
7940 (UNSPEC_P8V_MTVSRWZ): Likewise.
7941 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
7942 (UNSPEC_P8V_MTVSRD): Likewise.
7943 (UNSPEC_P8V_XXPERMDI): Likewise.
7944 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
7945 (UNSPEC_FUSION_GPR): Likewise.
7946 (FMOVE128_GPR): New iterator for direct move.
7947 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
7950 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
7951 loads and direct move instructions.
7952 (zero_extendsidi2_lfiwzx): Likewise.
7953 (extendsidi2_lfiwax): Likewise.
7954 (extendsidi2_nocell): Likewise.
7955 (floatsi<mode>2_lfiwax): Likewise.
7957 (floatunssi<mode>2_lfiwzx): Likewise.
7959 (fix_trunc<mode>_stfiwx): Likewise.
7960 (fixuns_trunc<mode>_stfiwx): Likewise.
7961 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
7962 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
7963 (parity<mode>2_cmpb): Set length/type attr.
7964 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
7965 for 'mr.' to fast_compare.
7966 (bpermd_<mode>): Change type attr to popcnt.
7967 (p8_fmrgow_<mode>): New insns for power8 direct move support.
7968 (p8_mtvsrwz_1): Likewise.
7969 (p8_mtvsrwz_2): Likewise.
7970 (reload_fpr_from_gpr<mode>): Likewise.
7971 (p8_mtvsrd_1): Likewise.
7972 (p8_mtvsrd_2): Likewise.
7973 (p8_xxpermdi_<mode>): Likewise.
7974 (reload_vsx_from_gpr<mode>): Likewise.
7975 (reload_vsx_from_gprsf): Likewise.
7976 (p8_mfvsrd_3_<mode>): LIkewise.
7977 (reload_gpr_from_vsx<mode>): Likewise.
7978 (reload_gpr_from_vsxsf): Likewise.
7979 (p8_mfvsrd_4_disf): Likewise.
7980 (multi-word GPR splits): Do not split direct moves or quad memory
7983 2013-06-10 David Malcolm <dmalcolm@redhat.com>
7985 * tree-into-ssa.c (interesting_blocks): Make static.
7987 2013-06-10 Jakub Jelinek <jakub@redhat.com>
7990 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
7991 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
7992 Use DATA_ABI_ALIGNMENT for that case instead if defined.
7993 (get_variable_align): New function.
7994 (get_variable_section, emit_bss, emit_common,
7995 assemble_variable_contents, place_block_symbol): Use
7996 get_variable_align instead of DECL_ALIGN.
7997 (assemble_noswitch_variable): Add align argument, use it
7998 instead of DECL_ALIGN.
7999 (assemble_variable): Adjust caller. Use get_variable_align
8000 instead of DECL_ALIGN.
8001 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
8003 (DATA_ABI_ALIGNMENT): Define.
8004 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
8005 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
8006 opt is false, only return the psABI mandated alignment increase.
8007 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
8008 (DATA_ABI_ALIGNMENT): ... this.
8009 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
8010 (DATA_ABI_ALIGNMENT): ... this.
8011 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
8012 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
8013 (DATA_ABI_ALIGNMENT): ... this.
8014 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
8015 * doc/tm.texi: Regenerated.
8017 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
8019 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
8020 cmp_code to construct REG_EQUAL note.
8022 2013-06-09 Jakub Jelinek <jakub@redhat.com>
8025 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
8026 that operands[2] doesn't overlap with operands[0].
8028 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
8029 Jan Hubicka <jh@suse.cz>
8031 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
8032 hack to mark symbols as used.
8034 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
8036 PR rtl-optimization/57559
8037 * lra-constraints.c (process_alt_operands): Don't discourage
8038 memory with known offset for offsetable memory constraint.
8039 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
8041 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
8043 * varasm.c (struct oc_local_state): Reorder fields.
8044 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
8045 and adjust accordingly.
8046 (output_constructor): Reorder initialization code and adjust call to
8047 output_constructor_bitfield.
8049 2013-06-07 Jan Hubicka <jh@suse.cz>
8051 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
8053 2013-06-07 David Malcolm <dmalcolm@redhat.com>
8055 * tree-object-size.c (unknown): Make const.
8057 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8059 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
8060 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
8061 for last alternative in the cpu_facility attribute.
8063 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8066 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
8067 (xordi3): Change operand 2 constraint to arm_xordi_operand.
8068 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
8069 * config/arm/constraints.md (Dg): New constraint.
8070 * config/arm/neon.md (xordi3_neon): Remove.
8071 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
8072 * config/arm/predicates.md (arm_xordi_operand): New predicate.
8074 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8076 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
8077 Clean up alternatives.
8079 2013-06-07 Alan Modra <amodra@gmail.com>
8081 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
8084 2013-06-07 Alan Modra <amodra@gmail.com>
8086 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
8088 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8090 * config/arm/constraints.md (Df): New constraint.
8091 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
8092 Correct length attribute for last two alternatives.
8094 2013-06-07 Alan Modra <amodra@gmail.com>
8096 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8097 override user -mfp-in-toc.
8098 (offsettable_ok_by_alignment): Consider just the current access
8099 rather than the whole object, unless BLKmode. Handle
8100 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
8101 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
8102 for -mcmodel=medium.
8103 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
8104 override user -mfp-in-toc or -msum-in-toc. Default to
8105 -mno-fp-in-toc for -mcmodel=medium.
8107 2013-06-06 DJ Delorie <dj@redhat.com>
8109 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
8110 TARGET_VALID_POINTER_MODE.
8112 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
8113 Pat Haugen <pthaugen@us.ibm.com>
8114 Peter Bergner <bergner@vnet.ibm.com>
8116 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8117 Document new power8 builtins.
8119 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
8120 condition code register, to allow 128-bit logical operations to be
8121 done in the VSX or GPR registers.
8122 (nor<mode>3): Use the canonical form for nor.
8123 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
8124 vclz*, and vpopcnt* vector instructions.
8125 (nand<mode>3): Likewise.
8126 (orc<mode>3): Likewise.
8127 (clz<mode>2): LIkewise.
8128 (popcount<mode>2): Likewise.
8130 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
8131 that only the GPRs are recognized.
8133 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8134 support for new power8 builtins.
8136 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
8138 (xscvdpspn): Likewise.
8144 (vpopcnt): Likewise.
8145 (vpopcntb): Likewise.
8146 (vpopcnth): Likewise.
8147 (vpopcntw): Likewise.
8148 (vpopcntd): Likewise.
8153 (eqv_v16qi3): Likewise.
8154 (eqv_v8hi3): Likewise.
8155 (eqv_v4si3): Likewise.
8156 (eqv_v2di3): Likewise.
8157 (eqv_v4sf3): Likewise.
8158 (eqv_v2df3): Likewise.
8160 (nand_v16qi3): Likewise.
8161 (nand_v8hi3): Likewise.
8162 (nand_v4si3): Likewise.
8163 (nand_v2di3): Likewise.
8164 (nand_v4sf3): Likewise.
8165 (nand_v2df3): Likewise.
8167 (orc_v16qi3): Likewise.
8168 (orc_v8hi3): Likewise.
8169 (orc_v4si3): Likewise.
8170 (orc_v2di3): Likewise.
8171 (orc_v4sf3): Likewise.
8172 (orc_v2df3): Likewise.
8174 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
8175 allow power8 quad mode in 64-bit.
8176 (rs6000_builtin_vectorized_function): Add support to vectorize
8177 ISA 2.07 count leading zeros, population count builtins.
8178 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
8179 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
8180 (builtin_function_type): Add vgbbd builtin function which takes an
8182 (altivec_expand_vec_perm_const): Add support for new power8 merge
8185 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
8186 that does not include TImdoe for use with 32-bit.
8187 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
8189 (UNSPEC_VSX_CVDPSPN): Likewise.
8190 (vsx_xscvdpspn): Likewise.
8191 (vsx_xscvspdpn): Likewise.
8192 (vsx_xscvdpspn_scalar): Likewise.
8193 (vsx_xscvspdpn_directmove): Likewise.
8194 (vsx_and<mode>3): Split logical operations into 32-bit and
8195 64-bit. Add support to do logical operations on TImode as well as
8196 VSX vector types. Allow logical operations to be done in either
8197 VSX registers or in general purpose registers in 64-bit mode. Add
8198 splitters if GPRs were used. For AND, add clobber of CCmode to
8199 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
8201 (vsx_and<mode>3_32bit): Likewise.
8202 (vsx_and<mode>3_64bit): Likewise.
8203 (vsx_ior<mode>3): Likewise.
8204 (vsx_ior<mode>3_32bit): Likewise.
8205 (vsx_ior<mode>3_64bit): Likewise.
8206 (vsx_xor<mode>3): Likewise.
8207 (vsx_xor<mode>3_32bit): Likewise.
8208 (vsx_xor<mode>3_64bit): Likewise.
8209 (vsx_one_cmpl<mode>2): Likewise.
8210 (vsx_one_cmpl<mode>2_32bit): Likewise.
8211 (vsx_one_cmpl<mode>2_64bit): Likewise.
8212 (vsx_nor<mode>3): Likewise.
8213 (vsx_nor<mode>3_32bit): Likewise.
8214 (vsx_nor<mode>3_64bit): Likewise.
8215 (vsx_andc<mode>3): Likewise.
8216 (vsx_andc<mode>3_32bit): Likewise.
8217 (vsx_andc<mode>3_64bit): Likewise.
8218 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
8219 and xxlorc instructions.
8220 (vsx_eqv<mode>3_64bit): Likewise.
8221 (vsx_nand<mode>3_32bit): Likewise.
8222 (vsx_nand<mode>3_64bit): Likewise.
8223 (vsx_orc<mode>3_32bit): Likewise.
8224 (vsx_orc<mode>3_64bit): Likewise.
8226 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
8228 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
8230 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
8231 (p8_vmrgow): Likewise.
8232 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
8233 GPRs to be split under VSX.
8234 (p8v_clz<mode>2): Add power8 count leading zero support.
8235 (p8v_popcount<mode>2): Add power8 population count support.
8236 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
8239 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
8242 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
8244 (vec_nand): Likewise.
8245 (vec_vclz): Likewise.
8246 (vec_vclzb): Likewise.
8247 (vec_vclzd): Likewise.
8248 (vec_vclzh): Likewise.
8249 (vec_vclzw): Likewise.
8250 (vec_vgbbd): Likewise.
8251 (vec_vmrgew): Likewise.
8252 (vec_vmrgow): Likewise.
8253 (vec_vpopcnt): Likewise.
8254 (vec_vpopcntb): Likewise.
8255 (vec_vpopcntd): Likewise.
8256 (vec_vpopcnth): Likewise.
8257 (vec_vpopcntw): Likewise.
8259 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8261 PR rtl-optimization/57468
8262 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
8265 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8267 PR rtl-optimization/57459
8268 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
8269 type when setting live regs.
8271 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8273 * config/s390/s390.opt (mlra): New option.
8274 * config/s390/s390.c (s390_decompose_address): Check displacement
8275 for all registers for LRA.
8276 (s390_secondary_reload): Don't used secondary reloads for LRA.
8277 (s390_lra_p): New function.
8278 (TARGET_LRA_P): Define.
8279 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
8280 of attribute cpu_facility to zarch for the last alternative.
8281 (*cmpmem_short): Ditto.
8283 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
8285 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
8286 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
8287 (arm_expand_prologue): Likewise.
8289 2013-06-06 Teresa Johnson <tejohnson@google.com>
8292 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
8293 as this is now done by redirect_edge_and_branch_force.
8294 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
8295 barriers, and fix interaction with splitting.
8296 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
8297 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
8298 reflect changes made in the routine.
8299 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
8300 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
8301 since this is called in cfglayout mode, and replace partition fixup
8302 with assert as that is now done by force_nonfallthru_and_redirect.
8303 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
8304 already be marked with region crossing note.
8305 (insert_section_boundary_note): Make non-static, gate on flag
8306 has_bb_partition, rewrite to also check for multiple partitions.
8307 (rest_of_handle_reorder_blocks): Remove call to
8308 insert_section_boundary_note, now done later during free_cfg.
8309 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
8310 * bb-reorder.h (insert_section_boundary_note): Declare.
8311 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
8312 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
8313 invoke insert_section_boundary_note.
8314 (try_redirect_by_replacing_jump): Remove unnecessary
8315 check for region crossing note.
8316 (fixup_partition_crossing): New function.
8317 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
8318 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
8319 in non-cfglayout mode.
8320 (force_nonfallthru_and_redirect): Fixup partition boundaries,
8321 remove old code that tried to do this. Emit barrier correctly
8322 when we are in cfglayout mode.
8323 (last_bb_in_partition): New function.
8324 (rtl_split_edge): Correctly fixup partition boundaries.
8325 (commit_one_edge_insertion): Remove old code that tried to
8326 fixup region crossing edge since this is now handled in
8327 split_block, and set up insertion point correctly since
8328 block may now end in a jump.
8329 (verify_hot_cold_block_grouping): Guard against checking when not in
8330 linearized RTL mode.
8331 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
8333 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
8334 rtl_verify_flow_info, so not called in cfglayout mode.
8335 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
8336 (fixup_reorder_chain): Remove old code that attempted to fixup region
8337 crossing note as this is now handled in force_nonfallthru_and_redirect.
8338 (duplicate_insn_chain): Don't duplicate switch section notes.
8339 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
8341 * basic-block.h (emit_barrier_after_bb): Declare.
8343 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8345 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
8346 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
8347 arm_usatsihi): Adjust alternatives for arm_restrict_it.
8349 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8351 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
8353 * config/arm/ldmstm.md: Regenerate.
8355 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8357 * config/arm/sync.md (atomic_loaddi_1):
8358 Disable predication for arm_restrict_it.
8359 (arm_load_exclusive<mode>): Likewise.
8360 (arm_load_exclusivesi): Likewise.
8361 (arm_load_exclusivedi): Likewise.
8362 (arm_load_acquire_exclusive<mode>): Likewise.
8363 (arm_load_acquire_exclusivesi): Likewise.
8364 (arm_load_acquire_exclusivedi): Likewise.
8365 (arm_store_exclusive<mode>): Likewise.
8366 (arm_store_exclusive<mode>): Likewise.
8367 (arm_store_release_exclusivedi): Likewise.
8368 (arm_store_release_exclusive<mode>): Likewise.
8370 2013-06-06 Richard Biener <rguenther@suse.de>
8372 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
8374 (lto_tag_is_tree_code_p): Adjust.
8375 (lto_tag_is_gimple_code_p): Likewise.
8376 (lto_gimple_code_to_tag): Likewise.
8377 (lto_tag_to_gimple_code): Likewise.
8378 (lto_tree_code_to_tag): Likewise.
8379 (lto_tag_to_tree_code): Likewise.
8380 * data-streamer.h (streamer_write_hwi_in_range): Use
8381 uhwi streaming to stream the normalized range.
8382 (streamer_read_hwi_in_range): Likewise.
8384 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8386 * config/arm/arm.md (enabled_for_depr_it): New attribute.
8387 (predicable_short_it): Likewise.
8388 (predicated): Likewise.
8389 (enabled): Handle above.
8390 (define_cond_exec): Set predicated attribute to yes.
8392 2013-06-05 Mike Stump <mikestump@comcast.net>
8394 * gdbinit.in (__FUNCTION__): Add.
8396 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
8398 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
8399 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
8401 2013-06-05 Jan Hubicka <jh@suse.cz>
8403 * varasm.c (mark_decl_referenced): Revert the removal until targets
8406 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
8408 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
8409 instead of mark_decl_referenced.
8411 2013-06-05 Jan Hubicka <jh@suse.cz>
8413 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
8414 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
8415 and symtab_used_from_object_file_p.
8416 (cgraph_make_node_local_1): Clear forced_by_abi.
8417 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
8418 * cgraph.h (symtab_node_base): Add forced_by_abi.
8419 (decide_is_variable_needed): Remove.
8420 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
8421 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
8422 (decide_is_symbol_needed): ... this one; handle symbols in general;
8423 always analyze virtuals; honnor forced_by_abi.
8424 (cgraph_finalize_function): Update.
8425 (varpool_finalize_decl): Update.
8426 (symbol_defined_and_needed): Remove.
8427 (analyze_functions): Update.
8428 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8429 output_refs, input_overwrite_node): Handle forced_by_abi.
8430 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
8431 (address_taken_from_non_vtable_p): ... this one.
8432 (comdat_can_be_unshared_p_1): New function.
8433 (cgraph_comdat_can_be_unshared_p): Rename to ...
8434 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
8435 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
8436 (function_and_variable_visibility): Clear forced_by_abi as needed.
8437 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
8438 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
8439 * symtab.c (dump_symtab_base): Dump forced_by_abi.
8440 * varpool.c (decide_is_variable_needed): Remove.
8442 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8444 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
8445 (arm_option_override): Override arm_restrict_it where appropriate.
8446 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
8447 * config/arm/arm.opt (mrestrict-it): New command-line option.
8448 * doc/invoke.texi: Document -mrestrict-it.
8450 2013-06-05 David Malcolm <dmalcolm@redhat.com>
8452 * tsan.c (tsan_atomic_table): Make const.
8454 2013-06-05 Richard Biener <rguenther@suse.de>
8456 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
8457 index associated with the tree we are supposed to replace.
8458 * tree-streamer-out.c (pack_ts_base_value_fields): Output
8459 TREE_ASM_WRITTEN as zero for everything but SSA names.
8461 2013-06-05 David Malcolm <dmalcolm@redhat.com>
8463 * tree-ssa-structalias.c (call_stmt_vars): Make static.
8465 2013-06-04 Jan Hubicka <jh@suse.cz>
8467 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
8468 (input_node, input_varpool_node): Handle correctly external same
8470 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
8471 nodes at ltrans stage.
8473 2013-06-04 Jan Hubicka <jh@suse.cz>
8475 * ipa-inline.c (update_caller_keys): Fix availability test.
8476 (update_callee_keys): Likewise.
8477 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
8478 to follow ELF standard.
8480 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
8482 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
8483 (mips64r5900el-*-elf*): New configurations.
8484 * config/mips/mips-cpus.def (r5900): New processor.
8485 * config/mips/mips-tables.opt: Regenerate.
8486 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
8487 (mips_issue_rate): Handle PROCESSOR_R5900.
8488 (mips_reorg_process_insns): Force reorder mode for the R5900.
8489 * config/mips/mips.h (TARGET_MIPS5900): Define.
8490 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
8492 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
8494 * config/mips/mips.md (processor): Add r5900.
8495 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
8497 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8499 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
8500 into function to generate MOVI instruction.
8501 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
8502 (aarch64_preferred_simd_mode): Turn into wrapper.
8503 (aarch64_output_scalar_simd_mov_immediate): New function.
8504 * config/aarch64/aarch64-protos.h: Add prototype for above.
8506 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8508 * config/aarch64/aarch64.c (simd_immediate_info): Remove
8509 element_char member.
8510 (sizetochar): Return signed char.
8511 (aarch64_simd_valid_immediate): Remove elchar and other
8512 unnecessary variables.
8513 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
8514 Calculate element_char as required.
8515 * config/aarch64/aarch64-protos.h: Update and move prototype
8516 for aarch64_output_simd_mov_immediate.
8517 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
8520 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8522 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
8523 information completed by aarch64_simd_valid_immediate.
8524 (aarch64_legitimate_constant_p): Update arguments.
8525 (aarch64_simd_valid_immediate): Work with struct rather than many
8527 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
8528 (aarch64_simd_make_constant): Update arguments.
8529 (aarch64_output_simd_mov_immediate): Work with struct rather than
8530 many pointers. Output immediate directly rather than as operand.
8531 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8533 * config/aarch64/constraints.md (Dn): Update arguments.
8535 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8537 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
8539 (aarch64_simd_immediate_valid_for_move): Remove.
8540 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
8541 (aarch64_simd_make_constant): Update call.
8542 (aarch64_output_simd_mov_immediate): Update call.
8543 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8545 * config/aarch64/constraints.md (Dn): Update call.
8547 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8549 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
8550 return type to bool for prototype.
8551 (aarch64_legitimate_constant_p): Check for true instead of not -1.
8552 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
8553 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
8555 2013-06-04 Catherine Moore <clm@codesourcery.com>
8557 * config/mips/mips.opt (meva): New.
8558 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
8559 (ASM_SPEC): Handle -meva.
8560 * doc/invoke.texi (meva): Document.
8562 2013-06-04 Alan Modra <amodra@gmail.com>
8564 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
8567 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8569 * rtl.def: Add extra fourth optional field to define_cond_exec.
8570 * gensupport.c (process_one_cond_exec): Process attributes from
8572 * doc/md.texi: Document fourth field in define_cond_exec.
8574 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
8576 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
8577 out the processing order as in store_bit_field_1.
8579 2013-06-04 Jan Hubicka <jh@suse.cz>
8582 * cgraphunit.c (cgraph_process_same_body_aliases): Create
8583 non-VAR_DECL node if it does not exist yet.
8585 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
8587 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
8588 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
8589 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
8590 target_cpu_default setting.
8592 2013-06-03 Teresa Johnson <tejohnson@google.com>
8594 * dumpfile.c (opt_info_switch_p): Change -fopt-info
8595 default to -fopt-info=optimized instead of all.
8596 * doc/invoke.texi: Ditto.
8597 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
8598 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
8599 (execute_vect_slp): Emit BB vectorization success under
8600 MSG_OPTIMIZED_LOCATIONS.
8601 * tree-vect-slp.c (vect_slp_transform_bb): Change
8602 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
8603 * tree-vect-loop.c (vect_transform_loop): Ditto.
8605 2013-06-03 Jason Merrill <jason@redhat.com>
8608 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8609 Use TARGET_EXPR for C++.
8611 2013-06-03 Jakub Jelinek <jakub@redhat.com>
8613 PR rtl-optimization/57268
8614 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
8615 if DEBUG_INSN_P (insn).
8618 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8620 PR rtl-optimization/57268
8621 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8622 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8624 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8626 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
8627 (ix86_avoid_lea_for_addr): Likewise.
8628 (exact_dependency_1): Likewise.
8629 (ix86_adjust_cost): Likewise.
8630 (swap_top_of_ready_list): Fix formatting and !reload_completed check
8632 (do_reorder_for_imul): Fix typo, formatting and
8633 !reload_completed check removed.
8634 (ix86_sched_reorder): Fix typo and formatting.
8635 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
8638 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
8640 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
8642 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
8644 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
8645 <STRING_CST>: Likewise.
8646 <VECTOR_CST>: Likewise.
8648 2013-06-01 Janus Weil <janus@gcc.gnu.org>
8649 Mikael Morin <mikael@gcc.gnu.org>
8651 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
8652 * config.in: Regenerated.
8653 * configure: Regenerated.
8655 2013-06-01 Jan Hubicka <jh@suse.cz>
8658 * cgraphunit.c (compile): When weakref is not supported,
8659 set up transparent aliases before final output pass.
8660 * varasm.c (assemble_alias): Do not try to do it here.
8662 2013-06-01 Jan Hubicka <jh@suse.cz>
8665 * passes.c (for_per_function): Skip unanalyzed functions.
8667 2013-06-01 Jan Hubicka <jh@suse.cz>
8669 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8670 (lto_symtab_merge_symbols_1): ... this one.
8671 (lto_symtab_merge_cgraph_nodes): Rename to ...
8672 (lto_symtab_merge_symbols): ... this one; simplify.
8673 * cgraph.c (same_body_aliases_done): Rename to ...
8674 (cpp_implicit_aliases_done): ... this one.
8675 (cgraph_create_function_alias): Update.
8676 (cgraph_same_body_alias): Update.
8677 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8678 (verify_edge_corresponds_to_fndecl): Simplify.
8679 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8680 (cgraph_node): Remove same_body_alias.
8681 (varpool_node): Remove alias_of and extra_name_alias.
8682 (same_body_aliases_done): Rename to ..
8683 (cpp_implicit_aliases_done): ... this one.
8684 (symtab_alias_ultimate_target): Add default parameter.
8685 (symtab_resolve_alias): New function.
8686 (fixup_same_cpp_alias_visibility): Declare.
8687 (cgraph_function_node): Add default parameter.
8688 (cgraph_node_asm_name): Likewise.
8689 (cgraph_function_or_thunk_node): Add default parameter; do
8690 not ICE when it is NULL.
8691 (varpool_variable_node): Likewise.
8692 * tree-emutls.c (create_emultls_var): Update.
8693 (ipa_lower_emutls): Update.
8694 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8695 (cgraph_reset_node): Reset alias info.
8696 (cgraph_finalize_function): Update.
8697 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8698 (analyze_function): Simplify.
8699 (cgraph_process_same_body_aliases): Simplify.
8700 (analyze_functions): Fixup same body aliases.
8701 (handle_alias_pairs): Simplify.
8702 (assemble_thunk): Update.
8703 (assemble_thunks_and_aliases): Update.
8704 (output_weakrefs): Rewrite.
8705 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8706 (lto_output_varpool_node): Likewise.
8707 (compute_ltrans_boundary): Remve assert.
8708 (get_alias_symbol): New functoin.
8709 (input_node): Rewrite alias handling.
8710 (input_varpool_node): Likewise.
8711 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8712 * ipa.c (process_references): Handle weakrefs correctly.
8713 (symtab_remove_unreachable_nodes): Likewise.
8714 * trans-mem.c (get_cg_data): Update.
8715 (ipa_tm_create_version_alias): Update.
8716 (ipa_tm_execute): Update.
8717 * symtab.c (dump_symtab_base): Dump aliases.
8718 (verify_symtab_base): Verify aliases.
8719 (symtab_node_availability): New function.
8720 (symtab_alias_ultimate_target): Simplify.
8721 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8722 handle all the fixup cases.
8723 (symtab_resolve_alias): New function.
8724 * passes.c (ipa_write_summaries): Handle weakrefs.
8725 * varpool.c (varpool_analyze_node): Simplify.
8726 (assemble_aliases): Update.
8727 (varpool_create_variable_alias): Simplify.
8728 (varpool_extra_name_alias): Simplify.
8729 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8730 (lto_symtab_merge_symbols): ... this one.
8732 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8735 PR rtl-optimization/57268
8736 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8737 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8739 2013-06-01 Tobias Burnus <burnus@net-b.de>
8742 2013-05-31 Tobias Burnus <burnus@net-b.de>
8745 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8748 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8750 PR rtl-optimization/57268
8751 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8752 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8754 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8756 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8757 unordered comparison operators when -fno-trapping-math is in effect
8759 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8760 and implement unordered comparison operators properly on the e500.
8762 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8764 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8765 for constant scalar integers.
8766 (simplify_relational_operation_1): Likewise.
8768 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8770 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8771 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8774 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8775 Igor Zamyatin <igor.zamyatin@intel.com>
8777 Silvermont (SLM) architecture performance tuning.
8778 * config/i386/i386.h (enum ix86_tune_indices): Add
8779 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8780 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8782 * config/i386/i386.c (initial_ix86_tune_features)
8783 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8784 (ix86_lea_outperforms): Handle Silvermont tuning.
8785 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8787 (ix86_use_lea_for_mov): Likewise.
8788 (ix86_avoid_lea_for_addr): Likewise.
8789 (ix86_lea_for_add_ok): Likewise.
8790 (exact_dependency_1): New function.
8791 (exact_store_load_dependency): Likewise.
8792 (ix86_adjust_cost): Handle Silvermont tuning.
8793 (do_reoder_for_imul): Likewise.
8794 (swap_top_of_ready_list): New function.
8795 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8797 * config/i386/i386.md (peepholes that split memory operand in fp
8800 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8802 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8803 Remove un-necessary braces.
8805 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8807 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8808 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8810 2013-05-31 Tobias Burnus <burnus@net-b.de>
8813 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8815 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8818 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8819 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8820 * config/arm/neon.md (iordi3_neon): Remove.
8821 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8822 * config/arm/predicates.md (imm_for_neon_logic_operand):
8823 Move to earlier in the file.
8824 (neon_logic_op2): Likewise.
8825 (arm_iordi_operand_neon): New predicate.
8827 2013-05-31 Richard Biener <rguenther@suse.de>
8829 PR tree-optimization/57478
8830 PR tree-optimization/57453
8831 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8834 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8836 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8837 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
8839 2013-05-30 Tobias Burnus <burnus@net-b.de>
8840 Thomas Koenig <tkoenig@gcc.gnu.org>
8843 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
8844 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
8846 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
8848 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
8850 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
8852 * target.def (register_usage_leveling_p): New hook.
8853 * targhooks.c (default_register_usage_leveling_p): New.
8854 * targhooks.h (default_register_usage_leveling_p): New prototype.
8855 * lra-assigns.c (register_usage_leveling_p): Use the hook.
8856 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
8857 * doc/tm.texi: Update.
8858 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
8860 2013-05-30 Ian Bolton <ian.bolton@arm.com>
8862 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
8863 (*insv_reg<mode>): New define_insn.
8865 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
8867 PR rtl-optimization/57439
8868 * postreload.c (move2add_valid_value_p): Check that we have
8869 a zero subreg_regno_offset when accessing the register in
8872 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8873 Igor Zamyatin <igor.zamyatin@intel.com>
8875 Silvermont (SLM) architecture pipeline model, tuning and
8877 * config.gcc: Add slm config options and target.
8879 * config/i386/slm.md: New.
8881 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
8883 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
8885 (ix86_target_macros_internal): Likewise.
8887 * gcc/config/i386/i386.c (slm_cost): New cost.
8888 (m_SLM): New macro flag.
8889 (initial_ix86_tune_features): Set m_SLM.
8890 (x86_accumulate_outgoing_args): Likewise.
8891 (x86_arch_always_fancy_math_387): Likewise.
8892 (processor_target_table): Add slm cost.
8893 (cpu_names): Add slm cpu name.
8894 (x86_option_override_internal): Set SLM ISA.
8895 (ix86_issue_rate): New case PROCESSOR_SLM.
8896 (ia32_multipass_dfa_lookahead): Likewise.
8897 (fold_builtin_cpu): Add slm.
8899 * config/i386/i386.h (TARGET_SLM): New target macro.
8900 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
8901 (processor_type): Add PROCESSOR_SLM.
8903 * config/i386/i386.md (cpu): Add new value "slm".
8904 (slm.md): Include slm.md.
8906 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
8907 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8909 * config/arm/arm-protos.h: Add and update function protos.
8910 * config/arm/arm.c (use_simple_return_p): New added.
8911 (thumb2_expand_return): Check simple_return flag.
8912 * config/arm/arm.md: Add simple_return and conditional simple_return.
8913 * config/arm/iterators.md: Add iterator for return and simple_return.
8915 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8917 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
8918 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
8919 (arm_emit_vfp_multi_reg_pop): Likewise.
8920 (thumb2_emit_ldrd_pop): Likewise.
8921 (arm_expand_epilogue): Add misc REG_CFA notes.
8922 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
8924 2013-05-29 Lawrence Crowl <crowl@google.com>
8926 * config/arm/t-arm: Update for below.
8928 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
8929 Change type to hash_table. Update dependent calls and types.
8931 * config/i386/t-cygming: Update for below.
8933 * config/i386/t-interix: Update for below.
8935 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
8936 Change type to hash_table. Update dependent calls and types.
8937 (i386_find_on_wrapper_list::wrappers): Likewise.
8939 * config/ia64/t-ia64: Update for below.
8941 * config/ia64/ia64.c (bundle_state_table):
8942 Change type to hash_table. Update dependent calls and types.
8944 * config/mips/mips.c (mips_reorg_process_insns::htab):
8945 Change type to hash_table. Update dependent calls and types.
8947 * config/sol2.c (solaris_comdat_htab):
8948 Change type to hash_table. Update dependent calls and types.
8950 * config/t-sol2: Update for above.
8952 2013-05-29 Teresa Johnson <tejohnson@google.com>
8954 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
8955 functions are not yet marked as defined.
8957 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
8958 Pat Haugen <pthaugen@us.ibm.com>
8959 Peter Bergner <bergner@vnet.ibm.com>
8961 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
8965 (vrotl<mode>3): Likewise.
8966 (vashl<mode>3): Likewise.
8967 (vlshr<mode>3): Likewise.
8968 (vashr<mode>3): Likewise.
8970 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8971 support for power8 V2DI builtins.
8973 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
8974 power8 V2DI builtins.
8975 (vupkhsw): Likewise.
8976 (vupklsw): Likewise.
8977 (vaddudm): Likewise.
8982 (vpkudum): Likewise.
8983 (vpksdss): Likewise.
8984 (vpkudus): Likewise.
8985 (vpksdus): Likewise.
8990 (vsubudm): Likewise.
8991 (vcmpequd): Likewise.
8992 (vcmpgtsd): Likewise.
8993 (vcmpgtud): Likewise.
8994 (vcmpequd_p): Likewise.
8995 (vcmpgtsd_p): Likewise.
8996 (vcmpgtud_p): Likewise.
8997 (vupkhsw): Likewise.
8998 (vupklsw): Likewise.
8999 (vaddudm): Likewise.
9004 (vpksdss): Likewise.
9005 (vpksdus): Likewise.
9006 (vpkudum): Likewise.
9007 (vpkudus): Likewise.
9012 (vsubudm): Likewise.
9014 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
9015 support for power8 V2DI instructions.
9017 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
9018 power8 V2DI instructions. Combine pack and unpack insns to use an
9019 iterator for each mode. Check whether a particular mode supports
9020 Altivec instructions instead of just checking TARGET_ALTIVEC.
9021 (UNSPEC_VPKUWUM): Likewise.
9022 (UNSPEC_VPKSHSS): Likewise.
9023 (UNSPEC_VPKSWSS): Likewise.
9024 (UNSPEC_VPKUHUS): Likewise.
9025 (UNSPEC_VPKSHUS): Likewise.
9026 (UNSPEC_VPKUWUS): Likewise.
9027 (UNSPEC_VPKSWUS): Likewise.
9028 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
9029 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
9030 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
9031 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
9032 (UNSPEC_VUPKHSB): Likewise.
9033 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
9034 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
9035 (UNSPEC_VUPKHSH): Likewise.
9036 (UNSPEC_VUPKLSB): Likewise.
9037 (UNSPEC_VUPKLSH): Likewise.
9039 (VI_char): Likewise.
9040 (VI_scalar): Likewise.
9041 (VI_unit): Likewise.
9043 (VP_small): Likewise.
9044 (VP_small_lc): Likewise.
9045 (VU_char): Likewise.
9046 (add<mode>3): Likewise.
9047 (altivec_vaddcuw): Likewise.
9048 (altivec_vaddu<VI_char>s): Likewise.
9049 (altivec_vadds<VI_char>s): Likewise.
9050 (sub<mode>3): Likewise.
9051 (altivec_vsubcuw): Likewise.
9052 (altivec_vsubu<VI_char>s): Likewise.
9053 (altivec_vsubs<VI_char>s): Likewise.
9054 (altivec_vavgs<VI_char>): Likewise.
9055 (altivec_vcmpbfp): Likewise.
9056 (altivec_eq<mode>): Likewise.
9057 (altivec_gt<mode>): Likewise.
9058 (altivec_gtu<mode>): Likewise.
9059 (umax<mode>3): Likewise.
9060 (smax<mode>3): Likewise.
9061 (umin<mode>3): Likewise.
9062 (smin<mode>3): Likewise.
9063 (altivec_vpkuhum): Likewise.
9064 (altivec_vpkuwum): Likewise.
9065 (altivec_vpkshss): Likewise.
9066 (altivec_vpkswss): Likewise.
9067 (altivec_vpkuhus): Likewise.
9068 (altivec_vpkshus): Likewise.
9069 (altivec_vpkuwus): Likewise.
9070 (altivec_vpkswus): Likewise.
9071 (altivec_vpks<VI_char>ss): Likewise.
9072 (altivec_vpks<VI_char>us): Likewise.
9073 (altivec_vpku<VI_char>us): Likewise.
9074 (altivec_vpku<VI_char>um): Likewise.
9075 (altivec_vrl<VI_char>): Likewise.
9076 (altivec_vsl<VI_char>): Likewise.
9077 (altivec_vsr<VI_char>): Likewise.
9078 (altivec_vsra<VI_char>): Likewise.
9079 (altivec_vsldoi_<mode>): Likewise.
9080 (altivec_vupkhsb): Likewise.
9081 (altivec_vupkhs<VU_char>): Likewise.
9082 (altivec_vupkls<VU_char>): Likewise.
9083 (altivec_vupkhsh): Likewise.
9084 (altivec_vupklsb): Likewise.
9085 (altivec_vupklsh): Likewise.
9086 (altivec_vcmpequ<VI_char>_p): Likewise.
9087 (altivec_vcmpgts<VI_char>_p): Likewise.
9088 (altivec_vcmpgtu<VI_char>_p): Likewise.
9089 (abs<mode>2): Likewise.
9090 (vec_unpacks_hi_v16qi): Likewise.
9091 (vec_unpacks_hi_v8hi): Likewise.
9092 (vec_unpacks_lo_v16qi): Likewise.
9093 (vec_unpacks_hi_<VP_small_lc>): Likewise.
9094 (vec_unpacks_lo_v8hi): Likewise.
9095 (vec_unpacks_lo_<VP_small_lc>): Likewise.
9096 (vec_pack_trunc_v8h): Likewise.
9097 (vec_pack_trunc_v4si): Likewise.
9098 (vec_pack_trunc_<mode>): Likewise.
9100 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
9102 (vec_vmaxsd): Likewise.
9103 (vec_vmaxud): Likewise.
9104 (vec_vminsd): Likewise.
9105 (vec_vminud): Likewise.
9106 (vec_vpksdss): Likewise.
9107 (vec_vpksdus): Likewise.
9108 (vec_vpkudum): Likewise.
9109 (vec_vpkudus): Likewise.
9110 (vec_vrld): Likewise.
9111 (vec_vsld): Likewise.
9112 (vec_vsrad): Likewise.
9113 (vec_vsrd): Likewise.
9114 (vec_vsubudm): Likewise.
9115 (vec_vupkhsw): Likewise.
9116 (vec_vupklsw): Likewise.
9118 2013-05-29 Jan Hubicka <jh@suse.cz>
9120 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
9121 flags; reorder rest of fields in more consistent way.
9122 (varpool_node): Remove analyzed, finalized and alias.
9123 (cgraph_ndoe): Likewise.
9124 (symtab_alias_ultimate_target): New function.
9125 (cgraph_function_node): Move offline.
9126 (cgraph_reset_node): Declare.
9127 (cgraph_comdat_can_be_unshared_p): Remove.
9128 (varpool_remove_initializer): Declare.
9129 (varpool_first_defined_variable, varpool_next_defined_variable
9130 cgraph_first_defined_function, cgraph_next_defined_function): Update.
9131 (cgraph_function_with_gimple_body_p): Update.
9132 (varpool_all_refs_explicit_p): Update.
9133 (symtab_alias_target): New function.
9134 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
9135 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
9136 (cgraph_function_or_thunk_node): Simplify using
9137 symtab_alias_ultimate_target.
9138 (varpool_variable_node): Likewise.
9139 * cgraph.c (cgraph_create_function_alias): Update.
9140 (cgraph_add_thunk): Update.
9141 (cgraph_remove_node): Update.
9142 (dump_cgraph_node): Do not dump removed flags.
9143 (cgraph_function_body_availability): Update.
9144 (cgraph_propagate_frequency): Update.
9145 (verify_cgraph_node): Check sanity of local flag.
9146 (cgraph_function_node): Move here from cgraph.h; revamp for
9147 cgraph_function_or_thunk_node.
9148 * lto-symtab.c (lto_varpool_replace_node): Update.
9149 (lto_symtab_resolve_can_prevail_p): Update.
9150 (lto_symtab_merge_cgraph_nodes): Update.
9151 * ipa-cp.c (determine_versionability, initialize_node_lattices,
9152 propagate_constants_accross_call, devirtualization_time_bonus,
9153 ipcp_propagate_stage): Update.
9154 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
9155 * ipa-inline-transform.c (clone_inlined_nodes,
9156 preserve_function_body_p): Update.
9157 * ipa-reference.c (propagate): Update.
9158 (write_node_summary_p): Update.
9159 * toplev.c (wrapup_global_declaration_2): Update.
9160 * cgraphunit.c (cgraph_analyze_function): Rename to ...
9161 (analyze_function) ... this one.
9162 (cgraph_process_new_functions): Update.
9163 (cgraph_reset_node): Export.
9164 (cgraph_finalize_function): Update.
9165 (cgraph_add_new_function): Update.
9166 (process_function_and_variable_attributes): Update.
9167 (varpool_finalize_decl): Update.
9168 (symbol_finalized): Remove.
9169 (symbol_finalized_and_needed): Rename to ...
9170 (symbol_defined_and_needed): ... update.
9171 (cgraph_analyze_functions): Update.
9172 (handle_alias_pairs): Update.
9173 (mark_functions_to_output): Update.
9174 (assemble_thunk): Update.
9175 (output_in_order): Update.
9176 (output_weakrefs): Update.
9177 (finalize_compilation_unit): Update.
9178 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
9179 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
9180 input_node, input_varpool_node): Update.
9181 * dbxout.c (dbxout_expand_expr): Update.
9182 * cgraphclones.c (cgraph_clone_node): Update.
9183 (cgraph_copy_node_for_versioning): Update.
9184 (cgraph_materialize_clone): Update.
9185 (cgraph_materialize_all_clones): Update.
9186 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
9187 propagate_pure_const, propagate_nothrow): Update.
9188 * lto-streamer-out.c (lto_output, write_symbol): Update.
9189 * ipa-utils.c (ipa_reverse_postorder): Update.
9190 * ipa-inline.c (can_inline_edge_p): Update.
9191 (update_caller_keys, ipa_inline): Update.
9192 * dwarf2out.c (reference_to_unused,
9193 premark_types_used_by_global_vars_helper): Update.
9194 * tree-eh.c (tree_could_trap_p): Update.
9195 * ipa-split.c (consider_split, execute_split_functions): Update.
9196 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
9197 has_addr_references_p): Update; move ahead in file for better
9199 (process_references): Simplify.
9200 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
9202 (cgraph_comdat_can_be_unshared_p): Make static.
9203 (cgraph_externally_visible_p): Update.
9204 (varpool_externally_visible_p): Update.
9205 (function_and_variable_visibility): Update.
9206 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
9207 ipa_tm_mark_force_output_node): Update.
9208 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
9209 estimate_edge_devirt_benefit, inline_generate_summary,
9210 inline_write_summary): Update.
9211 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
9212 * ipa-prop.c (ipa_compute_jump_functions): Update.
9213 (ipa_print_node_params, ipa_prop_read_section,
9214 ipa_update_after_lto_read, read_replacements_section): Update.
9215 * varasm.c (mark_decl_referenced): Update.
9216 (assemble_alias, dump_tm_clone_pairs): Update.
9217 * tree-inline.c (copy_bb): Update.
9218 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
9220 * symtab.c (dump_symtab_base): Print new flags.
9221 (verify_symtab_base): Verify new flags.
9222 (symtab_alias_ultimate_target): New function.
9223 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
9224 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
9226 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
9228 * i386.c (ix86_get_function_versions_dispatcher,
9229 ix86_generate_version_dispatcher_body): Update.
9230 (fold_builtin_cpu): Use varpool_add_new_variable.
9231 * varpool.c (varpool_remove_initializer): Break out from ...
9232 (varpool_remove_node): ... this one.
9233 (dump_varpool_node, varpool_node_for_asm,
9234 cgraph_variable_initializer_availability, varpool_analyze_node,
9235 varpool_assemble_decl, varpool_remove_unreferenced_decls,
9236 varpool_finalize_named_section_flags, varpool_create_variable_alias):
9239 2013-05-29 Jan Hubicka <jh@suse.cz>
9241 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
9243 2013-05-29 Easwaran Raman <eraman@google.com>
9245 PR tree-optimization/57442
9246 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
9247 when control exits the main loop.
9249 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
9251 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
9253 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
9254 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
9255 * rx/t-rx: Add rx100 under multi library matches option for nofpu
9258 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9260 PR tree-optimization/57441
9261 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
9262 Don't limit size of incr_vec to number of candidates.
9264 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
9266 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
9267 and mips16 directories.
9268 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
9269 (MULTILIB_DIRNAMES): Ditto.
9270 (MULTILIB_EXCEPTIONS): Add new exceptions.
9271 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
9272 (MULTILIB_DIRNAMES): Ditto.
9273 (MULTILIB_EXCEPTIONS): Add new exceptions.
9275 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9276 Marcus Shawcroft <marcus.shawcroft@arm.com>
9278 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
9279 SYMBOL_TINY_ABSOLUTE.
9280 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
9281 SYMBOL_TINY_ABSOLUTE.
9282 (aarch64_expand_mov_immediate): Likewise.
9283 (aarch64_classify_symbol): Likewise.
9284 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
9285 Permit SYMBOL_TINY_ABSOLUTE.
9286 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
9288 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9289 Marcus Shawcroft <marcus.shawcroft@arm.com>
9291 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
9292 Refactor if/switch. Replace gcc_assert with if.
9294 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9296 * config/i386/i386.c (initial_ix86_tune_features): Enable
9297 FP Reassociation for AMD bdver1 and bdver2.
9299 2013-05-29 Martin Jambor <mjambor@suse.cz>
9301 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
9302 and IMAGPART_EXPR do not occur within other handled_components.
9304 2013-05-29 Richard Biener <rguenther@suse.de>
9306 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
9307 access on whether the use is in the BB we currently try to
9309 (vect_bb_vectorization_profitable_p): Pass the BB we currently
9310 vectorize to vect_bb_slp_scalar_cost.
9312 2013-05-29 Richard Biener <rguenther@suse.de>
9314 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
9315 computing scalar cost offsetted by stmts that are kept live
9317 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
9318 for computation of scalar cost.
9320 2013-05-28 Steve Ellcey <sellcey@mips.com>
9322 * config/mips/mips-cpus.def (mips32r2): Change processor type.
9324 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9326 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
9327 array notation built-in reduction functions.
9328 * doc/passes.texi (Passes): Added documentation about changes done
9330 * doc/invoke.texi (C Dialect Options): Added documentation about
9331 the -fcilkplus flag.
9332 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
9333 (BUILTINS_DEF): Depend on cilkplus.def.
9334 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
9335 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
9336 * cilkplus.def: New file.
9338 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
9340 PR rtl-optimization/57439
9341 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
9343 2013-05-28 Easwaran Raman <eraman@google.com>
9345 PR tree-optimization/57337
9346 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
9347 (find_insert_point): Correctly identify the insertion point
9348 when two statements with the same UID is compared.
9350 2013-05-28 Richard Biener <rguenther@suse.de>
9352 PR tree-optimization/56787
9353 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
9354 from the list of data references.
9355 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
9357 (vect_analyze_loop_operations): Likewise.
9358 (vect_transform_loop): Remove clobbers.
9360 2013-05-28 Martin Jambor <mjambor@suse.cz>
9362 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
9363 and REALPART_EXPRs have scalar type.
9365 2013-05-28 Richard Biener <rguenther@suse.de>
9367 PR tree-optimization/57411
9368 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
9370 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
9371 virtual operand propagation.
9373 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9375 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
9376 destination register for bmasksi_vis.
9377 (vector_init_bshuffle): Likewise.
9378 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
9380 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9382 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
9383 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
9384 mode if the instruction isn't available in the original mode.
9385 * config/sparc/sparc.opt (mfix-ut699): New option.
9386 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
9387 (divdf3): Turn into expander.
9388 (divdf3_nofix): New insn.
9389 (divdf3_fix): Likewise.
9390 (divsf3): Disable if -mfix-ut699.
9391 (sqrtdf2): Turn into expander.
9392 (sqrtdf2_nofix): New insn.
9393 (sqrtdf2_fix): Likewise.
9394 (sqrtsf2): Disable if -mfix-ut699.
9396 2013-05-27 Richard Biener <rguenther@suse.de>
9399 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
9400 block for the new loop.
9402 2013-05-27 Richard Biener <rguenther@suse.de>
9404 PR tree-optimization/57343
9405 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
9406 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
9407 (number_of_iterations_cond): Do not build the folded tree.
9409 2013-05-27 Richard Biener <rguenther@suse.de>
9413 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9414 OEP_CONSTANT_ADDRESS_OF retained.
9416 PR tree-optimization/57417
9417 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
9419 (set_ssa_val_to): Compare addresses using
9420 get_addr_base_and_unit_offset.
9422 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
9424 PR rtl-optimization/56833
9425 * postreload.c (move2add_record_mode): New function.
9426 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
9427 (move2add_use_add2_insn): Use move2add_record_sym_value.
9428 (move2add_use_add3_insn): Likewise.
9429 (reload_cse_move2add): Use move2add_valid_value_p and
9430 move2add_record_mode. Invalidate call-clobbered and REG_INC
9431 affected regs by setting reg_mode to VOIDmode.
9432 (move2add_note_store): Don't pretend the inside of a SUBREG is
9433 the actual destination. Invalidate single/leading registers by
9434 setting reg_mode to VOIDmode.
9435 Use move2add_record_sym_value, move2add_valid_value_p and
9436 move2add_record_mode.
9438 2013-05-27 Richard Biener <rguenther@suse.de>
9440 PR tree-optimization/57396
9441 * tree-affine.c (double_int_constant_multiple_p): Properly
9442 return false for val == 0 and div != 0.
9444 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9446 * config/mips/mips.h: Use #elif in preprocessor conditions.
9448 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9451 * config/mips/constraints.md (kl): New constraint.
9452 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
9453 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
9454 constraint for operand 0. Split after CSE for MIPS16. Emit a move
9456 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
9458 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9461 * config/mips/mips.c (mips_can_inline_p): New function.
9462 (TARGET_CAN_INLINE_P): Define.
9464 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
9466 * sched-int.h (ds_t, dw_t): Make unsigned int.
9467 Fix documentation that describes how all the ds_t bits are used.
9468 Reserve the last bit for delayed-branch scheduling.
9469 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
9470 (BITS_PER_DEP_WEAK): Fix definition and documentation.
9471 (gen_dep_weak_1): Remove prototype.
9472 * sched-deps.c (get_dep_weak_1): Make static.
9473 * target.def (speculate_insn, needs_block_p, gen_spec_check,
9474 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
9475 * doc/tm.texi: Regenerate.
9476 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
9478 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
9481 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
9483 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
9484 Sandra Loosemore <sandra@codesourcery.com>
9486 * config.gcc (powerpc-*): Allow native for with-cpu.
9488 2013-05-24 Jeff Law <law@redhat.com>
9490 PR tree-optimization/57124
9491 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
9492 conversion feeding a condition if the range has an overflow
9493 if -fstrict-overflow. Add warnings for when we do make the
9496 2013-05-24 Dehao Chen <dehao@google.com>
9498 * tree-cfg.c (locus_discrim_map): Fix the typo.
9499 (locus_discrim_hasher): Likewise.
9500 (locus_discrim_hasher::hash): Likewise.
9501 (locus_discrim_hasher::equal): Likewise.
9503 2013-05-24 Martin Jambor <mjambor@suse.cz>
9505 PR tree-optimization/57294
9506 * cgraph.h (ipa_record_stmt_references): Declare.
9507 * cgraphbuild.c (ipa_record_stmt_references): New function.
9508 (build_cgraph_edges): Use ipa_record_stmt_references.
9509 (rebuild_cgraph_edges): Likewise.
9510 (cgraph_rebuild_references): Likewise.
9511 * ipa-prop.c (ipa_modify_call_arguments): Discard references
9512 associated with the old statement and build references from the
9513 newly built statements.
9514 * ipa-ref.c (ipa_remove_stmt_references): New function.
9515 * ipa-ref.h (ipa_remove_stmt_references): Declare.
9517 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
9519 * lra-constraints.c (emit_spill_move): Use smaller mode for
9521 (check_and_process_move): Consider mem-reg moves for secondary
9523 (curr_insn_transform): Don't lose insns emitted before for
9524 secondary memory moves.
9525 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
9526 reg set up in the current insn.
9528 2013-05-24 Dehao Chen <dehao@google.com>
9530 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
9532 (locus_descrim_hasher::equal): Likewise.
9533 (build_gimple_cfg): New discriminator assignment algorithm.
9534 (make_edges): Likewise.
9535 (next_discriminator_for_locus): Likewise.
9536 (same_line_p): Likewise.
9537 (assign_discriminators): Likewise.
9538 (make_cond_expr_edges): Likewise.
9539 (make_gimple_switch_edges): Likewise.
9540 (make_goto_expr_edges): Likewise.
9541 (make_gimple_asm_edges): Likewise.
9543 2013-05-24 Ian Bolton <ian.bolton@arm.com>
9545 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
9546 X format specifier to only display bottom 16 bits.
9547 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
9548 immediate to match for operand 2, since it will be masked.
9550 2013-05-24 Richard Biener <rguenther@suse.de>
9552 PR tree-optimization/57287
9553 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
9554 all SSA names that occur in abnormal PHIs.
9556 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
9559 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
9560 that index is not negative.
9562 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
9564 PR rtl-optimization/55177
9565 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
9566 (simplify_byte_swapping_operation): New.
9567 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
9568 (simplify_relational_operation_1): Deal with BSWAP.
9570 2013-05-23 Richard Henderson <rth@redhat.com>
9573 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
9574 (ix86_reorg): Call it.
9576 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
9579 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
9580 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
9581 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
9583 2013-05-23 Christian Bruel <christian.bruel@st.com>
9586 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
9588 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9589 Marcus Shawcroft <marcus.shawcroft@arm.com>
9591 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
9592 * config/aarch64/constraints.md (Usa): Remove.
9593 * doc/md.texi (AArch64 Usa): Remove.
9595 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9596 Marcus Shawcroft <marcus.shawcroft@arm.com>
9598 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
9599 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
9600 * config/aarch64/predicates.md (aarch64_const_address): Remove.
9601 (aarch64_mov_operand): Use aarch64_mov_operand_p.
9603 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
9605 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
9606 instruction (AdvSIMD).
9607 * config/aarch64/aarch64-builtins.c
9608 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
9609 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
9611 2013-05-23 Martin Jambor <mjambor@suse.cz>
9614 * tree.h (contains_bitfld_component_ref_p): Declare.
9615 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
9616 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
9618 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
9619 not access a bit-field. Assert all final offsets are byte-aligned.
9621 2013-05-23 Richard Biener <rguenther@suse.de>
9623 PR tree-optimization/57380
9624 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
9625 least one invariant or re-used load.
9626 * passes.c (init_optimization_passes): Move pass_phiprop before
9629 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
9631 * config/aarch64/aarch64-simd.md
9632 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
9634 2013-05-23 Richard Biener <rguenther@suse.de>
9637 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9638 OEP_CONSTANT_ADDRESS_OF retained.
9640 2013-05-23 Jakub Jelinek <jakub@redhat.com>
9643 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
9644 don't lower unit. Handle unit not being always BITS_PER_WORD.
9646 2013-05-23 Richard Biener <rguenther@suse.de>
9648 PR rtl-optimization/57341
9649 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
9650 instead of true_dependence.
9652 2013-05-22 David Malcolm <dmalcolm@redhat.com>
9654 * bb-reorder.c (branch_threshold): Make const.
9655 (exec_threshold): Ditto.
9657 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9658 Pat Haugen <pthaugen@us.ibm.com>
9659 Peter Bergner <bergner@vnet.ibm.com>
9661 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9662 documentation for the power8 crypto builtins.
9664 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9666 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9667 macros for defining power8 builtin functions.
9668 (BU_P8V_AV_2): Likewise.
9669 (BU_P8V_AV_P): Likewise.
9670 (BU_P8V_VSX_1): Likewise.
9671 (BU_P8V_OVERLOAD_1): Likewise.
9672 (BU_P8V_OVERLOAD_2): Likewise.
9673 (BU_CRYPTO_1): Likewise.
9674 (BU_CRYPTO_2): Likewise.
9675 (BU_CRYPTO_3): Likewise.
9676 (BU_CRYPTO_OVERLOAD_1): Likewise.
9677 (BU_CRYPTO_OVERLOAD_2): Likewise.
9678 (XSCVSPDP): Fix typo, point to the correct instruction.
9679 (VCIPHER): Add power8 crypto builtins.
9680 (VCIPHERLAST): Likewise.
9681 (VNCIPHER): Likewise.
9682 (VNCIPHERLAST): Likewise.
9683 (VPMSUMB): Likewise.
9684 (VPMSUMH): Likewise.
9685 (VPMSUMW): Likewise.
9686 (VPERMXOR_V2DI): Likewise.
9687 (VPERMXOR_V4SI: Likewise.
9688 (VPERMXOR_V8HI: Likewise.
9689 (VPERMXOR_V16QI: Likewise.
9690 (VSHASIGMAW): Likewise.
9691 (VSHASIGMAD): Likewise.
9693 (VPERMXOR): Likewise.
9694 (VSHASIGMA): Likewise.
9696 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9697 __CRYPTO__ if the crypto instructions are available.
9698 (altivec_overloaded_builtins): Add support for overloaded power8
9701 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9702 support for power8 crypto builtins.
9703 (builtin_function_type): Likewise.
9704 (altivec_init_builtins): Add support for builtins that take vector
9705 long long (V2DI) arguments.
9707 * config/rs6000/crypto.md: New file, define power8 crypto
9710 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9711 Pat Haugen <pthaugen@us.ibm.com>
9712 Peter Bergner <bergner@vnet.ibm.com>
9714 * doc/invoke.texi (Option Summary): Add power8 options.
9715 (RS/6000 and PowerPC Options): Likewise.
9717 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9718 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9719 wm, wn, wr documentation.
9721 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9722 registers if direct move instructions are enabled.
9723 (wn): New constraint for no registers.
9724 (wq): New constraint for quad word even GPR registers.
9725 (wr): New constraint if 64-bit instructions are enabled.
9726 (wv): New constraint if power8 vector instructions are enabled.
9727 (wQ): New constraint for quad word memory locations.
9729 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9730 constraint for 0..15 for crypto instructions.
9731 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9732 as GPR and floating point registers.
9733 (int_reg_operand): New predicate to match only GPR registers.
9734 (base_reg_operand): New predicate to match base registers.
9735 (quad_int_reg_operand): New predicate to match even GPR registers
9736 for quad memory operations.
9737 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9738 operations in both GPR and VSX registers.
9739 (quad_memory_operand): New predicate for quad memory operations.
9740 (reg_or_indexed_operand): New predicate for direct move support.
9742 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9743 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9744 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9745 (POWERPC_MASKS): Add power8 options.
9746 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9749 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9750 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9752 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9753 (-mpower8-fusion): New power8 options.
9754 (-mpower8-fusion-sign): Likewise.
9755 (-mpower8-vector): Likewise.
9756 (-mcrypto): Likewise.
9757 (-mdirect-move): Likewise.
9758 (-mquad-memory): Likewise.
9760 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9762 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9764 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9765 (rs6000_debug_vector_unit): Add p8_vector.
9766 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9767 definitions. Also print fusion state.
9768 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9769 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9770 (rs6000_option_override_internal): Add support for power8.
9771 (rs6000_common_init_builtins): Add debugging for skipped builtins
9773 (rs6000_adjust_cost): Add power8 support.
9774 (rs6000_issue_rate): Likewise.
9775 (insn_must_be_first_in_group): Likewise.
9776 (insn_must_be_last_in_group): Likewise.
9777 (force_new_group): Likewise.
9778 (rs6000_register_move_cost): Likewise.
9779 (rs6000_opt_masks): Likewise.
9781 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9782 power8 capable assembler, default to power7 options.
9783 (TARGET_DIRECT_MOVE): Likewise.
9784 (TARGET_CRYPTO): Likewise.
9785 (TARGET_P8_VECTOR): Likewise.
9786 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9787 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9788 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9789 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9790 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9791 (TARGET_XSCVDPSPN): Likewise.
9792 (TARGET_XSCVSPDPN): Likewsie.
9793 (TARGET_SYNC_HI_QI): Likewise.
9794 (TARGET_SYNC_TI): Likewise.
9795 (MASK_CRYPTO): Likewise.
9796 (MASK_DIRECT_MOVE): Likewise.
9797 (MASK_P8_FUSION): Likewise.
9798 (MASK_P8_VECTOR): Likewise.
9799 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9800 temporary used by some of the direct move instructions to get two FP
9801 temporary registers does not force creation of a stack frame.
9802 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9803 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9804 that any VSX registers are tieable, even if they are also an
9805 Altivec vector mode.
9806 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9807 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9808 (RS6000_BTM_CRYPTO): Likewise.
9809 (RS6000_BTM_COMMON): Likewise.
9811 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9812 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9813 (enum rs6000_vector): Add power8 vector support.
9815 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9819 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9820 (any_sibcall_could_use_r3): this and handle indirect calls.
9821 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9823 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9825 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9827 2013-05-22 Richard Biener <rguenther@suse.de>
9830 * profile.c (branch_prob): Do not split blocks that are
9831 abnormally receiving from ECF_RETURNS_TWICE functions.
9833 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9835 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9836 address mode. Move pointer mode initialization to the same place.
9838 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9840 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
9841 while it has any effect.
9843 2013-05-21 Easwaran Raman <eraman@google.com>
9845 PR tree-optimization/57322
9846 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
9847 UID of the statement added to the BB to be 1.
9849 2013-05-21 Jakub Jelinek <jakub@redhat.com>
9851 PR tree-optimization/57331
9852 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
9853 of conversion from pointer type to integral type with integer.
9855 2013-05-21 Martin Jambor <mjambor@suse.cz>
9858 * ipa-prop.c (ipa_read_node_info): Process param_used and
9859 controlled_uses in the same order as when writing.
9861 2013-05-21 Magnus Granberg <baldrick@free.fr>
9864 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
9866 2013-05-21 Richard Biener <rguenther@suse.de>
9868 PR tree-optimization/57318
9869 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
9870 estimate stmts with side-effects as likely eliminated.
9872 2013-05-21 Richard Biener <rguenther@suse.de>
9874 PR tree-optimization/57330
9875 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
9876 preserve the call stmts fntype.
9878 2013-05-21 Richard Biener <rguenther@suse.de>
9880 PR tree-optimization/57303
9881 * tree-ssa-sink.c (statement_sink_location): Improve killing
9882 stmt detection and properly handle self-assignments.
9884 2013-05-21 Christian Bruel <christian.bruel@st.com>
9886 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
9887 spanning registers. LEAF_REG_REMAP is supported only for contiguous
9888 registers. Set register size out of the PARALLEL loop.
9890 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
9893 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
9894 (*fmasf4, *fmasf4_media): New insns.
9896 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9898 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
9899 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
9900 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
9901 (mips_idiv_insns): Update the comments to say that the returned
9902 instruction counts are in units of BASE_INSN_LENGTH.
9903 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
9904 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
9905 using 2 rather than 4 as the length of indirect MIPS16 and
9906 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
9907 length of a NOP. Don't divide MIPS16 lengths by 2.
9908 (mips16_split_long_branches): Assume a branch is long if the
9909 length is greater than 4 rather than 8.
9910 * config/mips/mips.md (length): Give MIPS16 lengths directly,
9911 rather than multiplying them by 2. Multiply instruction counts
9912 by BASE_INSN_LENGTH rather than 4.
9913 (*jump_mips16, tls_get_tp_mips16_<mode>)
9914 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
9916 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9918 * config/mips/mips.md (extended_mips16): Remove branch case.
9919 (length): Remove duplicated extended_mips16 test.
9921 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9923 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
9925 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
9927 * recog.h (Recog_data): Rename to...
9928 (recog_data_d): ...this.
9929 (recog_data): Update accordingly.
9930 * recog.c (recog_data): Likewise.
9931 * reload.c (save_recog_data): Likewise.
9932 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
9933 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
9935 2013-05-17 Julian Brown <julian@codesourcery.com>
9937 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
9938 found in a REG_EQUAL note, invalidate it.
9940 2013-05-17 Easwaran Raman <eraman@google.com>
9942 * tree-ssa-reassoc.c (find_insert_point): New function.
9943 (insert_stmt_after): Likewise.
9944 (get_def_stmt): Likewise.
9945 (ensure_ops_are_available): Likewise.
9946 (not_dominated_by): Likewise.
9947 (rewrite_expr_tree): Do not move statements beyond what is
9948 necessary. Remove call to swap_ops_for_binary_stmt...
9949 (reassociate_bb): ... and move it here.
9950 (build_and_add_sum): Assign UIDs for new statements.
9951 (linearize_expr): Likewise.
9952 (do_reassoc): Renumber gimple statement UIDs.
9954 2013-05-17 Jan Hubicka <jh@suse.cz>
9956 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
9958 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
9959 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
9961 (output_weakrefs): Update.
9963 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
9964 Martin Jambor <mjambor@suse.cz>
9967 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
9968 value that corresponds to the given aggval is found in values vector.
9970 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
9972 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
9973 sse, sse2, sse3, ssse3 and sse4a flags to options.
9975 2013-05-17 David Malcolm <dmalcolm@redhat.com>
9977 * gengtype-state.c: (s_expr_writer): New class, to handle
9978 prettifying of output layout of s-expressions.
9979 (state_writer): New class, to write out gtype.state.
9980 (state_written_type_count): Move this variable into member data of
9982 (s_expr_writer::s_expr_writer): New code: constructor for new class
9983 (state_writer::state_writer(): ditto
9984 (s_expr_writer::write_new_line): New function
9985 (s_expr_writer::write_any_indent): ditto
9986 (s_expr_writer::begin_s_expr): ditto
9987 (s_expr_writer::end_s_expr): ditto
9988 (write_state_fileloc): convert to method of state_writer...
9989 (state_writer:: write_state_fileloc): ...and use methods of
9990 s_expr_writer to write indentation into the gtype.state output file
9991 to visually represent the hierarchical structure of the list
9993 (write_state_fields): ditto, renaming to...
9994 (state_writer::write_state_fields)
9995 (write_state_a_string): ditto, renaming to...
9996 (state_writer::write_state_a_string)
9997 (write_state_string_option): ditto, renaming to...
9998 (state_writer::write_state_string_option)
9999 (write_state_type_option): ditto, renaming to...
10000 (state_writer::write_state_type_option)
10001 (write_state_nested_option): ditto, renaming to...
10002 (state_writer::write_state_nested_option)
10003 (write_state_option): ditto, renaming to...
10004 (state_writer::write_state_option)
10005 (write_state_options): ditto, renaming to...
10006 (state_writer::write_state_options)
10007 (write_state_lang_bitmap): ditto, renaming to...
10008 (state_writer::write_state_lang_bitmap)
10009 (write_state_version): ditto, renaming to...
10010 (state_writer::write_state_version)
10011 (write_state_scalar_type): ditto, renaming to...
10012 (state_writer::write_state_scalar_type)
10013 (write_state_string_type): ditto, renaming to...
10014 (state_writer::write_state_string_type)
10015 (write_state_undefined_type): ditto, renaming to...
10016 (state_writer::write_state_undefined_type)
10017 (write_state_struct_union_type): ditto, renaming to...
10018 (state_writer::write_state_struct_union_type)
10019 (write_state_struct_type): ditto, renaming to...
10020 (state_writer::write_state_struct_type)
10021 (write_state_user_struct_type): ditto, renaming to...
10022 (state_writer::write_state_user_struct_type)
10023 (write_state_lang_struct_type): ditto, renaming to...
10024 (state_writer::write_state_lang_struct_type)
10025 (write_state_param_struct_type): ditto, renaming to...
10026 (state_writer::write_state_param_struct_type)
10027 (write_state_pointer_type): ditto, renaming to...
10028 (state_writer::write_state_pointer_type)
10029 (write_state_array_type): ditto, renaming to...
10030 (state_writer::write_state_array_type)
10031 (write_state_gc_used): ditto, renaming to...
10032 (state_writer::write_state_gc_used)
10033 (write_state_common_type_content): ditto, renaming to...
10034 (state_writer::write_state_common_type_content)
10035 (write_state_type): ditto, renaming to...
10036 (state_writer::write_state_type)
10037 (write_state_pair_list): ditto, renaming to...
10038 (state_writer::write_state_pair_list)
10039 (write_state_pair): ditto, renaming to...
10040 (state_writer::write_state_pair)
10041 (write_state_typedefs): ditto, renaming to...
10042 (state_writer::write_state_typedefs)
10043 (write_state_structures): ditto, renaming to...
10044 (state_writer::write_state_structures)
10045 (write_state_param_structs): ditto, renaming to...
10046 (state_writer::write_state_param_structs)
10047 (write_state_variables): ditto, renaming to...
10048 (state_writer::write_state_variables)
10049 (write_state_srcdir): ditto, renaming to...
10050 (state_writer::write_state_srcdir)
10051 (write_state_files_list): ditto, renaming to...
10052 (state_writer::write_state_files_list)
10053 (write_state_languages): ditto, renaming to...
10054 (state_writer::write_state_languages)
10055 (write_state): create a state_writer instance and use it when
10056 writing out the state file
10058 2013-05-17 Mike Stump <mikestump@comcast.net>
10060 PR rtl-optimization/57304
10061 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
10062 accessing DF_REF_REAL_LOC.
10064 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10066 PR rtl-optimization/57281
10067 PR rtl-optimization/57300
10068 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
10069 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
10070 what the other splitter did if the registers are dead.
10072 2013-05-17 Richard Biener <rguenther@suse.de>
10074 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
10077 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10079 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
10082 2013-05-17 Marek Polacek <polacek@redhat.com>
10084 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
10085 length when doing non-zero store of storing '\0' to '\0'.
10087 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10089 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
10090 vect_external_def oprnd1 with loop_vinfo, try to emit
10091 optional cast, negation and and stmts on the loop preheader
10092 edge instead of into the pattern def seq.
10094 PR tree-optimization/57051
10095 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
10096 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
10098 2013-05-16 Nick Clifton <nickc@redhat.com>
10100 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
10101 (rl78_is_naked_func): New function.
10102 (rl78_expand_prologue): Skip prologue generation for naked functions.
10103 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
10104 * doc/extend.texi (naked): Add RL78 to the list of processors
10105 that supports this attribute.
10107 2013-05-16 Jeff Law <law@redhat.com>
10109 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
10111 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10113 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
10114 cache parameters using detect_caches_amd also for CYRIX,
10115 NSC and TM2 signatures.
10117 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10118 Dzianis Kahanovich <mahatma@eu.by>
10122 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10123 VIA/Centaur processors and determine their cache parameters
10124 using detect_caches_amd.
10126 2013-05-16 Teresa Johnson <tejohnson@google.com>
10128 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
10129 (rtl_verify_edges): New function.
10130 (rtl_verify_bb_insns): Ditto.
10131 (rtl_verify_bb_pointers): Ditto.
10132 (rtl_verify_bb_insn_chain): Ditto.
10133 (rtl_verify_fallthru): Ditto.
10134 (rtl_verify_bb_layout): Ditto.
10135 (rtl_verify_flow_info_1): Outline checks into new functions.
10136 (rtl_verify_flow_info): Ditto.
10138 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
10140 * cfghooks.c (copy_bbs): Add update_dominance argument.
10141 * cfghooks.h (copy_bbs): Update prototype.
10142 * tree-cfg.c (gimple_duplicate_sese_region):
10143 Add update_dominance argument.
10144 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
10145 * tree-ssa-loop-ch.c (copy_loop_headers): Update
10146 gimple_duplicate_sese_region call.
10147 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
10148 Update copy_bbs call.
10149 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
10150 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
10152 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10154 * tree-vectorizer.h (NUM_PATTERNS): Increment.
10155 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
10156 vect_recog_rotate_pattern.
10157 (vect_recog_rotate_pattern): New function.
10159 2013-05-16 Jason Merrill <jason@redhat.com>
10161 * Makefile.in (LLINKER): New variable.
10162 (mostlyclean): Remove link mutex.
10163 * configure.ac: Handle --enable-link-mutex.
10164 * lock-and-run.sh: New script.
10166 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10169 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
10172 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10174 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
10176 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
10178 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
10179 * config/arm/arm.c (next_consecutive_mem): New function.
10180 (gen_movmem_ldrd_strd): Likewise.
10181 * config/arm/arm.md (movmemqi): Update condition and code.
10182 (unaligned_loaddi, unaligned_storedi): New patterns.
10184 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10186 * config.gcc: Obsolete *-*-solaris2.9*.
10187 * doc/install.texi (Specific, *-*-solaris2*): Document it.
10189 2013-05-16 Richard Biener <rguenther@suse.de>
10191 * passes.c (init_optimization_passes): Move pass_parallelize_loops
10192 earlier, after GRAPHITE transforms and IV canonicalization.
10194 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10196 * omp-low.c (extract_omp_for_data): For collapsed loops,
10197 if at least one of the loops is known at compile time to
10198 iterate zero times, set count to 0.
10199 (expand_omp_regimplify_p): New function.
10200 (expand_omp_for_generic): For collapsed loops, if at least
10201 one of the loops isn't known to iterate at least once,
10202 add runtime check with setting count to 0.
10203 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
10204 For unsigned types if it isn't known at compile time that
10205 the loop will iterate at least once, add runtime check to bypass
10206 the whole loop if initial condition isn't true.
10208 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
10210 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
10212 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
10214 PR middle-end/57286
10215 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
10216 transformations to avoid an infinite loop.
10218 2013-05-16 Marek Polacek <polacek@redhat.com>
10220 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
10222 2013-05-15 Leif Ekblad <leif@rdos.net>
10224 * config/i386/i386.c (ix86_decompose_address): Use
10225 DEFAULT_TLS_SEG_REG to access TLS segment register.
10226 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
10227 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
10228 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
10230 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
10233 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
10234 sibling calls to functions that would normally be lazily bound,
10235 unless $gp is call-clobbered.
10237 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
10239 * config/i386/i386.c (ix86_option_override_internal): Update
10240 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
10241 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
10242 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
10243 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
10245 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
10247 2013-05-15 Andreas Schwab <schwab@suse.de>
10249 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
10250 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
10253 2013-05-15 Teresa Johnson <tejohnson@google.com>
10255 * loop-unroll.c (report_unroll_peel): Check decision before
10256 emitting unroll/peel message.
10258 2013-05-15 Teresa Johnson <tejohnson@google.com>
10260 * function.h (has_bb_partition): New rtl_data flag.
10261 (bb_reorder_complete): Ditto.
10262 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
10263 instead of flag_reorder_blocks_and_partition.
10264 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
10265 with some enhancements.
10266 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
10267 * bb-reorder.c (connect_traces): Check for has_bb_partition
10268 instead of flag_reorder_blocks_and_partition.
10269 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
10270 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
10271 verify_hot_cold_block_grouping.
10272 (partition_hot_cold_basic_blocks): Set has_bb_partition.
10274 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10277 * config/arm/predicates.md (call_insn_operand): New predicate.
10278 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
10279 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
10280 if insn is not a tail call.
10281 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
10283 * config/arm/arm.h (enum reg_class): New caller save register class.
10284 (REG_CLASS_NAMES): Likewise.
10285 (REG_CLASS_CONTENTS): Likewise.
10286 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
10289 2013-05-15 Richard Biener <rguenther@suse.de>
10291 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
10292 of MSG_OPTIMIZED_LOCATIONS.
10293 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
10294 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
10296 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
10297 of MSG_OPTIMIZED_LOCATIONS.
10298 (execute_vect_slp): Likewise.
10299 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
10300 (vect_create_cond_for_alias_checks): Likewise.
10301 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
10302 (vect_recog_widen_mult_pattern): Likewise.
10303 (vect_recog_widen_sum_pattern): Likewise.
10304 (vect_recog_over_widening_pattern): Likewise.
10305 (vect_recog_widen_shift_pattern): Likewise.
10306 (vect_recog_vector_vector_shift_pattern): Likewise.
10307 (vect_recog_divmod_pattern): Likewise.
10308 (vect_recog_mixed_size_cond_pattern): Likewise.
10309 (vect_recog_bool_pattern): Likewise.
10310 (vect_pattern_recog_1): Likewise.
10312 2013-05-15 Martin Jambor <mjambor@suse.cz>
10314 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
10315 non-functions to builtin_unreachable.
10316 * ipa-inline-transform.c (inline_call): Do not assert estimates were
10317 correct when new direct edges were discovered.
10319 2013-05-15 Martin Jambor <mjambor@suse.cz>
10321 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
10322 header, print symbol order instead of node uid, print more information
10323 about indirect edge targets.
10324 (ipa_make_edge_direct_to_target): Print symbol order instead of node
10326 (ipa_make_edge_direct_to_target): Likewise.
10327 (remove_described_reference): Likewise.
10328 (propagate_controlled_uses): Likewise.
10329 (ipa_print_node_params): Also print symbol order.
10330 (ipcp_transform_function): Print symbol order instead of node uids.
10331 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
10332 (cgraph_get_create_real_symbol_node): Likewise.
10333 * ipa-cp.c (print_lattice): Likewise.
10334 (print_all_lattices): Likewise.
10335 (determine_versionability): Likewise.
10336 (initialize_node_lattices): Likewise.
10337 (estimate_local_effects): Likewise.
10338 (update_profiling_info): Likewise.
10339 (create_specialized_node): Likewise.
10340 (perhaps_add_new_callers): Likewise.
10341 (decide_about_value): Likewise.
10342 (decide_whether_version_node): Likewise.
10343 (identify_dead_nodes): Likewise.
10344 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
10345 (dump_inline_summary): Likewise.
10346 (estimate_node_size_and_time): Likewise.
10347 (inline_analyze_function): Likewise.
10348 * ipa-inline.c (report_inline_failed_reason): Likewise.
10349 (want_early_inline_function_p): Likewise.
10350 (edge_badness): Likewise.
10351 (update_edge_key): Likewise.
10352 (inline_small_functions): Likewise. Add dumping of order to two other
10354 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
10355 instead of node uids.
10356 (propagate_pure_const): Likewise.
10357 (propagate_pure_const): Likewise.
10358 * ipa-utils.c (dump_cgraph_node_set): Likewise.
10359 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
10360 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
10362 * tree-pretty-print.c (dump_function_header): Likewise.
10363 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
10364 Print symbol order instead of node uids.
10366 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10368 * config/s390/s390.c (s390_register_move_cost): Don't impose the
10369 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
10371 2013-05-15 Richard Biener <rguenther@suse.de>
10373 PR tree-optimization/57275
10374 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
10375 return value for fail to do runtime alias checks for gather loads.
10377 2013-05-15 Jan Hubicka <jh@suse.cz>
10381 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
10382 weakrefs are not external.
10383 (lto_symtab_merge_decls): Fix thinko when dealing with
10384 non-lto_symtab decls.
10385 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
10386 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
10387 * varpool.c (dump_varpool_node): Dump more flags.
10389 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10391 * config/i386/i386.c (processor_alias_table): Add instruction
10392 FSGSBASE for AMD bdver3 architecture.
10394 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10396 * tree.c (warn_deprecated_use): Print file:line using locus color.
10397 * diagnostic.c (diagnostic_report_current_module): Print file:line
10398 and file:line:column using locus color.
10400 2013-05-14 Mike Stump <mikestump@comcast.net>
10402 * gdbinit.in: Add __null.
10404 2013-05-14 Mike Stump <mikestump@comcast.net>
10406 * recog.h: Rename struct recog_data to Recog_data.
10407 * recog.c: Likewise.
10408 * reload.c (can_reload_into): Likewise.
10409 * config/picochip/picochip.c: Likewise.
10411 2013-05-14 Mike Stump <mikestump@comcast.net>
10413 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
10415 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
10417 * resource.h (struct resources): Remove unch_memory member.
10418 (CLEAR_RESOURCE): Don't clear unch_memory.
10419 * resource.c (mark_referenced_resources): Don't set it.
10420 (mark_set_resources): Likewise.
10421 (mark_target_live_regs): Don't clear it.
10422 (init_resource_info): Likewise.
10423 * reorg.c (resource_conflicts_p): Don't compare it.
10424 (redundant_insn): Don't set it.
10426 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
10428 * emit-rtl.c (next_label): Remove unused function.
10429 (skip_consecutive_labels, link_cc0_insns): Move to ...
10430 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
10431 only place where these functions are used, and make them static.
10433 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
10435 * fold-const.c (fold_negate_expr): Handle vectors.
10436 (fold_truth_not_expr): Make it static.
10437 (fold_invert_truthvalue): New static function.
10438 (invert_truthvalue_loc): Handle vectors. Do not call
10439 fold_truth_not_expr directly.
10440 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
10441 <TRUTH_NOT_EXPR>: Do not cast to boolean.
10442 (fold_comparison): Handle vector constants.
10443 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
10444 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
10445 * tree.h (fold_truth_not_expr): Remove declaration.
10447 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
10449 * config/aarch64/aarch64-simd.md
10450 (aarch64_vcond_internal<mode>): Rename to...
10451 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
10452 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
10453 float modes. Clarify all iterator modes.
10454 (vcond<mode><mode>): Use new name for vcond expanders.
10455 (vcond<v_cmp_result><mode>): Likewise.
10456 (vcondu<mode><mode>: Likewise.
10457 * config/aarch64/iterators.md (VDQF_COND): New.
10459 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
10462 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
10463 variable for the shift amount. Check that we shift by non-negative
10466 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
10469 * config/arm/arm.h (EPILOGUE_USES): Only return true
10470 for LR_REGNUM after epilogue_completed.
10472 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
10474 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
10475 is error_mark_node.
10477 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10480 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
10481 and Solaris 11+/x86 with gld.
10482 * configure: Regenerate.
10484 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10486 * expmed.c (expand_shift_1): Canonicalize rotates by
10487 constant bitsize / 2 to bitsize - 1.
10488 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
10489 case ROTATERT>: Likewise.
10492 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10494 * config/i386/i386.md (rotateinv): New code attr.
10495 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10496 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10497 roll $31, %eax, etc.
10499 2013-05-14 Richard Biener <rguenther@suse.de>
10501 PR middle-end/57235
10502 * tree-eh.c (sink_clobbers): Give up for successors with
10503 multiple predecessors and no virtual uses.
10505 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
10507 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
10508 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
10510 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10512 PR middle-end/57251
10513 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
10514 the case when both op0 and op1 have VOIDmode.
10516 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10518 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
10519 in multiply-accumulate mode.
10521 2013-05-13 Guozhi Wei <carrot@google.com>
10523 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
10525 2013-05-13 Kai Tietz <ktietz@redhat.com>
10528 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
10529 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
10530 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
10531 * config/i386/i386.c (ix86_option_override_internal): Likewise.
10532 (ix86_expand_prologue): Likewise.
10533 (ix86_expand_split_stack_prologue): Likewise.
10534 (legitimate_pic_address_disp_p): Likewise.
10535 (legitimize_pic_address): Likewise.
10536 (legitimize_tls_address): Likewise.
10537 (legitimize_pe_coff_symbol): Likewise.
10538 (output_pic_addr_const): Likewise.
10539 (construct_plt_address): Likewise.
10540 (ix86_expand_call): Likewise.
10541 (x86_output_mi_thunk): Likewise.
10542 (x86_function_profiler): Likewise.
10544 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
10546 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
10547 similar switch cases.
10548 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
10549 (aarch64_simd_mov_to_<mode>low): Delete.
10550 (aarch64_simd_mov_to_<mode>high): Delete.
10551 (move_lo_quad_<mode>): Add w<-r alternative.
10552 (aarch64_simd_move_hi_quad_<mode>): Likewise.
10553 (aarch64_simd_mov_from_*): Update type attribute.
10554 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
10557 2013-05-13 Jan Hubicka <jh@suse.cz>
10559 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
10560 * config/i386/i386.c (ix86_compute_frame_layout,
10561 ix86_expand_epilogue, emit_i387_cw_initialization,
10562 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
10563 ix86_local_alignment): Fix use of size/speed predicates.
10565 2013-05-13 Jakub Jelinek <jakub@redhat.com>
10567 PR tree-optimization/45216
10568 PR tree-optimization/57157
10569 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
10570 the (-Y) & (B - 1) variant if OP is |.
10571 * expmed.c (expand_shift_1): For rotations by const0_rtx just
10572 return shifted. Use (-op1) & (prec - 1) as other_amount
10573 instead of prec - op1.
10575 2013-05-13 Martin Jambor <mjambor@suse.cz>
10577 PR middle-end/42371
10578 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
10579 (ipa_constant_data): New type.
10580 (ipa_jump_func): Use ipa_constant_data to hold information about
10581 constant jump functions.
10582 (ipa_get_jf_constant): Adjust to jump function type changes.
10583 (ipa_get_jf_constant_rdesc): New function.
10584 (ipa_param_descriptor): New field controlled_uses.
10585 (ipa_get_controlled_uses): New function.
10586 (ipa_set_controlled_uses): Likewise.
10587 * ipa-ref.h (ipa_find_reference): Declare.
10588 * ipa-prop.c (ipa_cst_ref_desc): New type.
10589 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
10591 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
10592 New parameter cs. Adjust all callers.
10593 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
10594 (remove_described_reference): New function.
10595 (jfunc_rdesc_usable): Likewise.
10596 (try_make_edge_direct_simple_call): Decrement controlled use count,
10597 attempt to remove reference if it hits zero.
10598 (combine_controlled_uses_counters): New function.
10599 (propagate_controlled_uses): Likewise.
10600 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
10601 (ipa_edge_duplication_hook): Duplicate reference descriptions.
10602 (ipa_print_node_params): Print described use counter.
10603 (ipa_write_jump_function): Adjust to jump function type changes.
10604 (ipa_read_jump_function): New parameter CS, pass it to
10605 ipa_set_jf_constant. Adjust caller.
10606 (ipa_write_node_info): Stream controlled use count
10607 (ipa_read_node_info): Likewise.
10608 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
10610 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
10611 count. Remove cloning-added reference if it reaches zero.
10612 * ipa-ref.c (ipa_find_reference): New function.
10614 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10616 * config/i386/i386.c (processor_target_table): Modified default
10617 alignment values for AMD BD and BT architectures.
10619 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
10621 * tree-vect-generic.c (uniform_vector_p): Move ...
10622 * tree.c (uniform_vector_p): ... here.
10623 * tree.h (uniform_vector_p): Declare it.
10624 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
10627 2013-05-13 Jakub Jelinek <jakub@redhat.com>
10629 PR tree-optimization/57230
10630 * tree-ssa-strlen.c (handle_char_store): Record length for
10631 array store from STRING_CST.
10633 PR tree-optimization/57230
10634 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
10637 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
10639 * config/epiphany/epiphany.c (epiphany_init): Check size of
10640 NUM_MODES_FOR_MODE_SWITCHING.
10641 (epiphany_expand_prologue):
10642 Remove CONFIG_REGNUM initial value handling code.
10643 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
10644 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
10645 (emit_set_fp_mode, epiphany_mode_after): Likewise.
10646 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
10647 Don't return 1 for FP_MODE_NONE.
10648 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
10649 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
10650 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
10651 * config/epiphany/epiphany.md (save_config): New pattern.
10653 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
10655 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
10657 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
10659 * config/i386/i386.md (memory): Handle sseishft1.
10660 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10661 (*vec_extractv2di_1): Ditto.
10663 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10665 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10668 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
10670 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10671 Add mthumb/march=armv7-a multilib.
10672 Add mthumb/march=armv7-r multilib.
10673 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10675 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10677 * config/v850/t-rtems: Add more multilibs.
10679 2013-05-10 Richard Biener <rguenther@suse.de>
10681 PR tree-optimization/57214
10682 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10683 not propagate from SSA names that occur in abnormal PHI nodes.
10685 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10687 * stor-layout.c (element_precision): New function.
10688 * machmode.h (element_precision): Declare it.
10689 * tree.c (build_minus_one_cst): New function.
10690 (element_precision): Likewise.
10691 * tree.h (build_minus_one_cst): Declare new function.
10692 (element_precision): Likewise.
10693 * fold-const.c (operand_equal_p): Use element_precision.
10694 (fold_binary_loc): Handle vector types.
10695 * convert.c (convert_to_integer): Use element_precision.
10696 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10699 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10701 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10702 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10703 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10704 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10705 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10706 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10707 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10708 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10709 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10710 (Uuw8): New constraints.
10711 (Usb4): Move into alphabetical order.
10712 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10713 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10714 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10715 previously unnamed patterns.
10716 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10717 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10718 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10719 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10720 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10721 of specific lengths.
10723 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10725 * config/i386/i386.md (rotateinv): New code attr.
10726 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10727 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10728 roll $31, %eax, etc.
10730 PR tree-optimization/45216
10731 PR tree-optimization/57157
10732 * tree-ssa-forwprop.c (simplify_rotate): New function.
10733 (ssa_forward_propagate_and_combine): Call it.
10735 2013-05-10 Richard Biener <rguenther@suse.de>
10737 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10738 disable peeling when we version for aliasing.
10739 (vector_alignment_reachable_p): Honor explicit user alignment.
10740 (vect_supportable_dr_alignment): Likewise.
10741 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10742 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10743 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10744 then peeling to arrange for the cost-model check to come first.
10746 2013-05-10 Alan Modra <amodra@gmail.com>
10748 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10749 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10750 * configure: Regenerate.
10752 2013-05-10 Alan Modra <amodra@gmail.com>
10755 * varasm.c (default_elf_select_section): Move !DECL_P check..
10756 (get_named_section): ..to here before calling get_section_name.
10758 (default_section_type_flags): Add DECL_P check.
10759 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10760 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10762 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10764 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10765 When using gen_stack_adjust_str with a register offset, add a
10766 REG_FRAME_RELATED_EXPR note.
10768 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10770 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10771 (*vec_extractv4si_zext_mem): Ditto.
10772 (*vec_extractv2di): Add 0->x and x->x alternatives.
10773 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10774 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10776 2013-05-09 Jason Merrill <jason@redhat.com>
10778 N3639 C++1y VLA support
10779 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10782 * tree.c (build_constructor_va): New.
10783 * tree.h: Declare it.
10785 2013-05-09 Martin Jambor <mjambor@suse.cz>
10788 * gimple-fold.c (canonicalize_constructor_val): Call
10789 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10791 2013-05-09 Jan Hubicka <jh@suse.cz>
10792 Richard Biener <rguenther@suse.de>
10795 * symtab.c (symtab_make_decl_local): Do not add private names.
10797 2013-05-09 Jan Hubicka <jh@suse.cz>
10800 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10801 (unlink_from_assembler_name_hash): Likewise.
10802 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10803 symtab_unregister_node, symtab_initialize_asm_name_hash,
10804 change_decl_assembler_name): Update.
10806 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10808 * config/aarch64/aarch64.md: New movtf split.
10809 (*movtf_aarch64): Update.
10810 (aarch64_movdi_tilow): Handle TF modes and rename to
10811 aarch64_movdi_<mode>low.
10812 (aarch64_movdi_tihigh): Handle TF modes and rename to
10813 aarch64_movdi_<mode>high
10814 (aarch64_movtihigh_di): Handle TF modes and rename to
10815 aarch64_mov<mode>high_di
10816 (aarch64_movtilow_di): Handle TF modes and rename to
10817 aarch64_mov<mode>low_di
10818 (aarch64_movtilow_tilow): Remove spurious whitespace.
10819 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10821 (aarch64_print_operand): Update.
10823 2013-05-09 Alan Modra <amodra@gmail.com>
10825 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10827 * configure: Regenerate.
10829 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10831 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10832 splitter preparation statements.
10833 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10834 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10837 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10839 * gimple-ssa-strength-reduction.c (count_candidates): Change
10840 return value to int.
10841 (analyze_candidates_and_replace): Change type of length to int.
10843 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10845 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
10846 (*vec_extract<mode>): Use VI12_128 mode iterator.
10847 (*vec_extract<mode>_mem): Ditto.
10848 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
10851 2013-05-08 Diego Novillo <dnovillo@google.com>
10856 2012-08-17 Diego Novillo <dnovillo@google.com>
10859 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
10860 * config.in: Regenerate.
10861 * configure: Regenerate.
10862 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
10864 2013-05-08 Jan Hubicka <jh@suse.cz>
10867 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
10868 * cgraph.h (symtab_node_base): Add unique_name.
10869 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10870 input_overwrite_node, input_varpool_node): Stream unique_name.
10871 * cgraphclones.c (cgraph_create_virtual_clone,
10872 cgraph_function_versioning): Set unique_name.
10873 * ipa.c (function_and_variable_visibility): Set unique_name.
10875 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10877 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
10878 (alloc_cand_and_find_basis): Restrict conditional candidate
10879 processing to CAND_MULTs.
10881 2013-05-08 Jan Hubicka <jh@suse.cz>
10884 lto-symtab.c (lto_symtab_symbol_p): New function.
10885 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
10886 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
10887 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
10888 Skip static symbols.
10890 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
10892 PR tree-optimization/57200
10893 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
10894 Only call inform if the preceding warning_at returns true.
10896 2013-05-07 Han Shen <shenhan@google.com>
10898 * cfgexpand.c (record_or_union_type_has_array_p): New function.
10899 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
10900 * common.opt (fstack-protector-strong): New option.
10901 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
10902 * doc/invoke.texi (Optimization Options): Document
10903 "-fstack-protector-strong".
10904 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
10906 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10908 * config/mips/mips.c (mips_machine_reorg2): Return 0.
10910 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
10912 * ira.c (update_equiv_regs): Add insn having equiv memory even if
10913 it is not lhs of the insn.
10914 (setup_reg_equiv): Remove insn having equiv memory which it is not
10916 * lra-constraints.c (process_address): Try to improve generation
10917 code for address base + disp.
10918 (lra_constraints): Make correct the code for checking insn setting
10919 up backward equivalence. Remove insn only if it is in the init
10921 * lra-eliminations.c (update_reg_eliminate): Change return value.
10922 (lra_eliminate): Use the result.
10924 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
10926 * config/i386/sse.md (ssescalarnummask): New mode attribute.
10927 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
10928 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
10929 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
10930 register target operands.
10931 (*vec_extractv8hi_sse2): New pattern.
10932 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
10933 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
10934 (*vec_extract<mode>_mem): New insn and split pattern.
10936 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
10938 * config/arm/arm.c (arm_asan_shadow_offset): New function.
10939 (TARGET_ASAN_SHADOW_OFFSET): Define.
10940 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
10941 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
10943 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10945 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
10946 (incr_vec_index): Return -1 if increment not found.
10947 (create_add_on_incoming_edge): Assert if increment not found.
10948 (record_increment): Limit number of increments recorded.
10949 (all_phi_incrs_profitable): Return false if an increment not found.
10950 (replace_profitable_candidates): Don't process increments that were
10952 (analyze_candidates_and_replace): Limit size of incr_vec.
10954 2013-05-07 Richard Biener <rguenther@suse.de>
10956 * calls.c (special_function_p): setjmp-like functions are leaf.
10957 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
10958 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
10960 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
10962 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
10963 (aarch64_simd_mov<mode>): New expander.
10964 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
10965 (aarch64_simd_mov_to_<mode>high): Likewise.
10966 (aarch64_simd_mov_from_<mode>low): Likewise.
10967 (aarch64_simd_mov_from_<mode>high): Likewise.
10968 (aarch64_dup_lane<mode>): Update.
10969 (aarch64_dup_lanedi): New instruction pattern.
10970 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
10971 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
10973 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10975 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
10976 (replace_mult_candidate): Remove unnecessary argument; remove
10977 unnecessary parameter from call to introduce_cast_before_cand.
10978 (replace_unconditional_candidate): Remove unnecessary parameter
10979 from call to replace_mult_candidate.
10980 (replace_conditional_candidate): Likewise.
10981 (insert_initializers): Use make_temp_ssa_name.
10982 (introduce_cast_before_cand): Remove unnecessary argument; use
10983 make_temp_ssa_name.
10984 (replace_one_candidate): Remove unnecessary argument; remove
10985 unnecessary parameter from calls to introduce_cast_before_cand.
10986 (replace_profitable_candidates): Remove unnecessary parameters
10987 from calls to replace_one_candidate.
10989 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10991 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
10992 phi def as possibly hiding a basis for a CAND_ADD whose operands
10993 have been commuted in the analysis.
10994 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
10996 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10998 * config/aarch64/aarch64.md
10999 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
11000 shift value between 0-4.
11002 2013-05-07 Richard Biener <rguenther@suse.de>
11004 * double-int.h (rshift): New overload.
11005 * double-int.c (rshift): New function.
11006 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
11007 (create_reference_ops_from_ref): Remove.
11008 (vn_reference_insert): Use shared ops for constructing the
11009 reference and copy it.
11011 2013-05-07 Richard Biener <rguenther@suse.de>
11013 PR middle-end/57190
11014 * tree-eh.c (sink_clobbers): Properly propagate
11015 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
11017 2013-05-07 Jakub Jelinek <jakub@redhat.com>
11019 PR tree-optimization/57149
11020 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
11021 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
11022 collect_phi_def_edges, execute_late_warn_uninitialized): Use
11023 uninit_undefined_value_p instead of ssa_undefined_value_p.
11026 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
11027 for modifier == EXPAND_INITIALIZER.
11029 2013-05-07 Anton Blanchard <anton@samba.org>
11031 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
11032 for powerpc64 little endian.
11033 * configure: Regenerate.
11035 2013-05-06 Graham Stott <grahams@btinternet.com>
11037 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
11038 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
11039 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
11040 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
11042 2013-05-06 Graham Stott <grahams@btinternet.com>
11044 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
11045 codes which allow non-lvalues.
11047 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
11049 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
11050 components are all 1s.
11051 (integer_minus_onep): New function.
11052 * tree.h (integer_minus_onep): Declare it.
11053 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
11054 integer_minus_onep instead of integer_all_onesp.
11056 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11059 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
11060 variations of these patterns.
11062 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11064 * config/i386/i386.md (isa): Add x64_sse4 member.
11065 (enabled): Handle x64_sse4.
11066 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
11067 instruction for 64bit SSE4_1 targets. Update insn attributes.
11068 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
11069 instruction for SSE4_1 targets. Update insn attributes.
11070 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11071 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
11072 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
11074 (*vec_extractv4si): Rename from *sse4_1_pextrd.
11075 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
11076 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
11078 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11081 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
11083 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
11085 * final.c (do_assembler_dialects): Don't handle curly braces and
11086 vertical bar escaped by % as dialect delimiters.
11087 (output_asm_insn): Print curly braces and vertical bar if escaped
11088 by % and ASSEMBLER_DIALECT defined.
11089 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
11090 * doc/tm.texi: Regenerated.
11092 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11094 * config/mips/mips.c: Include tree-pass.h.
11095 (mips_reorg): Split in pre- and post-dbr_schedule parts.
11096 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
11097 (pass_mips_machine_reorg2): New machine specific pass.
11098 (insert_pass_mips_machine_reorg2): New pass plugin definition.
11099 (mips_option_override): Register the new pass.
11100 * rtl.h (cleanup_barriers): Remove prototype.
11101 (dbr_schedule): Likewise.
11102 * jump.c (cleanup_barriers): Make static.
11103 * reorg.c (dbr_schedule): Likewise.
11105 2013-05-06 Richard Biener <rguenther@suse.de>
11107 PR tree-optimization/57185
11108 * tree-parloops.c (add_field_for_reduction): Handle anonymous
11109 SSA names properly.
11111 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11114 * config/i386/i386.c (add_parameter_dependencies): Add dependence
11115 between "first_arg" and "insn", not "last" and "insn".
11117 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11119 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
11120 (find_candidates_in_block): Re-enable slsr_process_phi.
11121 (create_phi_basis): Fix double counting of candidate adjustment.
11123 2013-05-06 Richard Biener <rguenther@suse.de>
11125 PR middle-end/57147
11126 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
11127 the edge is also fallthru, preserve it and just clear the
11129 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
11130 also complex, preserve that and just clear the fallthru flag.
11131 * tree-inline.c (update_ssa_across_abnormal_edges): Also
11132 update virtual operands.
11134 2013-05-06 Alan Modra <amodra@gmail.com>
11136 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
11137 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
11138 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
11139 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
11140 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11141 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
11143 2013-05-06 Alan Modra <amodra@gmail.com>
11145 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
11146 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
11147 (DEFAULT_ASM_ENDIAN): Define.
11148 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11149 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
11150 Update -K PIC clause from sysv4.h.
11151 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
11152 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
11154 2013-05-06 Alan Modra <amodra@gmail.com>
11156 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
11157 twice for little-endian.
11158 (ashrdi3_no_power, ashrdi3): Support little-endian.
11160 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11163 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
11164 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
11166 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
11167 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
11168 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
11171 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
11172 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11174 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
11175 * config/pa/pa.opt: Make mbig-switch a no-op.
11176 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
11177 (CASE_VECTOR_MODE): Always return SImode.
11178 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
11179 for the !TARGET_BIG_SWITCH case.
11180 * config/pa/pa-linux.h: Likewise.
11181 * config/pa/pa-openbsd.h: Likewise.
11182 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
11183 * config/pa/pa.md (short_jump): Remove define_insn.
11184 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
11185 (casesi0): Remove define_insn.
11186 (type): Remove btable_branch.
11187 (pa_combine_type): Likewise.
11188 (in_nullified_branch_delay): Likewise.
11189 (in_call_delay): Likewise.
11190 (define_delay): Likewise.
11191 (define_insn_reservation "Z3"): Likewise.
11192 (define_insn_reservation "Z4"): Likewise.
11193 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
11194 (pa_adjust_insn_length): Remove adjustment for btable branches.
11195 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
11198 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
11200 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11201 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
11202 Add m->r,x alternatives.
11203 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
11204 splitters using SWI48x mode iterator.
11205 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
11206 TARGET_64BIT. Add m->x alternative.
11207 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
11208 Add o->x alternative. Enable for TARGET_SSE.
11209 (sse_storeq): Remove expander.
11210 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
11211 with memory input operand.
11212 (*vec_extractv2di_1 splitter): New.
11213 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
11214 * config/i386/i386.md (ssevecmodelower): New mode attribute.
11216 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11218 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
11219 (INT_LOWPART): Delete.
11220 (extract_MB): Adjust.
11221 (extract_ME): Adjust.
11222 (print_operand): Adjust.
11224 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11226 * config/rs6000/predicates.md (reg_or_add_cint_operand,
11227 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11228 (reg_or_logical_cint_operand, easy_fp_constant,
11229 logical_const_operand): Delete "CONST_DOUBLE" case.
11230 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
11231 "HOST_BITS_PER_WIDE_INT == 64" test.
11232 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
11233 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
11234 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
11235 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
11237 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
11238 CONST_DOUBLE DImode/VOIDmode case.
11239 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
11240 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11241 CONST_DOUBLE VOIDmode case.
11242 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11243 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
11244 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
11245 Delete CONST_DOUBLE case.
11246 (splitters for mov FMOVE64 const_double): Delete
11247 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11248 "HOST_BITS_PER_WIDE_INT >= 64" test.
11249 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
11251 (mov DI const_double): Delete.
11253 2013-05-04 Jakub Jelinek <jakub@redhat.com>
11255 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
11256 on op shows all bits zero in mode of a lowpart subreg, return zero.
11258 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
11261 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
11262 to save TFmode registers and DImode to save TImode registers for
11263 caller save operations.
11264 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
11265 mark being partially clobbered since they only use the first
11268 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
11269 and TDmode only use the upper 64-bits of each VSX register.
11271 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11273 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
11274 (find_candidates_in_block): Disable slsr_process_phi.
11276 2013-05-03 Guozhi Wei <carrot@google.com>
11278 * coverage.c (coverage_obj_init): Move the construction of gcov
11280 (build_init_ctor): ... here.
11282 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11284 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
11285 (slsr_cand_d): Redefine def_phi.
11286 (stride_status, phi_adjust_status, count_phis_status): New enums.
11287 (find_phi_def): New.
11288 (find_basis_for_base_expr): New.
11289 (find_basis_for_candidate): Handle hidden bases.
11290 (alloc_cand_and_find_basis): Handle phi candidates.
11291 (slsr_process_phi): New.
11292 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
11293 (create_mul_imm_cand): Likewise.
11294 (create_add_ssa_cand): Exclude phi base candidates.
11295 (create_add_imm_cand): Likewise.
11296 (slsr_process_cast): Likewise.
11297 (slsr_process_copy): Likewise.
11298 (find_candidates_in_block): Handle phi candidates.
11299 (dump_candidate): Likewise.
11300 (unconditional_cands): Delete.
11301 (unconditional_cands_with_known_stride_p): Delete.
11302 (phi_dependent_cand_p): New.
11303 (cand_increment): Handle phi-dependent candidates.
11304 (replace_dependent): Delete.
11305 (replace_mult_candidate): New.
11306 (replace_unconditional_candidate): New.
11307 (incr_vec_index): Move to avoid forward reference.
11308 (create_add_on_incoming_edge): New.
11309 (create_phi_basis): New.
11310 (replace_dependents): Delete.
11311 (replace_conditional_candidate): New.
11312 (phi_add_costs): New.
11313 (replace_uncond_cands_and_profitable_phis): New.
11314 (record_increment): Handle phi adjustments.
11315 (record_phi_increments): New.
11316 (record_increments): Handle phi adjustments.
11317 (phi_incr_cost): New.
11318 (lowest_cost_path): Handle phis.
11319 (total_savings): Likewise.
11320 (analyze_increments): Likewise.
11321 (ncd_with_phi): New.
11322 (ncd_of_cand_and_phis): New.
11323 (nearest_common_dominator_for_cands): Handle phi increments.
11324 (all_phi_incrs_profitable): New.
11325 (replace_profitable_candidates): Handle phi-dependent candidates.
11326 (analyze_candidates_and_replace): Likewise.
11328 2013-05-03 Teresa Johnson <tejohnson@google.com>
11331 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
11332 do not exceed REG_BR_PROB_BASE.
11334 2013-05-03 Jeff Law <law@redhat.com>
11336 PR tree-optimization/57144
11337 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
11338 operand of the condition will bit into the new type when eliminating
11339 a cast feeding a condition.
11341 2013-05-03 Jakub Jelinek <jakub@redhat.com>
11343 PR rtl-optimization/57130
11344 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
11345 of COMPARE as in_code to the recursive call if needed.
11347 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
11349 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
11350 (enabled): Handle new members.
11351 * config/i386/sse.md (*vec_concatv2si): Merge from
11352 *vec_concatv2si_sse2 and vec_concatv2si_sse.
11353 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
11355 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
11357 PR tree-optimization/57027
11358 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
11359 for fnms opportunity, check we got the prerequisite kind
11360 of tree / gimple before using accessor functions.
11362 2013-05-03 Richard Biener <rguenther@suse.de>
11364 * double-int.h (lshift): New overload without precision
11365 and arith argument.
11366 (operator *=, operator +=, operator -=): Move ...
11367 * double-int.c (operator *=, operator +=, operator -=): ... here
11368 and implement more efficiently.
11369 (mul_double_with_sign): Remove.
11370 (lshift_double): Adjust to take unsinged shift argument, push
11371 dispatching code to callers.
11372 (mul_double_wide_with_sign): Add early out for callers that
11373 are not interested in high parts or overflow.
11374 (lshift): New function.
11375 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
11376 dispatch code here.
11377 (lrotate, rrotate): Use logical shifts.
11378 * expr.c (get_inner_reference): Use lshift.
11379 * fixed-value.c (do_fixed_divide): Likewise.
11380 * tree-dfa.c (get_ref_base_and_extent): Likewise.
11381 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
11382 (indirect_refs_may_alias_p): Likewise.
11383 (stmt_kills_ref_p_1): Likewise.
11385 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11387 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
11389 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11391 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
11392 scalar form of FABD instruction.
11394 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
11396 * lra-constraints.c (process_alt_operands): Add checking alt
11397 number to choose the best alternative.
11399 2013-05-02 Richard Biener <rguenther@suse.de>
11401 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
11402 bitmap and its handling.
11403 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
11405 2013-05-02 Richard Biener <rguenther@suse.de>
11407 PR middle-end/57140
11408 * tree-inline.c (copy_loops): Properly handle removed loops.
11409 (copy_cfg_body): Mark destination loops for fixup if source
11410 loops needed fixup.
11412 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
11415 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
11416 generating simple_return for naked functions.
11418 2013-05-02 Martin Jambor <mjambor@suse.cz>
11420 PR middle-end/56988
11421 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
11422 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
11424 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
11425 ipa_agg_replacement_value structures.
11426 (known_aggs_to_agg_replacement_list): Likewise.
11427 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
11428 (read_agg_replacement_chain): Likewise.
11429 (ipcp_transform_function): Also check that by_ref flags match.
11431 2013-05-02 Richard Biener <rguenther@suse.de>
11433 * graphds.h (struct graph): Add obstack member.
11434 * graphds.c (new_graph): Initialize obstack and allocate
11436 (add_edge): Allocate edge from the obstack.
11437 (free_graph): Free the obstack instead of all edges and vertices.
11439 2013-05-02 Teresa Johnson <tejohnson@google.com>
11441 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
11443 * cfg.c (update_bb_profile_for_threading): Ditto.
11444 * tree-inline.c (copy_bb): Ditto.
11445 (copy_edges_for_bb): Ditto.
11446 (initialize_cfun): Ditto.
11447 (copy_cfg_body): Ditto.
11448 (expand_call_inline): Ditto.
11449 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
11450 (estimate_node_size_and_time): Ditto.
11451 (inline_merge_summary): Ditto.
11452 * cgraphclones.c (cgraph_clone_edge): Ditto.
11453 (cgraph_clone_node): Ditto.
11454 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11455 (compute_trg_info): Ditto.
11457 2013-05-02 Ian Bolton <ian.bolton@arm.com>
11459 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
11460 S reg when fp attribute set.
11461 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
11463 2013-05-02 Ian Bolton <ian.bolton@arm.com>
11465 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
11467 (*and_one_cmplsi3_compare0_uxtw): Likewise.
11468 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
11469 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
11471 2013-05-02 Richard Biener <rguenther@suse.de>
11473 * tree-scalar-evolution.c (scev_info_hasher): Remove.
11474 (struct instantiate_cache_entry): New type.
11475 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
11476 (struct instantiate_cache_type): New type.
11477 (set_instantiated_value, get_instantiated_value): Remove.
11478 (get_instantiated_value_entry): New function.
11479 (instantiate_scev_name): Use the new cache and adjust.
11480 (instantiate_scev_poly): Adjust.
11481 (instantiate_scev_binary): Likewise.
11482 (instantiate_array_ref): Likewise.
11483 (instantiate_scev_convert): Likewise.
11484 (instantiate_scev_not): Likewise.
11485 (instantiate_scev_3): Likewise.
11486 (instantiate_scev_2): Likewise.
11487 (instantiate_scev_r): Likewise.
11488 (instantiate_scev): Likewise.
11489 (resolve_mixers): Likewise.
11491 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
11494 * lra-constraints.c (best_small_class_operands_num): Remove.
11495 (process_alt_operands): Remove small_class_operands_num. Take
11496 small classes operands into losers and only if the operand is not
11497 matched. Modify debugging output.
11498 (curr_insn_transform): Remove best_small_class_operands_num.
11501 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11503 * config/aarch64/aarch64-builtins.c
11504 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
11505 * config/aarch64/aarch64-simd-builtins.def
11506 (reduc_splus_): Add new modes.
11507 (reduc_uplus_): New.
11508 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
11509 (reduc_uplus_v4sf): Likewise.
11510 (reduc_splus_v4sf): Likewise.
11511 (aarch64_addv<mode>): Likewise.
11512 (reduc_uplus_<mode>): Likewise.
11513 (reduc_splus_<mode>): Likewise.
11514 (aarch64_addvv2di): Likewise.
11515 (reduc_uplus_v2di): Likewise.
11516 (reduc_splus_v2di): Likewise.
11517 (aarch64_addvv2si): Likewise.
11518 (reduc_uplus_v2si): Likewise.
11519 (reduc_splus_v2si): Likewise.
11520 (reduc_<sur>plus_<mode>): New.
11521 (reduc_<sur>plus_v2di): Likewise.
11522 (reduc_<sur>plus_v2si): Likewise.
11523 (reduc_<sur>plus_v4sf): Likewise.
11524 (aarch64_addpv4sf): Likewise.
11525 * config/aarch64/arm_neon.h
11526 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
11527 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
11528 add UNSPEC_SADDV, UNSPEC_UADDV.
11530 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
11532 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11534 * config/aarch64/arm_neon.h
11535 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
11537 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11539 * config/aarch64/aarch64-builtins
11540 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
11542 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11544 * config/aarch64/aarch64-simd-builtins.def
11545 (reduc_smax_): New.
11546 (reduc_smin_): Likewise.
11547 (reduc_umax_): Likewise.
11548 (reduc_umin_): Likewise.
11549 (reduc_smax_nan_): Likewise.
11550 (reduc_smin_nan_): Likewise.
11553 (smax): Update for V2SF, V4SF and V2DF modes.
11556 (smin_nan): Likewise.
11557 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
11558 (<su><maxmin><mode>3): ...This, refactor.
11559 (s<maxmin><mode>3): New.
11560 (<maxmin_uns><mode>3): Likewise.
11561 (reduc_<maxmin_uns>_<mode>): Refactor.
11562 (reduc_<maxmin_uns>_v4sf): Likewise.
11563 (reduc_<maxmin_uns>_v2si): Likewise.
11564 (aarch64_<fmaxmin><mode>: Remove.
11565 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
11567 (vmin<q>_f<32,64>): Likewise.
11568 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11570 (su): Add mappings for smax, smin, umax, umin.
11572 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11573 (FMAXMIN): Rename as...
11574 (FMAXMIN_UNS): ...This.
11576 (fmaxminv): Likewise.
11577 (fmaxmin): Likewise.
11579 (maxmin_uns_op): Likewise.
11581 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11583 * config/aarch64/arm_neon.h
11584 (vac<ge, gt><sd>_f<32, 64>): Rename to...
11585 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
11586 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
11588 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11590 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
11591 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
11593 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11595 * config/aarch64/aarch64-simd.md
11596 (vcond<mode>_internal): Handle special cases for constant masks.
11597 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
11598 (vcondu<mode><mode>): Likewise.
11599 (vcond<v_cmp_result><mode>): New.
11601 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11603 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
11604 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
11605 * config/aarch64/aarch64-simd-builtins.def
11606 (cmeq): Update to BUILTIN_VALLDI.
11611 * config/aarch64/arm_neon.h
11612 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
11613 to builtins or C as appropriate.
11615 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11617 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
11619 (cmhi): Rename to...
11621 * config/aarch64/aarch64-simd.md
11622 (simd_mode): Add SF.
11623 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
11624 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
11625 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
11626 (cstore<mode>_neg): ...This.
11627 * config/aarch64/iterators.md
11629 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
11630 (COMPARISONS): New.
11631 (UCOMPARISONS): Likewise.
11632 (optab): Add missing comparisons.
11638 (VCMP_S): Likewise.
11639 (VCMP_U): Likewise.
11640 (V_cmp_result): Add DF, SF modes.
11641 (v_cmp_result): Likewise.
11643 (vmtype): Likewise.
11644 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
11646 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
11648 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
11649 define_insn to define_insn_and_split.
11650 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
11651 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
11652 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
11653 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
11654 (thumb2_negscc): Likewise.
11656 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11658 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11660 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11662 * config/arm/thumb2.md: Remove trailing whitespaces.
11664 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11666 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11667 Use gen_int_mode rather than GEN_INT.
11669 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11671 * value-prof.c (stream_in_histogram_value): Remove the strayed
11674 2013-04-30 Richard Biener <rguenther@suse.de>
11676 PR middle-end/57122
11677 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11679 2013-04-30 Richard Biener <rguenther@suse.de>
11681 PR middle-end/57107
11682 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11684 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11686 PR rtl-optimization/56957
11687 PR rtl-optimization/57105
11688 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11689 variable. Use just INSN_UID for determining whether an insn
11690 should be only disconnected from the insn stream.
11691 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11693 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11695 PR tree-optimization/57104
11696 * tsan.c (instrument_expr): Don't instrument accesses to
11697 DECL_HARD_REGISTER VAR_DECLs.
11699 2013-04-30 Richard Biener <rguenther@suse.de>
11701 * function.h (loops_for_fn): New inline function.
11702 (set_loops_for_fn): Likewise.
11703 * cfgloop.h (place_new_loop): Add struct function parameter.
11704 (get_loop): Likewise.
11705 (get_loops): Likewise.
11706 (number_of_loops): Likewise.
11707 (fel_next): Adjust.
11708 (fel_init): Likewise.
11709 * cfg.c (get_loop_copy): Adjust.
11710 * cfgloop.c (flow_loops_dump): Likewise.
11711 (record_loop_exits): Likewise.
11712 (verify_loop_structure): Likewise.
11713 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11714 (estimate_reg_pressure_cost): Likewise.
11715 (mark_loop_exit_edges): Likewise.
11716 * cfgloopmanip.c (place_new_loop): Likewise.
11717 (add_loop): Likewise.
11718 (duplicate_loop): Likewise.
11719 * graph.c (draw_cfg_nodes): Likewise.
11720 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11721 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11722 (extract_affine_chrec): Likewise.
11723 (build_scop_iteration_domain): Likewise.
11724 * graphite.c (graphite_initialize): Likewise.
11725 * ira-build.c (create_loop_tree_nodes): Likewise.
11726 (more_one_region_p): Likewise.
11727 (rebuild_regno_allocno_maps): Likewise.
11728 (mark_loops_for_removal): Likewise.
11729 (mark_all_loops_for_removal): Likewise.
11730 (remove_unnecessary_regions): Likewise.
11731 (ira_build): Likewise.
11732 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11733 * loop-init.c (fix_loop_structure): Likewise.
11734 (gate_rtl_move_loop_invariants): Likewise.
11735 (gate_rtl_unswitch): Likewise.
11736 (gate_rtl_unroll_and_peel_loops): Likewise.
11737 (rtl_doloop): Likewise.
11738 * lto-streamer-in.c (input_cfg): Likewise.
11739 * lto-streamer-out.c (output_cfg): Likewise.
11740 * modulo-sched.c (sms_schedule): Likewise.
11741 * predict.c (tree_estimate_probability): Likewise.
11742 (tree_estimate_probability_driver): Likewise.
11743 (estimate_loops): Likewise.
11744 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11745 (move_sese_region_to_fn): Likewise.
11746 (debug_loop_num): Likewise.
11747 * tree-chrec.c (chrec_evaluate): Likewise.
11748 (hide_evolution_in_other_loops_than_loop): Likewise.
11749 (chrec_component_in_loop_num): Likewise.
11750 (reset_evolution_in_loop): Likewise.
11751 (evolution_function_is_invariant_rec_p): Likewise.
11752 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11753 * tree-inline.c (copy_loops): Likewise.
11754 (copy_cfg_body): Likewise.
11755 (tree_function_versioning): Likewise.
11756 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11757 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11759 (add_to_evolution_1): Likewise.
11760 (scev_const_prop): Likewise.
11761 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11762 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11763 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11764 (tree_ssa_lim_initialize): Likewise.
11765 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11766 (verify_loop_closed_ssa): Likewise.
11767 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11768 (tree_ssa_loop_im): Likewise.
11769 (tree_ssa_loop_unswitch): Likewise.
11770 (tree_vectorize): Likewise.
11771 (check_data_deps): Likewise.
11772 (tree_ssa_loop_ivcanon): Likewise.
11773 (tree_ssa_loop_bounds): Likewise.
11774 (tree_complete_unroll): Likewise.
11775 (tree_complete_unroll_inner): Likewise.
11776 (tree_parallelize_loops): Likewise.
11777 (tree_ssa_loop_prefetch): Likewise.
11778 (tree_ssa_loop_ivopts): Likewise.
11779 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11780 * tree-vectorizer.c (vectorize_loops): Likewise.
11782 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11784 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11785 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11786 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11787 with EABI_LINK_SPEC.
11789 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11792 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11795 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11798 * lra-constraints.c (process_alt_operands): Discourage a bit more
11799 using memory for pseudos. Print cost dump for alternatives.
11800 Modify cost values for conflicts with early clobbers.
11801 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11803 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11806 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11808 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11810 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11811 from/to S register.
11812 (movdi_aarch64): Support LDR/STR from/to D register.
11814 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11816 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11817 or higher by default.
11819 2013-04-29 Richard Biener <rguenther@suse.de>
11821 PR middle-end/57075
11822 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11823 even if not adding abnormal edges for calls that can make
11826 2013-04-29 Richard Biener <rguenther@suse.de>
11828 PR middle-end/57103
11829 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11831 (move_stmt_r): Remove redundant checking.
11833 2013-04-29 Teresa Johnson <tejohnson@google.com>
11836 * basic-block.h (apply_scale): New function.
11837 (apply_probability): Use apply_scale.
11838 * gimple-streamer-in.c (input_bb): Ditto.
11839 * lto-streamer-in.c (input_cfg): Ditto.
11840 * lto-cgraph.c (merge_profile_summaries): Ditto.
11841 * tree-optimize.c (execute_fixup_cfg): Ditto.
11842 * tree-inline.c (copy_bb): Update comment to use apply_scale.
11843 (copy_edges_for_bb): Ditto.
11844 (copy_cfg_body): Ditto.
11846 2013-04-29 Tom de Vries <tom@codesourcery.com>
11848 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
11849 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
11850 (tail_merge_optimize): Handle current_loops == NULL.
11852 2013-04-26 Jeff Law <law@redhat.com>
11854 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
11855 (simplify_cond_using_ranges): Generalize code to simplify
11856 COND_EXPRs where one argument is a constant and the other
11857 is an SSA_NAME created by an integral type conversion.
11859 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11861 * config/arm/arm.md (store_minmaxsi): Use only when
11862 optimize_insn_for_size_p.
11864 2013-04-29 Christian Bruel <christian.bruel@st.com>
11867 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
11869 2013-04-29 Richard Biener <rguenther@suse.de>
11871 PR middle-end/57089
11872 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
11873 loop tree make sure to schedule a fixup for the child as well.
11874 (expand_omp_for_generic): Properly add loops.
11875 (expand_omp_for_static_nochunk): Likewise.
11876 (expand_omp_for_static_chunk): Likewise.
11877 (expand_omp_for): For the degenerate case fixup loops.
11878 (expand_omp_sections): Fix default bb placement in loops.
11879 (expand_omp_atomic_pipeline): Properly add loops.
11881 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11883 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
11885 2013-04-29 Tom de Vries <tom@codesourcery.com>
11887 * tree-ssa-tail-merge.c: Update header comment.
11889 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11891 * config/aarch64/arm_neon.h
11892 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
11893 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
11894 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
11895 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
11896 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
11897 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
11898 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
11899 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
11901 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11903 * config/aarch64/aarch64-simd.md
11904 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
11905 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
11906 fix_trunc, fixuns_trunc.
11907 (ftrunc<VDQF:mode>2): New.
11908 * config/aarch64/iterators.md (optab): Add fix, fixuns.
11909 (fix_trunc_optab): New.
11911 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11913 * config/aarch64/aarch64-builtins.c
11914 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
11915 iceilf, lround, iroundf.
11917 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11920 * config/i386/i386.h (enum ix86_tune_indices)
11921 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
11922 New, split from X86_TUNE_INTER_UNIT_MOVES.
11923 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
11924 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
11925 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
11926 (TARGET_INTER_UNIT_MOVES): Remove.
11927 * config/i386/i386.c (initial_ix86_tune_features): Update.
11928 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
11929 (ix86_expand_convert_uns_didf_sse): Use
11930 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11931 (ix86_expand_vector_init_one_nonzero): Ditto.
11932 (ix86_expand_vector_init_interleave): Ditto.
11933 (inline_secondary_memory_needed): Return true for moves from SSE class
11934 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
11935 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
11936 * config/i386/constraints.md (Yi, Ym): Depend on
11937 TARGET_INTER_UNIT_MOVES_TO_VEC.
11938 (Yj, Yn): New constraints.
11939 * config/i386/i386.md (*movdi_internal): Change constraints of
11940 operand 1 from Yi to Yj and from Ym to Yn.
11941 (*movsi_internal): Ditto.
11942 (*movdf_internal): Ditto.
11943 (*movsf_internal): Ditto.
11944 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
11945 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11946 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
11947 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
11948 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
11949 * config/i386/sse.md (movdi_to_sse): Ditto.
11950 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
11951 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
11952 TARGET_INTER_UNIT_MOVES.
11953 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
11954 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
11955 instead of TARGET_INTER_UNIT_MOVES.
11956 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
11957 operand 1 from Yi to Yj and from Ym to Yn.
11959 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11961 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
11962 (float_truncate_hi_): Likewise.
11963 (float_extend_lo_): Likewise.
11964 (float_truncate_lo_): Likewise.
11965 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
11966 (aarch64_float_extend_lo_v2df): Likewise.
11967 (vec_unpacks_hi_v4sf): Likewise.
11968 (aarch64_float_truncate_lo_v2sf): Likewise.
11969 (aarch64_float_truncate_hi_v4sf): Likewise.
11970 (vec_pack_trunc_v2df): Likewise.
11971 (vec_pack_trunc_df): Likewise.
11973 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11975 * config/aarch64/aarch64-builtins.c
11976 (aarch64_fold_builtin): Fold float conversions.
11977 * config/aarch64/aarch64-simd-builtins.def
11978 (floatv2si, floatv4si, floatv2di): New.
11979 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
11980 * config/aarch64/aarch64-simd.md
11981 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
11982 * config/aarch64/iterators.md (FLOATUORS): New.
11983 (optab): Add float, floatuns.
11984 (su_optab): Likewise.
11986 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11988 * config/aarch64/aarch64-builtins.c
11989 (aarch64_builtin_vectorized_function): Use new names for
11991 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
11992 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
11993 (fcvtzu): Split as...
11994 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
11995 (fcvtas): Split as...
11996 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
11997 (fcvtau): Split as...
11998 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
11999 (fcvtps): Split as...
12000 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
12001 (fcvtpu): Split as...
12002 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
12003 (fcvtms): Split as...
12004 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
12005 (fcvtmu): Split as...
12006 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
12007 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
12008 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
12009 (lfrintnusf, lfrintnudf): Likewise.
12010 * config/aarch64/aarch64-simd.md
12011 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
12013 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
12014 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
12015 (fcvt_pattern): Likewise.
12017 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12019 * config/aarch64/aarch64-simd.md
12020 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
12021 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
12023 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12025 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
12026 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
12027 (vrnd<a,m,n,p>_f32): Implement using builtins.
12028 (vrnd<i,x><q>_f<32, 64>): New.
12030 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12032 * config/aarch64/aarch64-builtins.c
12033 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
12034 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
12035 (frintz): Rename to...
12037 (frintp): Rename to...
12039 (frintm): Rename to...
12041 (frinti): Rename to...
12042 (nearbyint): ...this.
12043 (frintx): Rename to...
12045 (frinta): Rename to...
12047 * config/aarch64/aarch64-simd.md
12048 (aarch64_frint<frint_suffix><mode>): Delete.
12049 (<frint_pattern><mode>2): Convert to insn.
12050 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
12051 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
12052 (frint_pattern): Likewise.
12053 (frint_suffix): Likewise.
12055 2013-04-29 Richard Biener <rguenther@suse.de>
12057 PR tree-optimization/57081
12058 * loop-init.c: Include tree-flow.h.
12059 (loop_optimizer_finalize): Free number of iteration estimates.
12060 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
12062 2013-04-29 Jakub Jelinek <jakub@redhat.com>
12064 PR tree-optimization/57083
12065 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
12066 non-singleton shift count range, zero extend low_bound for uns case.
12068 * config/i386/predicates.md (general_vector_operand): New predicate.
12069 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
12070 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
12071 if they aren't nonimmediate operands. If their original values
12072 satisfy const_vector_equal_evenodd_p, don't shift them.
12073 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
12074 predicates. For the SSE4.1 case force operands[{1,2}] into registers
12075 if not nonimmediate_operand.
12076 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
12077 instead of register_operand.
12078 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
12080 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
12082 * stor-layout.c (finalize_size_functions): Allocate a structure and
12083 reset cfun before dumping the functions.
12085 2013-04-27 Jakub Jelinek <jakub@redhat.com>
12087 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
12090 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
12091 use xop_pmacsdqh if uns_p.
12092 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
12093 the immediate rotate count.
12095 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
12097 * rtl.h (struct rtx_def): Add comment for field jump.
12098 (LRA_SUBREG_P): New macro.
12099 * recog.c (register_operand): Check LRA_SUBREG_P.
12100 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12102 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12103 (lra_final_code_change): Skip subreg change for operators.
12104 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12105 if there are no operand changes.
12106 * lra-constraints.c (curr_insn_set): New.
12107 (match_reload): Set LRA_SUBREG_P.
12108 (emit_spill_move): Ditto.
12109 (check_and_process_move): Use curr_insn_set. Process only single
12110 set insns. Don't initialize sec_mem_p and change_p.
12111 (simplify_operand_subreg): Use LRA_SUBREG_P.
12112 (reg_in_class_p): New function.
12113 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12114 of #ifdef. Add code to remove cycling.
12115 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12116 non-null disp. Reload inner instead of disp when base and index
12117 are null. Try to put lo_sum into register.
12118 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12119 (check_and_process_move): Move code for move cost check to
12120 simple_move_p. Remove equiv_substitution.
12121 (simple_move_p): New function.
12122 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12123 curr_insn_set. Call check_and_process_move only for single set
12124 insns. Use the new function. Move call of check_and_process_move
12125 after operand equiv substitution and address process.
12127 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12130 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
12131 with nonlocal goto receivers or returns twice calls, ignore
12132 unininitialized values from abnormal edges to nl goto receiver
12133 or returns twice call.
12135 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12137 PR tree-optimization/57051
12138 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
12139 and VEC_RSHIFT_EXPR if shift count is a multiple of element
12142 2013-04-26 Richard Biener <rguenther@suse.de>
12144 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
12145 (expand_omp_taskreg): Likewise. Mark loops for fixup.
12146 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
12147 (fixup_loop_arrays_after_move): New function.
12148 (move_sese_region_to_fn): Properly outline the loop tree parts
12149 of the SESE region.
12151 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
12153 * config/i386/i386.md (type, unit): Fix long lines.
12155 2013-04-26 Richard Biener <rguenther@suse.de>
12157 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
12158 (lto-streamer-out.o): Likewise.
12159 * cfgloop.c (init_loops_structure): Export, add struct function
12160 argument and adjust.
12161 (flow_loops_find): Adjust.
12162 * cfgloop.h (enum loop_estimation): Add EST_LAST.
12163 (init_loops_structure): Declare.
12164 * lto-streamer-in.c: Include cfgloop.h.
12165 (input_cfg): Input the loop tree.
12166 * lto-streamer-out.c: Include cfgloop.h.
12167 (output_cfg): Output the loop tree.
12168 (output_struct_function_base): Do not drop PROP_loops.
12170 2013-03-26 Richard Biener <rguenther@suse.de>
12172 * tree-cfg.c (execute_build_cfg): Build the loop tree.
12173 (pass_build_cfg): Provide PROP_loops.
12174 (move_sese_region_to_fn): Remove loops that are outlined into fn
12176 * tree-inline.c: Include cfgloop.h.
12177 (initialize_cfun): Do not drop PROP_loops.
12178 (copy_loops): New function.
12179 (copy_cfg_body): Copy loop structure.
12180 (tree_function_versioning): Initialize destination loop tree.
12181 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
12182 (pass_parallelize_loops): Do IL verification.
12183 * loop-init.c (loop_optimizer_init): Fixup loops if required.
12184 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
12185 the CFG make sure we fixup loops as well.
12186 * tree-ssa-tail-merge.c: Include cfgloop.h.
12187 (replace_block_by): When merging loop latches mark loops for fixup.
12188 * lto-streamer-out.c (output_struct_function_base): Drop
12189 PROP_loops for now.
12190 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
12191 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
12192 * ipa-split.c: Include cfgloop.h.
12193 (split_function): Add the new return block to the loop tree root.
12194 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
12195 whether we have removed the forwarder block.
12196 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
12197 * cfgloop.h (place_new_loop): Declare.
12198 * cfgloopmanip.c (place_new_loop): Export.
12199 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
12200 (tree-switch-conversion.o): Likewise.
12201 (tree-complex.o): Likewise.
12202 (tree-inline.o): Likewise.
12203 (tree-ssa-tailmerge.o): Likewise.
12204 (ipa-split.o): Likewise.
12205 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
12206 (tree-ssa-copy.o): Likewise.
12207 * tree-switch-conversion.c: Include cfgloop.h
12208 (process_switch): If we emit a bit-test cascade, schedule loops
12210 * tree-complex.c: Include cfgloop.h.
12211 (expand_complex_div_wide): Properly add new basic-blocks to loops.
12212 * asan.c: Include cfgloop.h.
12213 (create_cond_insert_point): Properly add new basic-blocks to
12214 loops, schedule loop fixup.
12215 * cfgloop.c (verify_loop_structure): Check that looks are not
12217 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
12219 (expand_omp_for_generic): Likewise.
12220 (expand_omp_sections): Likewise.
12221 (expand_omp_atomic_pipeline): Schedule loops for fixup.
12222 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
12223 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
12224 is initialized, not when loops are present.
12225 * tree-parloops.c (parallelize_loops): Remove checking here.
12226 * passes.c (init_optimization_passes): Schedule a copy-propagation
12227 pass before complete unrolling of inner loops.
12229 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12231 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
12232 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
12233 (colorize_init): Add argument to _WIN32 version.
12234 * toplev.c: Include diagnostic-color.h.
12235 (process_options): Default to -fdiagnostics-color=auto if
12236 GCC_COLORS env var is in the environment.
12237 * common.opt (fdiagnostics-color=): Add Var and Init.
12238 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
12239 env var is in the environment, the default is auto rather than never.
12241 * diagnostic.h (file_name_as_prefix): Add context argument.
12242 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
12243 the string as locus.
12244 * langhooks.c (lhd_print_error_function): Adjust caller.
12246 2013-04-25 Lawrence Crowl <crowl@google.com>
12248 * var-tracking.c (shared_hash_def::htab):
12249 Change type to hash_table. Update dependent calls and types.
12251 2013-04-25 Lawrence Crowl <crowl@google.com>
12253 * Makefile.in: Update as needed below.
12255 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
12256 Move declaration to after the type's method definitons.
12258 * attribs.c (htab_t scoped_attributes::attribute_hash):
12259 Change type to hash_table. Update dependent calls and types.
12261 * bitmap.c (htab_t bitmap_desc_hash):
12262 Change type to hash_table. Update dependent calls and types.
12264 * cselib.c (htab_t cselib_hash_table):
12265 Change type to hash_table. Update dependent calls and types.
12267 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
12268 (hash_string_slot_node): Move implementation into lto-streamer.h
12269 struct string_slot_hasher.
12270 (eq_string_slot_node): Likewise.
12272 * data-streamer-out.c: Update output_block::string_hash_table
12273 dependent calls and types.
12275 * dwarf2cfi.c (htab_t trace_index):
12276 Change type to hash_table. Update dependent calls and types.
12278 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
12279 Change type to hash_table. Update dependent calls and types.
12280 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
12281 (htab_t optimize_external_refs::map): Likewise.
12282 (htab_t output_comp_unit::extern_map): Likewise.
12283 (htab_t output_comdat_type_unit::extern_map): Likewise.
12284 (htab_t output_macinfo::macinfo_htab): Likewise.
12285 (htab_t optimize_location_lists::htab): Likewise.
12286 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
12288 * except.c (htab_t ehspec_hash_type):
12289 Change type to hash_table. Update dependent calls and types.
12290 (assign_filter_values::ttypes): Likewise.
12291 (assign_filter_values::ehspec): Likewise.
12292 (sjlj_assign_call_site_values::ar_hash): Likewise.
12293 (convert_to_eh_region_ranges::ar_hash): Likewise.
12295 * gcse.c (htab_t pre_ldst_table):
12296 Change type to hash_table. Update dependent calls and types.
12298 * ggc-common.c (htab_t saving_htab):
12299 Change type to hash_table. Update dependent calls and types.
12300 (htab_t loc_hash): Likewise.
12301 (htab_t ptr_hash): Likewise.
12302 (call_count): Rename ggc_call_count.
12303 (call_alloc): Rename ggc_call_alloc.
12304 (loc_descriptor): Rename make_loc_descriptor.
12305 (add_statistics): Rename ggc_add_statistics.
12307 * ggc-common.c (saving_htab):
12308 Change type to hash_table. Update dependent calls and types.
12310 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
12311 (push_gimplify_context): Likewise.
12312 (pop_gimplify_context): Likewise.
12313 (struct gimple_temp_hash_elt): Added.
12314 (struct gimplify_hasher): Likewise.
12315 (struct gimplify_ctx.temp_htab):
12316 Change type to hash_table. Update dependent calls and types.
12318 * gimple-fold.c: Include gimplify-ctx.h.
12320 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
12321 Change type to hash_table. Update dependent calls and types.
12322 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
12323 avoid potential global name collision.
12325 * gimplify.c: Include gimplify-ctx.h.
12326 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
12327 (htab_t gimplify_ctx::temp_htab):
12328 Update dependent calls and types for new type hash_table.
12329 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
12330 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
12332 * gimplify-ctx.h: New.
12333 (struct gimple_temp_hash_elt): Move from gimplify.c.
12334 (class gimplify_hasher): New.
12335 (struct gimplify_ctx): Move from gimple.h.
12336 (htab_t gimplify_ctx::temp_htab):
12337 Change type to hash_table. Update dependent calls and types.
12339 * graphite-clast-to-gimple.c: Include graphite-htab.h.
12340 (htab_t ivs_params::newivs_index):
12341 Change type to hash_table. Update dependent calls and types.
12342 (htab_t ivs_params::params_index): Likewise.
12343 (htab_t print_generated_program::params_index): Likewise.
12344 (htab_t gloog::newivs_index): Likewise.
12345 (htab_t gloog::params_index): Likewise.
12347 * graphite.c: Include graphite-htab.h.
12348 4htab_t graphite_transform_loops::bb_pbb_mapping):
12349 Change type to hash_table. Update dependent calls and types.
12351 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
12352 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
12353 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
12355 * graphite-dependences.c: Include graphite-htab.h.
12356 (loop_is_parallel_p): Change hash table type of parameter.
12358 * graphite-htab.h: New.
12359 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
12360 (extern find_pbb_via_hash): Move from graphite-poly.h.
12361 (extern loop_is_parallel_p): Move from graphite-poly.h.
12362 (extern get_loop_body_pbbs): Move from graphite-poly.h.
12364 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
12365 (extern loop_is_parallel_p): Move to graphite-htab.h.
12366 (extern get_loop_body_pbbs): Move to graphite-htab.h.
12368 * haifa-sched.c (htab_t delay_htab):
12369 Change type to hash_table. Update dependent calls and types.
12370 (htab_t delay_htab_i2): Likewise.
12372 * ira-color.c (htab_t allocno_hard_regs_htab):
12373 Change type to hash_table. Update dependent calls and types.
12375 * ira-costs.c (htab_t cost_classes_htab):
12376 Change type to hash_table. Update dependent calls and types.
12378 * loop-invariant.c (htab_t merge_identical_invariants::eq):
12379 Change type to hash_table. Update dependent calls and types.
12381 * loop-iv.c (htab_t bivs):
12382 Change type to hash_table. Update dependent calls and types.
12384 * loop-unroll.c (htab_t opt_info::insns_to_split):
12385 Change type to hash_table. Update dependent calls and types.
12386 (htab_t opt_info::insns_with_var_to_expand): Likewise.
12388 * lto-streamer.h (struct string_slot): Move from data-streamer.h
12389 (struct string_slot_hasher): New.
12390 (htab_t output_block::string_hash_table):
12391 Change type to hash_table. Update dependent calls and types.
12393 * lto-streamer-in.c (freeing_string_slot_hasher): New.
12394 (htab_t file_name_hash_table):
12395 Change type to hash_table. Update dependent calls and types.
12397 * lto-streamer-out.c: Update output_block::string_hash_table dependent
12400 * lto-streamer.c (htab_t tree_htab):
12401 Change type to hash_table. Update dependent calls and types.
12403 * omp-low.c: Include gimplify-ctx.h.
12405 * passes.c (htab_t name_to_pass_map):
12406 Change type to hash_table. Update dependent calls and types.
12407 (pass_traverse): Rename to passes_pass_traverse.
12409 * plugin.c (htab_t event_tab):
12410 Change type to hash_table. Update dependent calls and types.
12412 * postreload-gcse.c (htab_t expr_table):
12413 Change type to hash_table. Update dependent calls and types.
12414 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
12416 * sese.c (debug_rename_map_1): Make extern.
12417 (htab_t copy_bb_and_scalar_dependences::rename_map):
12418 Change type to hash_table. Update dependent calls and types.
12420 * sese.h (extern debug_rename_map): Move to .c file.
12422 * store-motion.c (htab_t store_motion_mems_table):
12423 Change type to hash_table. Update dependent calls and types.
12425 * trans-mem.c (htab_t tm_new_mem_hash):
12426 Change type to hash_table. Update dependent calls and types.
12428 * tree-browser.c (htab_t TB_up_ht):
12429 Change type to hash_table. Update dependent calls and types.
12431 * tree-cfg.c (htab_t discriminator_per_locus):
12432 Change type to hash_table. Update dependent calls and types.
12434 * tree-complex.c: Include tree-hasher.h
12435 (htab_t complex_variable_components):
12436 Change type to hash_table. Update dependent calls and types.
12438 * tree-eh.c (htab_t finally_tree):
12439 Change type to hash_table. Update dependent calls and types.
12441 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
12442 struct int_tree_hasher.
12443 (extern int_tree_map_eq): Likewise.
12444 (uid_decl_map_hash): Removed.
12445 (extern decl_tree_map_eq): Likewise.
12447 * tree-hasher.h: New.
12448 (struct int_tree_hasher): New.
12449 (typedef int_tree_htab_type): New.
12451 * tree-inline.c: Include gimplify-ctx.h.
12453 * tree-mudflap.c: Include gimplify-ctx.h.
12455 * tree-parloops.c: Include tree-hasher.h.
12456 (htab_t eliminate_local_variables_stmt::decl_address):
12457 Change type to hash_table. Update dependent calls and types.
12458 (htab_t separate_decls_in_region::decl_copies): Likewise.
12460 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
12461 Change type to hash_table. Update dependent calls and types.
12463 * tree-sra.c (candidates):
12464 Change type to hash_table. Update dependent calls and types.
12466 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
12468 (int_tree_map_hash): Likewise.
12470 * tree-ssa-dom.c (htab_t avail_exprs):
12471 Change type to hash_table. Update dependent calls and types.
12473 * tree-ssa-live.c (var_map_base_init::tree_to_index):
12474 Change type to hash_table. Update dependent calls and types.
12476 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
12477 Change type to hash_table. Update dependent calls and types.
12479 * tree-ssa-phiopt.c (seen_ssa_names):
12480 Change type to hash_table. Update dependent calls and types.
12482 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
12483 Change type to hash_table. Update dependent calls and types.
12485 * tree-ssa-uncprop.c (equiv):
12486 Change type to hash_table. Update dependent calls and types.
12488 2013-04-25 Jakub Jelinek <jakub@redhat.com>
12490 PR rtl-optimization/57003
12491 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
12492 call note_stores with kill_clobbered_value callback again after
12493 killing regs_invalidated_by_call.
12495 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12497 * config/aarch64/aarch64-simd.md
12498 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
12499 (aarch64_simd_bsl<mode>): Likewise.
12500 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
12502 2013-04-25 Marek Polacek <polacek@redhat.com>
12504 PR tree-optimization/57066
12505 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
12507 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
12509 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
12511 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12513 * config/aarch64/aarch64-builtins.c
12514 (aarch64_fold_builtin): New.
12515 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
12516 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
12517 * config/aarch64/aarch64-simd-builtins.def (abs): New.
12518 * config/aarch64/arm_neon.h
12519 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
12521 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12522 Tejas Belagod <tejas.belagod@arm.com>
12524 * config/aarch64/aarch64-builtins.c
12525 (aarch64_gimple_fold_builtin): New.
12526 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
12527 * config/aarch64/aarch64-simd-builtins.def (addv): New.
12528 * config/aarch64/aarch64-simd.md (addpv4sf): New.
12529 (addvv4sf): Update.
12530 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
12532 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12534 * config/aarch64/aarch64.md
12535 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
12537 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12539 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
12540 (*ngcsi_uxtw): New pattern.
12542 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12543 Julian Brown <julian@codesourcery.com>
12545 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
12546 (TB_DREG): Add T_V4HF.
12547 (v4hf_UP): New macro.
12548 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12549 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12550 Handle initialisation of V4HF. Adjust initialisation of reinterpret
12552 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12553 (arm_vector_mode_supported_p): Handle V4HF.
12554 (arm_mangle_map): Handle V4HFmode.
12555 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
12556 * config/arm/arm_neon_builtins.def: Add entries for
12557 vcvtv4hfv4sf, vcvtv4sfv4hf.
12558 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
12559 (neon_vcvtv4hfv4sf): Likewise.
12560 * config/arm/neon-gen.ml: Handle half-precision floating point
12562 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
12563 * config/arm/arm_neon.h: Regenerate.
12564 * config/arm/neon.ml (type elts): Add F16.
12565 (type vectype): Add T_float16x4, T_floatHF.
12566 (type vecmode): Add V4HF.
12567 (type features): Add Requires_FP_bit feature.
12568 (elt_width): Handle F16.
12569 (elt_class): Likewise.
12570 (elt_of_class_width): Likewise.
12571 (mode_of_elt): Refactor.
12572 (type_for_elt): Handle F16, fix error messages.
12573 (vectype_size): Handle T_float16x4.
12574 (vcvt_sh): New function.
12575 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
12576 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
12577 (string_of_mode): Handle V4HF.
12578 * doc/arm-neon-intrinsics.texi: Regenerate.
12580 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12582 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
12583 format specifier in 'X' case.
12585 2013-04-25 Alan Modra <amodra@gmail.com>
12588 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
12589 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
12590 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
12591 Repeat for many other rotate/shift and mask patterns using subregs.
12592 Name lshiftrt insns.
12593 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
12594 on WORDS_BIG_ENDIAN.
12596 2013-04-25 Alan Modra <amodra@gmail.com>
12598 * config.gcc: Support little-endian powerpc-linux targets.
12599 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
12600 (LINK_OS_LINUX_SPEC): Define.
12601 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
12602 Preserve MASK_LITTLE_ENDIAN.
12603 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
12604 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
12605 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
12606 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
12607 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
12608 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
12609 Correct fp word order for little-endian. Don't shift toc entries
12610 smaller than a word for little-endian.
12611 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
12612 (bswapdi2 splits): Correct low-part subreg for little-endian.
12613 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
12614 low/high where such is correct only for be.
12615 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
12616 little-endian for -mcall-aixdesc.
12618 2013-04-25 Alan Modra <amodra@gmail.com>
12620 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
12621 replace_equiv_address_nv.
12623 2013-04-25 Alan Modra <amodra@gmail.com>
12625 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
12627 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12630 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12631 * rtl.h (struct rtx_def): ...
12633 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12635 PR rtl-optimizations/57046
12636 * lra-constraints (split_reg): Set up lra_risky_transformations_p
12637 for multi-reg splits.
12639 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
12641 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
12643 2013-04-24 Sterling Augustine <saugustine@google.com>
12645 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
12646 (comp_dir_string, debug_str_dwo_section): New.
12647 (DEBUG_STR_DWO_SECTION): Rename to ...
12648 (DEBUG_DWO_STR_SECTION): ... this.
12649 (DEBUG_NORM_STR_SECTION): Delete.
12650 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
12651 (DEBUG_STR_DWO_SECTION_FLAGS): New.
12652 (find_AT_string): Move most logic to ...
12653 (find_AT_string_in_table): ... here. New.
12654 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
12655 add_skeleton_AT_string. Delete logic.
12656 (output_skeleton_debug_sections): Remove call to
12657 add_top_level_skeleton_die_attrs.
12658 (add_comp_dir_attribute): Move logic to comp_dir_string.
12659 (dwarf2out_init): Initialize debug_str_dwo_section.
12660 (output_indirect_string): Call find_string_form.
12661 (output_indirect_strings): Rewrite.
12662 (prune_unused_types): Empty skeleton_debug_str_hash.
12663 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12664 (dwarf2out_finish): Call output_indirect_strings.
12666 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12668 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12670 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12672 * rtl.h (struct rtx_def): Add comment for field jump.
12673 (LRA_SUBREG_P): New macro.
12674 * recog.c (register_operand): Check LRA_SUBREG_P.
12675 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12677 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12678 (lra_final_code_change): Skip subreg change for operators.
12679 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12680 if there are no operand changes.
12681 * lra-constraints.c (curr_insn_set): New.
12682 (match_reload): Set LRA_SUBREG_P.
12683 (emit_spill_move): Ditto.
12684 (check_and_process_move): Use curr_insn_set. Process only single
12685 set insns. Don't initialize sec_mem_p and change_p.
12686 (simplify_operand_subreg): Use LRA_SUBREG_P.
12687 (reg_in_class_p): New function.
12688 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12689 of #ifdef. Add code to remove cycling.
12690 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12691 non-null disp. Reload inner instead of disp when base and index
12692 are null. Try to put lo_sum into register.
12693 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12694 (check_and_process_move): Move code for move cost check to
12695 simple_move_p. Remove equiv_substitution.
12696 (simple_move_p): New function.
12697 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12698 curr_insn_set. Call check_and_process_move only for single set
12699 insns. Use the new function. Move call of check_and_process_move
12700 after operand equiv substitution and address process.
12702 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12704 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12705 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12706 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12708 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12710 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12712 2013-04-24 Marek Polacek <polacek@redhat.com>
12714 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12715 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12716 (select_loops_exit_conditions): Likewise.
12717 (number_of_iterations_for_all_loops): Likewise.
12718 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12719 (scev_analysis): Likewise.
12721 2013-04-02 Catherine Moore <clm@codesourcery.com>
12722 Chao-ying Fu <fu@mips.com>
12724 * config/mips/micromips.md (jraddiusp): New pattern.
12725 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12726 instruction if possible.
12728 2013-04-24 Alan Modra <amodra@gmail.com>
12730 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12732 2013-04-24 Julian Brown <julian@codesourcery.com>
12733 Chung-Lin Tang <cltang@codesourcery.com>
12735 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12736 dependency behavior in enumeration type DIE generation. Add TODO note
12737 to comments about future DW_FORM_sdata/udata re-work of related code.
12739 2013-04-23 Lawrence Crowl <crowl@google.com>
12741 * Makefile.in: Update as needed below.
12743 * hash-table.h (class hash_table):
12744 Correct many methods with parameter types compare_type to the correct
12745 value_type. (Correct code was unlikely to notice the change.)
12746 (hash_table::elements_with_deleted) New.
12747 (class hashtable::iterator): New.
12748 (hashtable::begin()): New.
12749 (hashtable::end()): New.
12750 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12752 * statistics.c (statistics_hashes):
12753 Change type to hash_table. Update dependent calls and types.
12755 * tree-into-ssa.c (var_infos):
12756 Change type to hash_table. Update dependent calls and types.
12758 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12759 Change type to hash_table. Update dependent calls and types.
12761 * tree-ssa-loop-im.c (struct mem_ref.refs):
12762 Change type to hash_table. Update dependent calls and types.
12764 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12765 Change type to hash_table. Update dependent calls and types.
12767 * tree-ssa-sccvn.c (vn_tables_s::nary):
12768 Change type to hash_table. Update dependent calls and types.
12769 (vn_tables_s::phis): Likewise.
12770 (vn_tables_s::references): Likewise.
12772 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12773 (vn_reference_eq): Update parameter and return types.
12775 * tree-ssa-structalias.c (pointer_equiv_class_table):
12776 Change type to hash_table. Update dependent calls and types.
12777 (location_equiv_class_table): Likewise.
12779 * tree-vect-data-refs.c: Consequential changes for making
12780 peeling a hash_table.
12782 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12783 (destroy_loop_vec_info): Dependent hash_table update.
12785 * tree-vectorizer.h (peeling_htab):
12786 Change type to hash_table. Update dependent calls and types.
12788 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12790 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12791 to check the register content is equal or not.
12792 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12793 to assign register content record.
12794 * lra-eliminations.c (update_reg_eliminate): Use
12795 lra_update_reg_val_offset to update register content offset.
12796 * lra-int.h (struct lra_reg): Add offset member.
12797 (lra_reg_val_equal_p): New static inline function.
12798 (lra_update_reg_val_offset): New static inline function.
12799 (lra_assign_reg_val): New static inline function.
12800 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12801 to assign register content record.
12802 (initialize_lra_reg_info_element): Initial offset to zero.
12804 2013-04-23 Catherine Moore <clm@codesourcery.com>
12806 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12807 operands. Record compression.
12809 2013-04-23 Xinliang David Li <davidxl@google.com>
12811 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12813 2013-04-23 Richard Biener <rguenther@suse.de>
12815 PR middle-end/57036
12816 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12817 parameter, only add abnormal goto edges from the copied body
12818 if the call could perform abnormal gotos.
12819 (copy_cfg_body): Adjust.
12821 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12823 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12825 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12827 * coretypes.h (gimple_stmt_iterator): Add struct to make
12830 2013-04-23 Richard Biener <rguenther@suse.de>
12832 PR tree-optimization/57026
12833 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12834 from SSA names occuring in abnormal PHI nodes.
12836 2013-04-22 Andi Kleen <ak@linux.intel.com>
12838 * lto/lto.c (print_lto_report_1): Fix LTO report names.
12840 2013-04-22 Andi Kleen <ak@linux.intel.com>
12842 * lto/lto.c (print_lto_report_1): Declare early.
12843 (read_cgraph_and_symbols): Call print_lto_report_1 early.
12845 2013-04-22 Andi Kleen <ak@linux.intel.com>
12847 * common.opt (-flto-report-wpa): Add.
12848 * doc/invoke.texi (-flto-report-wpa): Add.
12849 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
12852 2013-04-22 Xinliang David Li <davidxl@google.com>
12854 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
12855 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
12856 * Makefile.in: New dependency
12858 David Daney <ddaney.cavm@gmail.com>
12860 * configure.ac (gcc_cv_as_micromips_support): Use the
12861 --fatal-warnings option.
12862 * configure: Regenerate.
12864 2013-04-22 Marek Polacek <polacek@redhat.com>
12867 * tsan.c (instrument_expr): Don't instrument expression
12868 in case its size is zero.
12870 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
12874 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
12876 * config/alpha/alpha.c (TARGET_LRA_P): New define.
12878 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12880 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
12881 (gimple_stmt_iterator): New typedef.
12882 * gimple.h (gimple_stmt_iterator): Rename to...
12883 (gimple_stmt_iterator_d): ... This.
12884 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
12885 trees be valid for GIMPLE and GENERIC.
12886 (TARGET_GIMPLE_FOLD_BUILTIN): New.
12887 * gimple-fold.c (gimple_fold_call): Call target hook
12888 gimple_fold_builtin.
12889 * hooks.c (hook_bool_gsiptr_false): New.
12890 * hooks.h (hook_bool_gsiptr_false): New.
12891 * target.def (fold_stmt): New.
12892 * doc/tm.texi: Regenerate.
12894 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
12897 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
12898 a set sp if no stack realignment.
12900 2013-04-22 Nick Clifton <nickc@redhat.com>
12902 * config.gcc (tilegx-linux): Extend extra_objs rather than
12904 (tilepro-linux): Likewise.
12906 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12908 * config/aarch64/aarch64-builtins.c
12910 (CF0, CF1, CF2, CF3, CF4, CF10): New.
12911 (VAR<1-12>): Add MAP parameter.
12912 (BUILTIN_*): Likewise.
12913 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
12914 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
12915 (aarch64_ushl_n<mode>): Likewise.
12916 (aarch64_sshr_n<mode>): Likewise.
12917 (aarch64_ushr_n<mode>): Likewise.
12918 (aarch64_<maxmin><mode>): Likewise.
12919 (aarch64_sqrt<mode>): Likewise.
12920 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
12921 (vshr<q>_n_*): Likewise.
12923 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12925 * config/aarch64/aarch64-builtins.c
12926 (aarch64_simd_builtin_type_mode): Handle SF types.
12928 (BUILTIN_GPF): Define.
12929 (aarch64_init_simd_builtins): Handle SF types.
12930 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
12931 (frecps): Likewise.
12932 (frecpx): Likewise.
12933 * config/aarch64/aarch64-simd.md
12934 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
12935 (aarch64_frecpe<mode>): New.
12936 (aarch64_frecps<mode>): Likewise.
12937 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
12938 (v8type): Add frecp<esx>.
12939 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
12940 (aarch64_frecps<mode>): Likewise.
12941 * config/aarch64/iterators.md (FRECP): New.
12942 (frecp_suffix): Likewise.
12943 * config/aarch64/arm_neon.h
12944 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
12946 2013-04-22 Christian Bruel <christian.bruel@st.com>
12949 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
12950 (REG_CLASS_NAMES): Idem.
12951 (REG_CLASS_CONTENTS): Idem.
12952 (REGCLASS_HAS_FP_REG): Idem.
12953 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
12954 (sh_conditional_register_usage): Idem.
12956 2013-04-21 Jeff Law <law@redhat.com>
12958 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
12959 (ssa_forward_propagate_and_combine): Use it.
12961 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12963 * lra.c: Update the flow chart diagram.
12965 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12967 PR rtl-optimization/56847
12968 * lra-constraints.c (process_alt_operands): Discourage alternative
12969 with non-matche doffsettable memory constraint fro memory with
12972 2013-04-19 Richard Biener <rguenther@suse.de>
12974 PR tree-optimization/56982
12975 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
12977 * gimplify.c (gimplify_call_expr): Notice special calls.
12978 (gimplify_modify_expr): Likewise.
12979 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
12980 abnormal control flow receivers.
12981 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
12982 in the same way as cfun->has_nonlocal_labels.
12983 (gimple_purge_dead_abnormal_call_edges): Likewise.
12984 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
12985 receivers start a basic-block.
12987 2013-04-19 Richard Biener <rguenther@suse.de>
12989 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
12991 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
12992 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
12993 (SLP_TREE_LOAD_PERMUTATION): Add.
12994 (vect_transform_slp_perm_load): Adjust prototype.
12995 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
12996 (vect_free_slp_instance): Likewise.
12997 (vect_create_new_slp_node): Likewise.
12998 (vect_supported_slp_permutation_p): Remove.
12999 (vect_slp_rearrange_stmts): Adjust.
13000 (vect_supported_load_permutation_p): Likewise. Inline
13001 vect_supported_slp_permutation_p here.
13002 (vect_analyze_slp_instance): Compute load permutations per
13003 slp node instead of per instance.
13004 (vect_get_slp_defs): Adjust.
13005 (vect_transform_slp_perm_load): Likewise.
13006 (vect_schedule_slp_instance): Remove redundant code.
13007 (vect_schedule_slp): Remove hack for PR56270, add it ...
13008 * tree-vect-stmts.c (vectorizable_load): ... here, do not
13009 CSE loads for SLP. Adjust.
13011 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13013 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
13014 spelling in two comments.
13016 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13019 * config/arm/arm.c (load_multiple_sequence): Require SP
13020 as base register for loads if SP is in the register list.
13022 2013-04-19 Martin Jambor <mjambor@suse.cz>
13024 PR tree-optimization/56718
13025 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
13026 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
13027 and made public. Adjusted all callers.
13028 (ipa_intraprocedural_devirtualization): New function.
13029 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
13030 (ipa_intraprocedural_devirtualization): Likewise.
13031 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
13033 2013-04-19 Richard Biener <rguenther@suse.de>
13035 PR tree-optimization/57000
13036 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
13038 2013-04-19 Terry Guo <terry.guo@arm.com>
13040 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
13042 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
13043 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
13044 (cortex_m4_fmacs): Use new reservations.
13045 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
13047 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
13049 PR rtl-optimization/56999
13050 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
13052 (lra_coalesce): Remove split_origin_bitmap and related code.
13053 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
13054 ranges if necessary.
13056 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
13058 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
13060 (ix86_expand_call): Remove clobbered_registers array and use
13061 x86_64_ms_sysv_extra_clobbered_registers instead.
13062 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
13064 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
13066 * config/i386/i386.md (*call_rex64_ms_sysv): Use
13067 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
13068 (*call_value_rex64_ms_sysv): Ditto.
13070 2013-04-18 Cary Coutant <ccoutant@google.com>
13072 * dwarf2out.c (output_pubnames): Check die_perennial_p of
13073 parent instead of die_mark.
13075 2013-04-18 Diego Novillo <dnovillo@google.com>
13077 * gimple.c (create_gimple_tmp): New.
13078 (get_expr_type): New.
13079 (build_assign): New.
13080 (build_type_cast): New.
13081 * gimple.h (enum ssa_mode): Define.
13082 (gimple_seq_set_location): New.
13083 * asan.c (build_check_stmt): Change some gimple_build_* calls
13084 to use build_assign and build_type_cast.
13086 2013-04-18 Richard Biener <rguenther@suse.de>
13088 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
13089 handle negative step. Remove redundant checks.
13090 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
13091 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
13092 for negative step and grouped loads fail to vectorize.
13094 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13096 * emit-rtl.c (reset_insn_used_flags): New function.
13097 (reset_all_used_flags): Use it.
13098 (verify_insn_sharing): New function.
13099 (verify_rtl_sharing): Fix verification for SEQUENCEs.
13101 2013-04-18 Jakub Jelinek <jakub@redhat.com>
13103 PR tree-optimization/56984
13104 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
13105 and (x >> M) >= N don't register any assertion if N << M is the
13108 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13110 * lower-subreg.c (resolve_simple_move): If called self-recursive,
13111 do not delete_insn insns that have not yet been emitted, only
13112 unlink them with remove_insn.
13113 * df-scan.c (df_insn_delete): Revert r197492.
13115 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
13117 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
13118 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
13120 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13122 * config/arm/arm.md (movsicc_insn): Convert define_insn into
13123 define_insn_and_split.
13124 (and_scc,ior_scc,negscc): Likewise.
13125 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
13127 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13129 * config/arm/arm.c (use_return_insn): Return 0 for targets that
13130 can benefit from using a sequence of LDRD instructions in epilogue
13131 instead of a single LDM instruction.
13133 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
13136 * doc/extend.texi: Fix typo.
13138 2013-04-17 Richard Biener <rguenther@suse.de>
13140 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
13141 (vect_build_slp_tree): ... here.
13142 (vect_build_slp_tree_1): Compute which stmts of the SLP group
13143 match. Remove special-casing of mismatched complex loads.
13144 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
13145 re-try the match with swapped commutative operands.
13146 (vect_supported_load_permutation_p): Remove special-casing of
13147 mismatched complex loads.
13148 (vect_analyze_slp_instance): Adjust.
13150 2013-04-17 Richard Biener <rguenther@suse.de>
13152 PR rtl-optimization/56921
13153 * cfgloop.h (struct loop): Add simple_loop_desc member.
13154 (struct niter_desc): Mark with GTY(()).
13155 (simple_loop_desc): Do not use aux field but simple_loop_desc.
13156 * loop-iv.c (get_simple_loop_desc): Likewise.
13157 (free_simple_loop_desc): Likewise.
13160 2013-04-16 Richard Biener <rguenther@suse.de>
13162 PR rtl-optimization/56921
13163 * loop-init.c (pass_rtl_move_loop_invariants): Add
13164 TODO_do_not_ggc_collect to todo_flags_finish.
13165 (pass_rtl_unswitch): Same.
13166 (pass_rtl_unroll_and_peel_loops): Same.
13167 (pass_rtl_doloop): Same.
13169 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
13171 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
13172 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
13173 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
13175 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
13176 * tree-streamer.c (record_common_node): Adjust reference in comment.
13178 2013-04-17 Terry Guo <terry.guo@arm.com>
13180 * config/arm/cortex-m4.md: Add a new bypass.
13182 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13184 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
13186 (*subs_<optab><mode>_multp2): New pattern.
13187 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13188 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13190 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13192 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
13193 (*subs_mul_imm_<mode>): New pattern.
13195 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
13198 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
13199 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
13200 (vsx_movti_32bit): Change j->wa to O->wa.
13202 2013-04-16 Richard Biener <rguenther@suse.de>
13204 PR rtl-optimization/56921
13205 * loop-init.c (pass_rtl_move_loop_invariants): Add
13206 TODO_do_not_ggc_collect to todo_flags_finish.
13207 (pass_rtl_unswitch): Same.
13208 (pass_rtl_unroll_and_peel_loops): Same.
13209 (pass_rtl_doloop): Same.
13211 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
13213 * config/arm/arm.c (emit_multi_reg_push): New declaration
13214 for an existing function.
13215 (arm_emit_strd_push): New function.
13216 (arm_expand_prologue): Used here.
13217 (arm_emit_ldrd_pop): New function.
13218 (arm_expand_epilogue): Used here.
13219 (arm_get_frame_offsets): Update condition.
13220 (arm_emit_multi_reg_pop): Add a special case for load of a single
13221 register with writeback.
13223 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
13225 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
13228 2013-04-16 Richard Biener <rguenther@suse.de>
13230 PR tree-optimization/56756
13231 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
13232 (first_mem_ref_loc): New.
13233 (execute_sm): Place the load temporarily before a previous
13234 access instead of in the latch edge to ensure its SSA dependencies
13235 are defined at points dominating the load.
13237 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
13239 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
13240 correct fix by moving header and footer insn to the footer of
13241 the merged basic block. Clear BB_END of the merged-away block.
13243 PR middle-end/43631
13244 * emit-rtl.c (make_note_raw): New function.
13245 (link_insn_into_chain): New static inline function.
13246 (add_insn): Use it.
13247 (add_insn_before, add_insn_after): Factor insn chain linking code...
13248 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
13249 using link_insn_into_chain.
13250 (note_outside_basic_block_p): New helper function for emit_note_after
13251 and emit_note_before.
13252 (emit_note_after): Use nobb variant of add_insn_after if the note
13253 should not be contained in a basic block.
13254 (emit_note_before): Use nobb variant of add_insn_before if the note
13255 should not be contained in a basic block.
13256 (emit_note_copy): Use make_note_raw.
13257 (emit_note): Likewise.
13258 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
13259 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
13260 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
13261 the moved barrier the tail of the basic block it follows.
13262 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
13264 2013-04-15 Jakub Jelinek <jakub@redhat.com>
13266 PR tree-optimization/56962
13267 * gimple-ssa-strength-reduction.c (record_increment): Only set
13268 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
13269 either rhs1 or rhs2 is equal to c->base_expr.
13271 2013-04-15 Richard Biener <rguenther@suse.de>
13273 PR tree-optimization/56933
13274 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
13276 (GROUP_READ_WRITE_DEPENDENCE): Remove.
13277 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
13278 * tree-vect-data-refs.c (vect_analyze_group_access): Move
13279 dependence check ...
13280 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
13282 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
13283 GROUP_READ_WRITE_DEPENDENCE.
13285 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13287 * emit-rtl.c (reset_all_used_flags): New function.
13288 (verify_rtl_sharing): Call reset_all_used_flags before and after
13289 performing the checks.
13291 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13293 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
13294 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
13295 * config/arm/constraints.md (De): New constraint.
13296 * config/arm/neon.md (anddi3_neon): Delete.
13297 (neon_vand<mode>): Expand to standard anddi3 pattern.
13298 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
13299 Move earlier in the file.
13300 (neon_inv_logic_op2): Likewise.
13301 (arm_anddi_operand_neon): New predicate.
13303 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13305 * configure.ac (gcc_cv_ld_as_needed): Set
13306 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
13307 Use -z ignore, -z record on *-*-solaris2*.
13308 (HAVE_LD_AS_NEEDED): Update comment.
13309 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
13310 * configure: Regenerate.
13311 * config.in: Regenerate.
13312 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
13313 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
13314 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
13315 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
13316 equivalents. Fix markup.
13317 * doc/tm.texi: Regenerate.
13319 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
13321 * config/i386/i386.opt: New option mstack-protector-guard=.
13322 * config/i386/i386-opts.h: Add enum stack_protector_guard.
13323 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
13324 TARGET_SSP_TLS_GUARD.
13325 * config/i386/i386.c (ix86_option_override_internal): Set
13326 ix86_stack_protector_guard.
13327 * config/i386/i386.md (stack_protect_set): Enable for
13328 TARGET_SSP_TLS_GUARD only.
13329 (stack_protect_set_<mode>): Ditto.
13330 (stack_protect_test): Ditto.
13331 (stack_protect_test_<mode>): Ditto.
13332 * doc/invoke.texi (i386 Option): Document.
13334 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
13337 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
13338 (S_MODES): Set H_MODE bit.
13339 (SF_MODES): Set only S_MODE and SF_MODE bits.
13340 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
13341 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
13342 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
13343 <MODE_FLOAT>: Likewise.
13345 2013-04-15 Joey Ye <joey.ye@arm.com>
13347 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
13349 2013-04-15 Joey Ye <joey.ye@arm.com>
13351 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
13353 (thumb_far_jump_used_p): Count instruction size and set
13356 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
13358 * reorg.c (fill_simple_delay_slots): Reindent block of code.
13359 * resource.c (mark_target_live_regs): Reformat conditional block.
13361 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13363 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
13364 notes, they are emitted only just before final.
13365 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
13367 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13369 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
13370 * cfgrtl.c (delete_insn): Call it here instead.
13371 * lra-spills.c (lra_final_code_change): Use delete_insn.
13372 * haifa-sched.c (sched_remove_insn): Likewise.
13373 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
13374 returning to the nop pool.
13375 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
13376 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
13378 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13380 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
13381 * doc/tm.texi: Regenerated.
13383 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
13385 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
13388 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13390 * df-core.c (df_find_def): Compare register numbers.
13391 (df_find_use): Likewise.
13393 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
13396 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
13397 lra_in_progress for return.
13399 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13401 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
13402 define_insn into define_insn_and_split and emit movsicc patterns.
13404 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13406 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
13408 2013-04-12 Richard Biener <rguenther@suse.de>
13410 * tree-pass.h (TODO_do_not_ggc_collect): New.
13411 * passes.c (execute_one_ipa_transform_pass): Honor
13412 TODO_do_not_ggc_collect.
13413 (execute_one_pass): Likewise.
13416 2013-04-10 Richard Biener <rguenther@suse.de>
13418 * passes.c (init_optimization_passes): Remove reload pass.
13419 * ira.c (do_reload): Merge into ...
13421 (rest_of_handle_reload): Remove.
13422 (pass_reload): Likewise.
13423 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13424 of reload for vzeroupper pass placement.
13426 2013-04-12 Jakub Jelinek <jakub@redhat.com>
13428 PR tree-optimization/56918
13429 PR tree-optimization/56920
13430 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
13431 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
13432 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
13433 use wide_mul_with_sign method.
13435 2013-04-12 Richard Biener <rguenther@suse.de>
13437 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
13438 not be considered a gimple constant.
13440 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
13442 * fold-const.c (const_binop): Handle vector shifts by a scalar.
13443 (fold_binary_loc): Call const_binop also for mixed vector-scalar
13446 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
13447 Jakub Jelinek <jakub@redhat.com>
13449 * opts.c: Include diagnostic-color.h.
13450 (common_handle_option): Handle OPT_fdiagnostics_color_.
13451 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
13452 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
13453 (diagnostic-color.o): New.
13454 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
13455 (diagnostic_color_rule): New enum.
13456 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
13457 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
13458 the location string.
13459 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
13460 either NULL, or color kind.
13461 * diagnostic-color.c: New file.
13462 * diagnostic-color.h: New file.
13463 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
13465 * doc/invoke.texi (-fdiagnostics-color): Document.
13466 * pretty-print.h (pp_show_color): Define.
13467 (struct pretty_print_info): Add show_color field.
13468 * diagnostic.c: Include diagnostic-color.h.
13469 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
13470 macros. Colorize error:, warning: etc. strings and also the location
13472 (diagnostic_show_locus): Colorize the caret line.
13473 * pretty-print.c: Include diagnostic-color.h.
13474 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
13475 inside of %< %> quotes or quoted through q format modifier.
13477 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13479 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
13481 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13483 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
13484 code in CC_NZ mode.
13485 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
13488 2013-04-11 Marek Polacek <polacek@redhat.com>
13490 PR tree-optimization/48184
13491 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
13493 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
13495 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
13496 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
13497 (skip_simple_arithmetic): Tidy up.
13498 * tree.h (skip_simple_constant_arithmetic): Declare.
13500 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13502 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
13504 2013-04-11 Richard Biener <rguenther@suse.de>
13506 * tree-vect-loop.c (get_initial_def_for_induction): Properly
13507 generate vector constants.
13509 2013-04-11 Richard Biener <rguenther@suse.de>
13511 PR tree-optimization/56878
13512 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
13513 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
13515 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
13516 Prefer to align the DR with the most invariant base address.
13518 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
13520 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
13523 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
13525 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
13526 floating-point vector comparisons against 0.
13528 2013-04-11 Jakub Jelinek <jakub@redhat.com>
13530 PR tree-optimization/56899
13531 * fold-const.c (extract_muldiv_1): Apply distributive law
13532 only if TYPE_OVERFLOW_WRAPS (ctype).
13534 2013-04-11 Bin Cheng <bin.cheng@arm.com>
13537 * ira-costs.c (scan_one_insn): Check whether the source rtx of
13538 loading has side effect.
13540 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
13542 * config/sparc/sparc.c: Include tree-pass.h.
13543 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
13544 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
13545 head of file. Change return type. Split off gate function.
13546 (sparc_gate_work_around_errata): New function.
13547 (pass_work_around_errata): New pass definition.
13548 (insert_pass_work_around_errata) New pass insert definition to
13549 insert pass_work_around_errata just after delayed-branch scheduling.
13550 (sparc_option_override): Insert the pass.
13551 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
13553 2013-04-10 David S. Miller <davem@davemloft.net>
13555 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
13556 or -mcpu=hypersparc.
13558 * target.def (cstore_mode): New hook.
13559 * target.h: Include insn-codes.h
13560 * targhooks.c: Likewise.
13561 (default_cstore_mode): New function.
13562 * targhooks.h: Declare it.
13563 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
13564 * doc/tm.texi: Rebuild.
13565 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
13566 target hook, rather than inspecting the insn_data.
13567 * config/sparc/sparc.c (sparc_cstore_mode): New function.
13568 (TARGET_CSTORE_MODE): Redefine.
13569 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
13571 * config/sparc/predicates.md (cstore_result_operand): New special
13573 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
13574 Use it for operand 0.
13575 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
13576 (*snesi_special): Likewise.
13577 (*snesi_zero): Likewise.
13578 (*seqsi_zero): Likewise.
13579 (*sltu_insn): Likewise.
13580 (*sgeu_insn): Likewise.
13581 (*seqdi_special): Make operand 0 and comparison operation be of
13583 (*snedi_special): Likewise.
13584 (*snedi_special_vis3): Likewise.
13585 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
13586 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
13587 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
13588 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
13589 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
13590 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
13591 (*sltu_extend_sp64): Likewise.
13592 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
13593 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
13594 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
13595 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
13596 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
13598 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
13600 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
13601 (aarch64_start_file): Use the new function.
13603 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
13604 Jason Merrill <jason@redhat.com>
13606 * common.opt: Add -gdwarf.
13607 * opts.c (common_handle_option): Handle it.
13608 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
13610 2013-04-10 Richard Biener <rguenther@suse.de>
13612 * passes.c (execute_todo): Do not call ggc_collect conditional here.
13613 (execute_one_ipa_transform_pass): But unconditionally here.
13614 (execute_one_pass): And here.
13615 (init_optimization_passes): Remove reload pass.
13616 * tree-pass.h (TODO_ggc_collect): Remove.
13617 (pass_reload): Likewise.
13618 * ira.c (do_reload): Merge into ...
13620 (rest_of_handle_reload): Remove.
13621 (pass_reload): Likewise.
13622 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13623 of reload for vzeroupper pass placement.
13624 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
13625 and todo_flags_finish of all passes.
13627 2013-04-10 Richard Biener <rguenther@suse.de>
13629 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
13630 first_const_oprnd field, rename first_def_type to first_op_type.
13631 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
13632 (vect_get_and_check_slp_defs): Always use the type of the
13633 operand. Allow mixed vect_external_def, vect_constant_def types.
13634 (vect_get_constant_vectors): Handle mixed vect_external_def,
13635 vect_constant_def types.
13637 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
13639 PR tree-optimization/55524
13640 * tree-ssa-math-opts.c
13641 (convert_mult_to_fma): Don't use an fms construct
13642 when we don't have an fms operation, but fnma, and it looks
13643 likely that we'll be able to use the latter.
13645 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13647 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
13649 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
13650 inline fail caused by overwritable functions.
13652 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
13654 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
13655 unnecessary bits in the constant power of two case.
13657 2013-04-10 Richard Biener <rguenther@suse.de>
13659 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13660 broken code swapping operands.
13661 (vect_build_slp_tree): Do not compute load permutations here.
13662 (vect_analyze_slp_instance): Compute load permutations here,
13663 after building the SLP tree.
13665 2013-04-09 Christian Bruel <christian.bruel@st.com>
13667 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13668 of next/prev_real_insn.
13670 2013-04-09 Jan Hubicka <jh@suse.cz>
13672 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13673 Drop aliased parameter.
13674 (function_and_variable_visibility): Do not handle alias pairs.
13675 * cgraph.c (varpool_externally_visible_p): Update prototype.
13676 * varpool.c (varpool_add_new_variable): Update.
13678 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13680 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13682 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13684 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13686 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13688 2013-04-09 Marek Polacek <polacek@redhat.com>
13690 PR tree-optimization/48762
13691 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13693 2013-04-09 Richard Biener <rguenther@suse.de>
13695 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13697 (vect_build_slp_tree): Likewise.
13698 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13699 calculating the cost of a SLP instance.
13700 (vect_analyze_slp_instance): Use it from here, after building
13703 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13705 PR middle-end/56883
13706 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13707 expand_omp_for_static_chunk): Use simple_p = true in
13708 force_gimple_operand_gsi calls when assigning to addressable decls.
13710 2013-04-09 Jeff Law <law@redhat.com>
13712 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13713 when the boolean was created by converting a wider object which
13714 had a boolean range.
13716 2013-04-09 Richard Biener <rguenther@suse.de>
13718 * tree-vectorizer.h (slp_void_p): Remove.
13719 (slp_tree): Typedef before _slp_tree declaration.
13720 (struct _slp_tree): Use a vector of slp_tree as children.
13721 (vect_get_place_in_interleaving_chain): Remove.
13722 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13724 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13726 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13727 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13728 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13729 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13730 Use slp_node instead of slp_void_p and adjust.
13732 2013-04-09 Richard Biener <rguenther@suse.de>
13734 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13735 work that is not necessary.
13737 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13739 PR tree-optimization/56854
13740 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13741 forward into clobber stmts if it would change MEM_REF lhs into
13744 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13746 * tree.c (type_hash_lookup, type_hash_add): Make static.
13747 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13749 2013-04-09 Richard Biener <rguenther@suse.de>
13751 * tree.h (unsave_expr_now): Remove.
13752 * tree-inline.c (mark_local_for_remap_r): Remove.
13753 (unsave_expr_1): Likewise.
13754 (unsave_r): Likewise.
13755 (unsave_expr_now): Likewise.
13756 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13757 (propagate_tree_value): Likewise.
13759 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13761 * doc/rtl.texi (sequence): Rewrite documentation to match the
13762 current use of SEQUENCE rtl objects.
13763 * rtl.def (SEQUENCE): Likewise.
13765 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13766 Update documentation.
13767 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13768 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13770 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13772 2013-04-08 Teresa Johnson <tejohnson@google.com>
13774 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13775 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13777 (estimate_edge_size_and_time): Add comment to suggest using rounding
13779 (estimate_node_size_and_time): Ditto.
13780 (remap_edge_change_prob): Use helper rounding divide methods.
13781 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13782 (gimple_mod_pow2_value_transform): Ditto.
13783 (gimple_mod_subtract_transform): Ditto.
13784 (gimple_ic_transform): Ditto.
13785 (gimple_stringops_transform): Ditto.
13786 * stmt.c (conditional_probability): Ditto.
13787 (emit_case_dispatch_table): Ditto.
13788 * lto-cgraph.c (merge_profile_summaries): Ditto.
13789 * tree-optimize.c (execute_fixup_cfg): Ditto.
13790 * cfgcleanup.c (try_forward_edges): Ditto.
13791 * cfgloopmanip.c (scale_loop_profile): Ditto.
13793 (duplicate_loop_to_header_edge): Ditto.
13794 (lv_adjust_loop_entry_edge): Ditto.
13795 * tree-vect-loop.c (vect_transform_loop): Ditto.
13796 * profile.c (compute_branch_probabilities): Ditto.
13797 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13798 * lto-streamer-in.c (input_cfg): Ditto.
13799 * gimple-streamer-in.c (input_bb): Ditto.
13800 * ipa-cp.c (update_profiling_info): Ditto.
13801 (update_specialized_profile): Ditto.
13802 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13803 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13805 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13806 (compute_trg_info): Ditto.
13807 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13808 (purge_dead_edges): Ditto.
13809 * loop-unswitch.c (unswitch_loop): Ditto.
13810 * cgraphclones.c (cgraph_clone_edge): Ditto.
13811 (cgraph_clone_node): Ditto.
13812 * tree-inline.c (copy_bb): Ditto.
13813 (copy_edges_for_bb): Ditto.
13814 (initialize_cfun): Ditto.
13815 (copy_cfg_body): Ditto.
13816 (expand_call_inline): Ditto.
13818 2013-04-08 Kai Tietz <ktietz@redhat.com>
13820 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13821 TARGET_CYGWIN64 by TARGET_64BIT.
13823 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13825 * config/epiphany/epiphany.md (GPR_1): New constant.
13826 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13827 * config/epiphany/epiphany.c (gen_compare_reg):
13828 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13829 is already in place.
13830 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13831 Don't require being called during rtl expansion; If y operlaps r0,
13833 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13834 (epiphany_expand_epilogue): Likewise.
13836 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13837 Don't use CC_FPmode for ORDERED / UNORDERED.
13838 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
13840 * config/epiphany/constraints.md (CnL): New constraint.
13841 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
13842 * config/epiphany/predicates.md (add_operand): Allow 1024.
13844 * config/epiphany/epiphany.md (logical_op): New code iterator.
13845 (op_mnc): New code attribute.
13846 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
13847 (mov_f+1, mov_f+2): New peephole2 patterns.
13849 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
13850 (cstoresi4): Also allow re-use of zero result when doing a NE
13851 comparison to a non-zero operand.
13852 Use (clobber (scratch)) for first insn if the gpr output is not needed.
13854 * config/epiphany/epiphany.md (<insn_opname>v2si3):
13855 Use gen_addsi3_i / gen_subsi3_i.
13857 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13861 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
13862 contain anything but clobbers, at most one __builtin_stack_restore,
13863 optionally debug stmts and final resx, and if it has at least one
13864 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
13865 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
13866 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
13867 which isn't defaut definition, remove them.
13868 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
13869 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
13870 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
13871 with MEM_REF LHS with SSA_NAME address.
13873 2013-04-08 Jeff Law <law@redhat.com>
13875 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
13877 2013-04-08 Richard Biener <rguenther@suse.de>
13879 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
13881 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
13882 determined vector type.
13883 (vect_analyze_data_refs): Likewise.
13884 (vect_get_new_vect_var): Adjust.
13885 (vect_create_destination_var): Preserve SSA name versions.
13886 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
13887 not dump anything here.
13889 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13891 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
13892 Add member lr_slot_known.
13893 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
13895 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13896 Remove code that sets lr_slot_offset according to what a previous
13897 version of epiphany_emit_save_restore used to do.
13898 (epiphany_emit_save_restore): When doing an lr save or restore,
13899 set/verify lr_slot_known and lr_slot_offset.
13901 2013-04-08 Xinyu Qi <xyqi@marvell.com>
13904 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
13907 2013-04-08 Richard Biener <rguenther@suse.de>
13909 * alias.c (find_base_term): Fix thinko in previous change.
13911 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13913 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
13914 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
13915 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
13916 if possible to compute val.
13917 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
13918 For QImode integers don't require anything about precision. Use
13919 const_with_all_bytes_same to find out if the constant doesn't have
13920 repeated bytes in it.
13922 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13924 * config/s390/s390.c (s390_expand_insv): Only accept insertions
13927 2013-04-08 Marek Polacek <polacek@redhat.com>
13929 PR rtl-optimization/48182
13930 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
13933 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13936 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
13937 nuses, make sure we have a label.
13939 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13942 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
13943 (rs6000_emit_swdiv_low_precision): Remove.
13944 (rs6000_emit_swdiv): Rewrite to handle between one and four
13945 iterations of Newton-Raphson generally; modify required number of
13946 iterations for some cases.
13947 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
13949 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
13951 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
13952 set-but-unused variable.
13954 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
13955 basic blocks of released function bodies garbage-collectable.
13957 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
13958 (struct rtl_opt_pass): Add TODO_df_finish.
13960 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
13962 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13964 * config/arm/constraints.md (q): New constraint.
13965 * config/arm/ldrdstrd.md: New file.
13966 * config/arm/arm.md (ldrdstrd.md) New include.
13967 (arm_movdi): Use "q" instead of "r" constraint
13968 for double-word memory access.
13969 (movdf_soft_insn): Likewise.
13970 * config/arm/vfp.md (movdi_vfp): Likewise.
13971 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
13972 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
13973 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
13974 (mem_ok_for_ldrd_strd): Likewise.
13975 (output_move_double): Update assertion.
13977 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13979 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
13981 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13983 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
13984 define_insn_and_split.
13985 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
13987 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13989 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
13990 define_insn_and_split.
13991 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
13992 (shiftsi3_compare): New pattern.
13993 (rrx): New pattern.
13994 * config/arm/unspecs.md (UNSPEC_RRX): New.
13996 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13998 * config/arm/arm.md (negdi_extendsidi): New pattern.
13999 (negdi_zero_extendsidi): Likewise.
14001 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14003 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
14004 define_insn_and_split.
14005 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
14006 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
14008 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14010 * config/arm/arm.md (arm_subdi3): Convert define_insn into
14011 define_insn_and_split.
14012 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
14013 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
14015 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14017 * config/arm/arm.md (subsi3_carryin): New pattern.
14018 (subsi3_carryin_const): Likewise.
14019 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
14020 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
14022 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14024 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
14026 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14028 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
14029 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
14031 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14033 * config/arm/arm.c (arm_expand_builtin): Change fcode
14034 type to unsigned int.
14036 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14038 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
14040 2013-04-04 Ian Lance Taylor <iant@google.com>
14042 * doc/standards.texi (Standards): The Go frontend supports the Go 1
14045 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
14047 PR middle-end/56729
14048 * df-scan.c (df_insn_delete): Disable failing assert.
14050 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14052 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
14053 New function prototype.
14054 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
14055 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
14056 (arm_builtin_vectorized_function): New function.
14058 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14060 * config/arm/arm_neon_builtins.def: New file.
14061 * config/arm/arm.c (neon_builtin_data): Move contents to
14062 arm_neon_builtins.def.
14063 (enum arm_builtins): Include neon builtin definitions.
14064 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
14065 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
14067 2013-04-04 Marek Polacek <polacek@redhat.com>
14069 PR tree-optimization/48186
14070 * predict.c (maybe_hot_frequency_p): Return false if
14071 HOT_BB_FREQUENCY_FRACTION is 0.
14072 (cgraph_maybe_hot_edge_p): Likewise.
14074 2013-04-04 Richard Biener <rguenther@suse.de>
14076 PR tree-optimization/56826
14077 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
14080 2013-04-04 Richard Biener <rguenther@suse.de>
14082 PR tree-optimization/56213
14083 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
14084 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
14086 2013-04-04 Richard Biener <rguenther@suse.de>
14088 PR tree-optimization/56837
14089 * tree-loop-distribution.c (classify_partition): For non-zero
14090 values require that the value has the same precision as its
14091 mode to be useful as memset value.
14093 2013-04-03 Nick Clifton <nickc@redhat.com>
14095 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
14096 (fmssf4): Use fmsf.s on E3V5 architectures.
14097 (fnmasf4): Use fnmaf.s on E3V5 architectures.
14098 (fnmssf4): Use fnmsf.s on E3V5 architectures.
14100 2013-04-03 Jeff Law <law@redhat.com>
14102 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
14103 (lra-eliminations.o): Likewise.
14105 2013-04-03 Teresa Johnson <tejohnson@google.com>
14107 * gcov-io.c (compute_working_sets): Moved most of body of old
14108 compute_working_sets here from profile.c.
14109 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
14110 (gcov_working_set_t): Moved typedef here from basic-block.h
14111 (compute_working_set): Declare.
14112 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
14113 (get_working_sets): Renamed from compute_working_set,
14114 replace most of body with call to new compute_working_sets.
14115 (get_exec_counts): Replace call to compute_working_sets
14116 to get_working_sets.
14117 * profile.h (get_working_sets): Renamed from compute_working_set.
14118 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
14119 to get_working_sets.
14120 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
14121 * gcov-dump.c (dump_working_sets): New function.
14123 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
14125 * hwint.c (sext_hwi, zext_hwi): New functions.
14126 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
14127 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
14128 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
14129 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
14130 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
14131 (sext_hwi, zext_hwi): New functions.
14133 2013-04-03 Jeff Law <law@redhat.com>
14135 PR tree-optimization/56799
14136 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
14137 back test for widening conversion erroneously dropped in prior change.
14139 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14142 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
14143 instead of next_real_insn.
14145 2013-04-03 Marek Polacek <polacek@redhat.com>
14148 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
14150 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14153 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
14155 (thumb1_output_casesi): Likewise.
14156 (thumb2_output_casesi): Likewise.
14158 2013-04-03 Richard Biener <rguenther@suse.de>
14160 PR tree-optimization/56817
14161 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
14163 (tree_unroll_loops_completely_1): ... new function to manually
14164 walk the loop tree, properly defering outer loops of unrolled
14165 loops to later iterations.
14167 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14169 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
14170 (vectorizable_load): Likewise.
14171 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
14172 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
14174 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14176 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
14179 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14181 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
14183 2013-04-03 Bin Cheng <bin.cheng@arm.com>
14185 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
14187 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14189 PR tree-optimization/56790
14190 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
14193 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14195 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
14197 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
14198 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
14201 2013-04-03 Jakub Jelinek <jakub@redhat.com>
14204 * tree.h (force_folding_builtin_constant_p): New decl.
14205 * builtins.c (force_folding_builtin_constant_p): New variable.
14206 (fold_builtin_constant_p): Fold immediately also if
14207 force_folding_builtin_constant_p.
14209 2013-04-03 Richard Biener <rguenther@suse.de>
14211 PR tree-optimization/56812
14212 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
14213 DRs of the same interleaving chain are independent.
14215 2013-04-02 Jason Merrill <jason@redhat.com>
14217 * gdbinit.in (pbb): Use debug fn.
14219 2013-04-02 Lawrence Crowl <crowl@google.com>
14221 * sese.h (struct ivtype_map_elt_s): Remove unused.
14222 (extern debug_ivtype_map): Remove unused.
14223 (extern eq_ivtype_map_elts): Remove unused.
14224 * sese.c (debug_ivtype_map): Removed unused.
14225 (debug_ivtype_map_1): Removed unused.
14226 (debug_ivtype_elt): Remove unused.
14227 (eq_ivtype_map_elts): Remove unused.
14230 2013-04-02 Kai Tietz <ktietz@redhat.com>
14233 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
14234 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
14235 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
14237 (legitimize_pe_coff_symbol): Likewise.
14238 (is_imported_p): New helper-function.
14239 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
14240 for Windows x64 targets.
14241 (ix86_expand_prologue): Optimize for pe-coff targets.
14242 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
14243 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
14244 medium/large code-model.
14245 (legitimize_pic_address): Likewise.
14246 (legitimize_tls_address): Likewise.
14247 (ix86_expand_call): Likewise.
14248 (x86_output_mi_thunk): Likewise.
14249 (get_dllimport_decl): Add new beimport argument.
14250 (construct_plt_address): Don't assert for x64 pe-coff targets.
14251 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
14253 (SYMBOL_FLAG_STUBVAR): New macro.
14254 (SYMBOL_REF_STUBVAR_P): Likewise.
14255 * config/i386/winnt.c (stub_list): New structure.
14256 (stub_head): New local variable.
14257 (i386_pe_record_stub): New function.
14258 (i386_pe_file_end): Emit refptr-stubs.
14260 2013-04-02 Jakub Jelinek <jakub@redhat.com>
14262 PR rtl-optimization/56745
14263 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
14264 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
14267 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
14268 and both of them are MEM_REFs, just compare first argument for
14269 equality and attempt to deal even with differing offsets.
14272 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
14273 of gimple_clobber_p to be MEM_REF.
14274 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
14275 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
14276 after gimplification.
14277 * asan.c (get_mem_ref_of_assignment): Don't instrument
14278 gimple_clobber_p stmts.
14279 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
14280 gimple_clobber_p stmt if they have MEM_REF lhs and
14281 are dead because of another gimple_clobber_p stmt.
14282 * tree-ssa-live.c (clear_unused_block_pointer): Treat
14283 gimple_clobber_p stmts like debug stmts.
14284 (remove_unused_locals): Remove clobbers with MEM_REF lhs
14285 that refer to unused VAR_DECLs or uninitialized values.
14286 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
14287 gimple_clobber_p stmts if they refer to removed parameters.
14288 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
14291 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
14293 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
14294 using SWI48 mode attribute.
14296 2013-04-02 Wei Mi <wmi@google.com>
14298 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
14299 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
14300 *<rotate_insn><mode>3_mask in i386.md.
14302 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
14304 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
14306 2013-04-02 Richard Biener <rguenther@suse.de>
14308 PR tree-optimization/56778
14309 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
14310 Runtime alias tests are not supported for gather loads.
14311 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
14312 stmts referenced from SSA operands before updating SSA form.
14314 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
14315 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14317 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
14318 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
14319 * config/arm/cortex-a53.md: New file.
14320 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
14321 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
14322 * config/arm/arm.c (arm_issue_rate): Likewise.
14323 * config/arm/arm-tune.md: Regenerate
14324 * config/arm/arm-tables.opt: Regenerate.
14325 * config/arm/arm-cores.def: Add cortex-a53.
14327 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
14329 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
14332 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14334 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
14335 scalar load/store operations using B/H registers.
14336 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
14338 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14340 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
14342 * config/aarch64/aarch64.c
14343 (aarch64_simd_scalar_immediate_valid_for_move): New.
14344 * config/aarch64/aarch64-protos.h
14345 (aarch64_simd_scalar_immediate_valid_for_move): New.
14346 * config/aarch64/constraints.md (Dh, Dq): New.
14347 * config/aarch64/iterators.md (hq): New.
14349 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
14351 * reorg.c (get_branch_condition): Deal with conditional returns.
14352 (fill_simple_delay_slots): Remove dead code dealing with jumps.
14354 2013-04-01 Wei Mi <wmi@google.com>
14356 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
14357 Truncate operand 2 using %b asm operand modifier.
14358 (*<shift_insn><mode>3_mask): Ditto.
14359 (*<rotate_insn><mode>3_mask): Ditto.
14361 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
14363 PR middle-end/56798
14364 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
14366 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
14368 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
14370 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
14372 2013-03-30 Lawrence Crowl <crowl@google.com>
14374 * dse.c (clear_alias_sets): Remove never set.
14375 (disqualified_clear_alias_sets): Remove never set.
14376 (clear_alias_mode_pool): Remove never set.
14377 (dse_step0): Remove condition that is never true.
14378 (canon_address): Remove condition that is never true.
14379 (dse_step7): Remove condition that is never true.
14380 (rest_of_handle_dse): Remove condition that is never true.
14381 (rest_of_handle_dse::did_global): Remove never read from above.
14382 (dse_step2_spill): Remove never called from above.
14383 (dse_step5_spill): Remove never called from above.
14385 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
14387 * doc/md.texi (Standard Names) <casesi>: Update documentation for
14388 JUMP_TABLE_DATA changes.
14389 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
14390 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
14391 (Insns) <jump_table_data>: New entry.
14392 * doc/tm.texi: Regenerate.
14394 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
14396 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
14397 for table jump at the end of a basic block using tablejump_p.
14398 * targhooks.c (default_invalid_within_doloop): Likewise.
14399 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
14400 target hook implementation that is identical to the default hook.
14401 (rs6000_invalid_within_doloop): Remove.
14403 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
14404 unused variable from tablejump_p call.
14406 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
14407 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
14408 (INSN_DELETED_P): Likewise.
14409 (emit_jump_table_data): New prototype.
14410 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
14411 after 4th as unused.
14412 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
14413 * sched-vis.c (print_insn): Likewise.
14414 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
14415 insn for compatibility with back ends that use next_active_insn to
14416 identify jump table data.
14417 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
14418 (remove_insn): Likewise.
14419 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
14421 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
14422 (emit_jump_table_data): New function.
14424 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
14425 basic block, a JUMP_TABLE_DATA never is.
14426 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
14427 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
14428 off from code handling real insns.
14429 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
14430 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
14431 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
14432 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
14433 is not a NONDEBUG_INSN_P.
14434 * ira-costs.c (scan_one_insn): Likewise.
14435 * jump.c (mark_all_labels): Likewise.
14436 (mark_jump_label_1): Likewise.
14437 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14438 * lra.c (get_insn_freq): Expect all insns reaching here to be in
14440 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
14441 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
14442 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
14443 JUMP_TABLE_DATA_P insns.
14444 (calculate_elim_costs_all_insns): Likewise.
14445 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
14446 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
14447 (delete_output_reload): Code style fixups.
14448 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
14449 insn flags on this non-insn.
14450 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
14451 as scheduling barriers, for pre-change compatibility.
14452 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
14453 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
14455 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
14456 redundant JUMP_TABLE_DATA_P test.
14457 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
14458 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14459 (frv_for_each_packet): Likewise.
14460 * config/i386/i386.c (min_insn_size): Likewise.
14461 (ix86_avoid_jump_mispredicts): Likewise.
14462 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14463 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
14464 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14465 (mips16_insn_length): Robustify.
14466 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
14467 (mips16_split_long_branches): Likewise.
14468 * config/pa/pa.c (pa_combine_instructions): Likewise.
14469 * config/rs6000/rs6000.c (get_next_active_insn): Treat
14470 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
14471 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
14472 as contributing to pool range lengths.
14473 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
14474 Remove redundant JUMP_TABLE_DATA_P test.
14475 (sh_loop_align): Likewise.
14476 (split_branches): Likewise.
14477 (sh_insn_length_adjustment): Likewise.
14478 * config/spu/spu.c (get_branch_target): Likewise.
14480 2013-03-29 Jan Hubicka <jh@suse.cz>
14482 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
14483 gcov streaming; stream hot bb threshold to ltrans.
14484 * predict.c (get_hot_bb_threshold): Break out from ....
14485 (maybe_hot_count_p): ... here.
14486 (set_hot_bb_threshold): New function.
14487 * lto-section-in.c (lto_section_name): Add profile.
14488 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
14489 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
14490 and data-streamer.h
14491 (histogram_entry): New structure.
14492 (histogram, histogram_pool): New global vars.
14493 (histogram_hash): New structure.
14494 (histogram_hash::hash): New method.
14495 (histogram_hash::equal): Likewise.
14496 (account_time_size): New function.
14497 (cmp_counts): New function.
14498 (dump_histogram): New function.
14499 (ipa_profile_generate_summary): New function.
14500 (ipa_profile_write_summary): New function.
14501 (ipa_profile_read_summary): New function.
14502 (ipa_profile): Decide on threshold.
14503 (pass_ipa_profile): Add ipa_profile_write_summary and
14504 ipa_profile_read_summary.
14505 * Makefile.in (ipa.o): Update dependencies.
14506 * lto-streamer.h (LTO_section_ipa_profile): New section.
14508 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
14510 * tree.h (VAR_P): New.
14512 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
14515 * doc/invoke.texi ([-fwhole-program]): Fix typo.
14517 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
14519 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
14520 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14521 (control_flow_insn_p): Likewise.
14522 * cfgrtl.c (duplicate_insn_chain): Likewise.
14523 * final.c (get_attr_length_1): Likewise.
14524 (shorten_branches): Likewise.
14525 (final_scan_insn): Likewise.
14526 * function.c (instantiate_virtual_regs): Likewise.
14527 * gcse.c (insert_insn_end_basic_block): Likewise.
14528 * ira-costs.c (scan_one_insn): Likewise.
14529 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14530 * lra.c (check_rtl): Likewise.
14531 * reload1.c (elimination_costs_in_insn): Likewise.
14532 * reorg.c (follow_jumps): Likewise.
14534 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
14535 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14536 (thumb_far_jump_used_p): Likewise.
14537 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
14538 (workaround_speculation): Likewise.
14539 (add_sched_insns_for_speculation): Likewise.
14540 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
14541 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14542 (frv_for_each_packet): Likewise.
14543 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
14544 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
14545 (final_emit_insn_group_barriers): Likewise.
14546 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14547 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14548 (mips16_insn_length): Likewise.
14549 * config/pa/pa.c (pa_reorg): Likewise.
14550 (pa_combine_instructions): Likewise.
14551 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
14552 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
14553 (sh_reorg): Likewise.
14554 (split_branches): Likewise.
14555 * config/spu/spu.c (get_branch_target): Likewise.
14557 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
14560 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
14562 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
14563 Fix declaration name.
14565 2013-03-28 Lawrence Crowl <crowl@google.com>
14567 * graphds.h (struct graph.indicies): Remove unused.
14568 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
14569 (SCOP_ORIGINAL_PDDRS): Remove unused.
14570 * sese.h (extern insert_loop_close_phis): Removed unused.
14571 (extern insert_guard_phis): Removed unused.
14572 (extern ivtype_map_elt_info): Removed unused.
14573 (new_ivtype_map_elt): Removed unused.
14574 * sese.c (ivtype_map_elt_info): Removed unused.
14576 2013-03-28 Lawrence Crowl <crowl@google.com>
14578 * Makefile.in: Add several missing include dependences.
14580 (test-dump.o): New. This object is not added to any executable,
14581 but is present for ad-hoc testing.
14583 (debug (const bitmap_head_def &)): New.
14584 (debug (const bitmap_head_def *)): New.
14586 (extern debug (const bitmap_head_def &)): New.
14587 (extern debug (const bitmap_head_def *)): New.
14589 (debug (edge_def &)): New.
14590 (debug (edge_def *)): New.
14592 (debug (basic_block_def &)): New.
14593 (debug (basic_block_def *)): New.
14595 (dump_node (const_tree, int, FILE *)): Correct source file.
14597 (debug (die_struct &)): New.
14598 (debug (die_struct *)): New.
14600 (extern debug (die_struct &)): New.
14601 (extern debug (die_struct *)): New.
14602 * gimple-pretty-print.c
14603 (debug (gimple_statement_d &)): New.
14604 (debug (gimple_statement_d *)): New.
14605 * gimple-pretty-print.h
14606 (extern debug (gimple_statement_d &)): New.
14607 (extern debug (gimple_statement_d *)): New.
14609 (debug (ira_allocno_copy &)): New.
14610 (debug (ira_allocno_copy *)): New.
14611 (debug (ira_allocno &)): New.
14612 (debug (ira_allocno *)): New.
14614 (extern debug (ira_allocno_copy &)): New.
14615 (extern debug (ira_allocno_copy *)): New.
14616 (extern debug (ira_allocno &)): New.
14617 (extern debug (ira_allocno *)): New.
14619 (debug (live_range &)): New.
14620 (debug (live_range *)): New.
14622 (debug (lra_live_range &)): New.
14623 (debug (lra_live_range *)): New.
14625 (debug (lra_live_range &)): New.
14626 (debug (lra_live_range *)): New.
14628 (debug (omega_pb_d &)): New.
14629 (debug (omega_pb_d *)): New.
14631 (extern debug (omega_pb_d &)): New.
14632 (extern debug (omega_pb_d *)): New.
14634 (debug (const rtx_def &)): New.
14635 (debug (const rtx_def *)): New.
14637 (debug_tree (tree): Move within file.
14638 (debug_raw (const tree_node &)): New.
14639 (debug_raw (const tree_node *)): New.
14640 (dump_tree_via_hooks (const tree_node *, int)): New.
14641 (debug (const tree_node &)): New.
14642 (debug (const tree_node *)): New.
14643 (debug_verbose (const tree_node &)): New.
14644 (debug_verbose (const tree_node *)): New.
14645 (debug_head (const tree_node &)): New.
14646 (debug_head (const tree_node *)): New.
14647 (debug_body (const tree_node &)): New.
14648 (debug_body (const tree_node *)): New.
14649 (debug_vec_tree (tree): Move and reimplement in terms of dump.
14650 (debug (vec<tree, va_gc> &)): New.
14651 (debug (vec<tree, va_gc> *)): New.
14653 (extern debug (const rtx_def &)): New.
14654 (extern debug (const rtx_def *)): New.
14656 (debug_raw (simple_bitmap_def &)): New.
14657 (debug_raw (simple_bitmap_def *)): New.
14658 (debug (simple_bitmap_def &)): New.
14659 (debug (simple_bitmap_def *)): New.
14661 (extern debug (simple_bitmap_def &)): New.
14662 (extern debug (simple_bitmap_def *)): New.
14663 (extern debug_raw (simple_bitmap_def &)): New.
14664 (extern debug_raw (simple_bitmap_def *)): New.
14666 (debug (vinsn_def &)): New.
14667 (debug (vinsn_def *)): New.
14668 (debug_verbose (vinsn_def &)): New.
14669 (debug_verbose (vinsn_def *)): New.
14670 (debug (expr_def &)): New.
14671 (debug (expr_def *)): New.
14672 (debug_verbose (expr_def &)): New.
14673 (debug_verbose (expr_def *)): New.
14674 (debug (vec<rtx> &)): New.
14675 (debug (vec<rtx> *)): New.
14677 (extern debug (vinsn_def &)): New.
14678 (extern debug (vinsn_def *)): New.
14679 (extern debug_verbose (vinsn_def &)): New.
14680 (extern debug_verbose (vinsn_def *)): New.
14681 (extern debug (expr_def &)): New.
14682 (extern debug (expr_def *)): New.
14683 (extern debug_verbose (expr_def &)): New.
14684 (extern debug_verbose (expr_def *)): New.
14685 (extern debug (vec<rtx> &)): New.
14686 (extern debug (vec<rtx> *)): New.
14688 (_list_iter_cond_expr): Make inline instead of static.
14690 (debug (sreal &)): New.
14691 (debug (sreal *)): New.
14693 (extern debug (sreal &)): New.
14694 (extern debug (sreal *)): New.
14696 (extern debug_raw (const tree_node &)): New.
14697 (extern debug_raw (const tree_node *)): New.
14698 (extern debug (const tree_node &)): New.
14699 (extern debug (const tree_node *)): New.
14700 (extern debug_verbose (const tree_node &)): New.
14701 (extern debug_verbose (const tree_node *)): New.
14702 (extern debug_head (const tree_node &)): New.
14703 (extern debug_head (const tree_node *)): New.
14704 (extern debug_body (const tree_node &)): New.
14705 (extern debug_body (const tree_node *)): New.
14706 (extern debug (vec<tree, va_gc> &)): New.
14707 (extern debug (vec<tree, va_gc> *)): New.
14709 (debug (struct loop &)): New.
14710 (debug (struct loop *)): New.
14711 (debug_verbose (struct loop &)): New.
14712 (debug_verbose (struct loop *)): New.
14713 * tree-dump.c: Add header dependence.
14715 (extern debug (struct loop &)): New.
14716 (extern debug (struct loop *)): New.
14717 (extern debug_verbose (struct loop &)): New.
14718 (extern debug_verbose (struct loop *)): New.
14720 (debug (data_reference &)): New.
14721 (debug (data_reference *)): New.
14722 (debug (vec<data_reference_p> &)): New.
14723 (debug (vec<data_reference_p> *)): New.
14724 (debug (vec<ddr_p> &)): New.
14725 (debug (vec<ddr_p> *)): New.
14727 (extern debug (data_reference &)): New.
14728 (extern debug (data_reference *)): New.
14729 (extern debug (vec<data_reference_p> &)): New.
14730 (extern debug (vec<data_reference_p> *)): New.
14731 (extern debug (vec<ddr_p> &)): New.
14732 (extern debug (vec<ddr_p> *)): New.
14734 (debug (pt_solution &)): New.
14735 (debug (pt_solution *)): New.
14737 (extern debug (pt_solution &)): New.
14738 (extern debug (pt_solution *)): New.
14740 (debug (_var_map &)): New.
14741 (debug (_var_map *)): New.
14742 (debug (tree_live_info_d &)): New.
14743 (debug (tree_live_info_d *)): New.
14745 (extern debug (_var_map &)): New.
14746 (extern debug (_var_map *)): New.
14747 (extern debug (tree_live_info_d &)): New.
14748 (extern debug (tree_live_info_d *)): New.
14750 2013-03-28 Jan Hubicka <jh@suse.cz>
14752 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14754 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14756 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14757 record only when desired or required.
14759 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14761 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14762 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14764 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14766 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14767 (*andsi3_compare0_uxtw): New pattern.
14768 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14769 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14771 2013-03-28 Jan Hubicka <jh@suse.cz>
14773 * data-streamer-in.c (streamer_read_gcov_count): New function.
14774 * gimple-streamer-out.c: Include value-prof.h.
14775 (output_gimple_stmt): Output histogram.
14776 (output_bb): Use streamer_write_gcov_count.
14777 * value-prof.c: Include data-streamer.h
14778 (dump_histogram_value): Add HIST_TYPE_MAX.
14779 (stream_out_histogram_value): New function.
14780 (stream_in_histogram_value): New function.
14781 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14782 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14783 * data-streamer-out.c (streamer_write_gcov_count): New function.
14784 (streamer_write_gcov_count_stream): New function.
14785 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14786 (lto_output_node): Likewise.
14787 (input_node, input_edge): Likewise.
14788 * lto-streamer-out.c (output_cfg): Update streaming.
14789 * lto-streamer-in.c (input_cfg): Likewise.
14790 * data-streamer.h (streamer_write_gcov_count,
14791 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14792 * gimple-streamer-in.c: Include value-prof.h
14793 (input_gimple_stmt): Input histograms.
14794 (input_bb): Update profile streaming.
14796 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14798 * genmodes.c (emit_max_int): New function.
14799 (emit_insn_modes_h): Added call to emit_max_function.
14800 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14802 * machmode.def: Fixed comment.
14804 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14806 * combine.c (try_combine): Removed useless assert.
14807 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14809 2013-03-28 Marek Polacek <polacek@redhat.com>
14810 Richard Biener <rguenther@suse.de>
14812 PR tree-optimization/56695
14813 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14814 build signed result of a vector comparison.
14815 * tree-cfg.c (verify_gimple_comparison): Check that a result
14816 of a vector comparison has signed type.
14818 2013-03-28 Richard Biener <rguenther@suse.de>
14820 PR tree-optimization/37021
14821 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14822 do not restrict gaps between groups.
14823 * tree-vect-stmts.c (vectorizable_load): Properly account for
14824 a gap between groups.
14826 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14828 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14829 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14832 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14834 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14835 * doc/extend.texi (Named Address Spaces): Ditto.
14836 (Variable Attributes): Ditto.
14838 2013-03-27 Kai Tietz <ktietz@redhat.com>
14840 * config.build: Add support for cygwin x64 target.
14841 * config.gcc: Likewise.
14842 * config.host: Likewise.
14843 * configure.ac: Likewise
14844 * configure: Regenerated.
14846 2013-03-27 Kai Tietz <ktietz@redhat.com>
14848 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
14849 * config/i386/t-cygwin-w64: New file.
14850 * config/i386/cygwin-w64.h: New file.
14851 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
14852 and add support for x64-cygwin target.
14853 (CPP_SPEC): Likewise.
14854 (CXX_WRAP_SPEC_LIST): Undefine before define.
14855 (LIBGCJ_SONAME): Use 15 as version.
14857 2013-03-27 Richard Biener <rguenther@suse.de>
14859 PR tree-optimization/56716
14860 * tree-ssa-structalias.c (perform_var_substitution): Adjust
14861 dumping for ref nodes.
14863 2013-03-27 Martin Jambor <mjambor@suse.cz>
14865 PR tree-optimization/55334
14866 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
14867 restricted pointers to arrays.
14869 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
14871 * Makefile.in (.SUFFIXES): Add .cc.
14872 (.c.o): Apply same recipe for implicit rule .cc.o.
14874 2013-03-27 Richard Biener <rguenther@suse.de>
14876 PR tree-optimization/37021
14877 * tree-vect-data-refs.c (vect_check_strided_load): Allow
14878 REALPART/IMAGPART_EXPRs around the supported refs.
14879 * tree-ssa-structalias.c (find_func_aliases): Assume that
14880 floating-point values are not used to transfer pointers.
14882 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
14884 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
14885 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
14886 * gcc/doc/tm.texi: Regenerate.
14887 * gcc/targhooks.h (default_has_ifunc_p): New.
14888 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
14889 * gcc/config/linux-protos.h: New file.
14890 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
14891 this hook for linux which disables support of indirect functions in
14893 * gcc/config/linux-android.c: New file.
14894 * gcc/config/t-linux-android.c: Ditto.
14895 * gcc/config.gcc: Added new object file linux-android.o.
14896 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
14897 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
14898 * gcc/varasm.c (do_assemble_alias): Likewise.
14899 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
14900 doesn't support indirect functions.
14901 * configure: Regenerate.
14903 2013-03-27 Bin Cheng <bin.cheng@arm.com>
14906 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
14907 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
14910 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14912 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
14914 2013-03-27 Terry Guo <terry.guo@arm.com>
14916 * config/arm/arm-cores.def: Added core cortex-r7.
14917 * config/arm/arm-tune.md: Regenerated.
14918 * config/arm/arm-tables.opt: Regenerated.
14919 * doc/invoke.texi: Added entry for core cortex-r7.
14921 2013-03-27 Walter Lee <walt@tilera.com>
14923 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
14924 double-decrement of next_scratch_regno.
14926 2013-03-27 Walter Lee <walt@tilera.com>
14928 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
14930 (insn_v1mulus): Ditto.
14931 (insn_v2muls): Ditto.
14933 2013-03-27 Walter Lee <walt@tilera.com>
14935 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
14936 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
14938 2013-03-27 Walter Lee <walt@tilera.com>
14940 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
14941 (*sibcall_value): Ditto.
14943 2013-03-27 Walter Lee <walt@tilera.com>
14945 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
14946 (insn_mnz_v8qi): ... this ...
14947 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
14949 (insn_v<n>mnz): Replaced by ...
14950 (insn_v1mnz): ... this ...
14951 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
14953 (insn_mz_<mode>): Replaced by ...
14954 (insn_mz_v8qi): ... this ...
14955 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
14957 (insn_v<n>mz): Replaced by ...
14958 (insn_v1mz): ... this ...
14959 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
14962 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
14964 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
14966 2013-03-26 Roland McGrath <mcgrathr@google.com>
14968 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
14969 than fprintf with a non-constant, non-format string.
14971 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14973 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
14974 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
14975 operand 0 predicate.
14976 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
14977 attribute. Use general_x64nomem_operand as operand 1 predicate.
14978 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
14979 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14980 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
14981 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14982 (mov<mode>_insv_1): Remove expander. Merge insn with
14983 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
14984 Use general_x64nomem_operand as operand 1 predicate.
14985 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
14986 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
14987 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14988 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
14989 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
14990 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
14991 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14992 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
14993 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
14994 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
14995 (general_x64nomem_operand): Ditto.
14997 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
14999 * config/rtems.opt: Add -pthread option.
15001 2013-03-26 Richard Biener <rguenther@suse.de>
15003 * alias.c (find_base_term): Avoid redundant and not used recursion.
15004 (base_alias_check): Get the initial base term from the caller.
15005 (true_dependence_1): Compute and pass base terms to base_alias_check.
15006 (write_dependence_p): Likewise.
15007 (may_alias_p): Likewise.
15009 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
15011 * config/aarch64/aarch64.c (aarch64_classify_address): Support
15012 PC-relative load in SI modes and above only.
15014 2013-03-26 Xinyu Qi <xyqi@marvell.com>
15016 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
15017 * config/arm/iwmmxt.md (WCGR0): Update.
15018 (WCGR1, WCGR2, WCGR3): Likewise.
15020 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
15022 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
15023 Use x64 and nox64 isa attributes.
15025 2013-03-26 Richard Biener <rguenther@suse.de>
15027 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
15028 alignment computations and rely on get_object_alignment_1
15029 for the !TYPE_P case.
15030 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
15032 2013-03-26 Walter Lee <walt@tilera.com>
15034 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
15035 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
15037 2013-03-25 Jeff Law <law@redhat.com>
15039 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
15040 check for INTEGRAL_TYPE_P that was missing due to checking in
15041 wrong version of prior patch.
15043 2013-03-25 Walter Lee <walt@tilera.com>
15045 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
15046 TILEGX_INSN_SHUFFLEBYTES1.
15047 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
15049 (tilegx_builtins): Ditto.
15050 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
15052 2013-03-25 Walter Lee <walt@tilera.com>
15054 * config/tilegx/tilegx.md (floatsisf2): New pattern.
15055 (floatunssisf2): New pattern.
15056 (floatsidf2): New pattern.
15057 (floatunssidf2): New pattern.
15059 2013-03-25 Walter Lee <walt@tilera.com>
15061 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
15062 tests for constraint J, K, N, P.
15064 2013-03-25 Walter Lee <walt@tilera.com>
15066 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
15067 Use indirect/pcrel encoding.
15068 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
15071 2013-03-25 Steve Ellcey <sellcey@mips.com>
15073 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
15074 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
15075 * config/mips/mips.c (mips_option_override): Set IMADD default.
15076 * config/mips/mips.h (PTF_AVOID_IMADD): New.
15077 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
15078 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
15079 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
15080 * doc/invoke.texi (-mimadd/-mno-imadd): New.
15082 2013-03-25 Jeff Law <law@redhat.com>
15084 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
15085 slightly to avoid creating and folding useless trees. Simplify
15086 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
15088 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
15090 * config/i386/i386.md (*zero_extendsidi2): Merge with
15091 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
15092 * config/i386/predicates.md (x86_64_zext_operand): Rename from
15093 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
15094 targets. Clarify comment.
15096 2013-03-25 Martin Jambor <mjambor@suse.cz>
15098 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
15099 pass-through jump functions differently.
15100 (ipa_read_jump_function): Likewise. Also use setter functions to set
15103 2013-03-25 Martin Jambor <mjambor@suse.cz>
15105 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
15106 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
15108 (ipa_get_indirect_edge_target): New function.
15109 (devirtualization_time_bonus): New parameter known_aggs, pass it to
15110 ipa_get_indirect_edge_target. Update all callers.
15111 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
15112 ipa_get_indirect_edge_target_1 instead of calling
15113 ipa_get_indirect_edge_target.
15114 (create_specialized_node): Pass aggvlas to
15115 ipcp_discover_new_direct_edges.
15117 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15119 * config/arm/arm.md (f_sels, f_seld): New types.
15120 (*cmov<mode>): New pattern.
15121 * config/arm/predicates.md (arm_vsel_comparison_operator): New
15124 2013-03-25 Kai Tietz <ktietz@redhat.com>
15126 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
15127 POSIX-printf for mingw-hosted builds.
15129 2013-03-25 Richard Biener <rguenther@suse.de>
15131 PR middle-end/56694
15132 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
15133 must-not-throw stmt location.
15135 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15137 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
15138 Emit load-acquire versions when acq is true.
15139 (arm_emit_store_exclusive): Add rel parameter.
15140 Emit store-release versions when rel is true.
15141 (arm_split_compare_and_swap): Use acquire-release instructions
15143 of barriers when appropriate.
15144 (arm_split_atomic_op): Likewise.
15145 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
15146 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
15147 (VUNSPEC_SLX): Likewise.
15148 (VUNSPEC_LDA): Likewise.
15149 (VUNSPEC_STL): Likewise.
15150 * config/arm/sync.md (atomic_load<mode>): New pattern.
15151 (atomic_store<mode>): Likewise.
15152 (arm_load_acquire_exclusive<mode>): Likewise.
15153 (arm_load_acquire_exclusivesi): Likewise.
15154 (arm_load_acquire_exclusivedi): Likewise.
15155 (arm_store_release_exclusive<mode>): Likewise.
15157 2013-03-25 Catherine Moore <clm@codesourcery.com>
15159 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
15160 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
15161 * config/mip/predicates.md (lwsp_swsp_operand,
15162 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
15163 sb16_operand, db4_operand, db7_operand, ib3_operand,
15164 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
15165 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
15166 andi16_operand): New predicates.
15167 * config/mips/mips.md (compression): New attribute.
15168 (enabled): New attribute.
15169 (length): Consider compression in computing length.
15170 (shift_compression): New code attribute.
15171 (*add<mode>3): New operands. Record compression.
15172 (sub<mode>3): Likewise.
15173 (one_cmpl<mode>2): Likewise.
15174 (*and<mode>3): Likewise.
15175 (*ior<mode>3): Likewise.
15176 (unnamed pattern for xor): Likewise.
15177 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
15178 (*<optab><mode>3): Likewise.
15179 (*mov<mode>_internal: Likewise.
15180 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
15181 (mips_unsigned_immediate_p): New.
15182 (umips_lwsp_swsp_address_p): New.
15183 (m16_based_address_p): New.
15184 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
15185 (mips_unsigned_immediate_p): New prototype.
15186 (lwsp_swsp_address_p): New prototype.
15187 (m16_based_address_p): New prototype.
15188 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
15189 (mips_signed_immediate_p): New function.
15190 (m16_based_address_p): New function.
15191 (lwsp_swsp_address_p): New function.
15192 (mips_print_operand_punctuation): Recognize short delay slot insns
15193 for microMIPS.add<mode>3"
15195 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15198 * config/arm/iterators.md (v_cmp_result): New mode attribute.
15199 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
15201 2013-03-25 Richard Biener <rguenther@suse.de>
15203 PR tree-optimization/56689
15204 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
15207 2013-03-25 Richard Biener <rguenther@suse.de>
15209 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
15211 (memory_references): Likewise.
15212 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
15213 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
15214 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
15215 (gather_mem_refs_in_loops): Fold into ...
15216 (analyze_memory_references): ... this. Move initialization
15217 to tree_ssa_lim_initialize.
15218 (fill_always_executed_in): Rename to ...
15219 (fill_always_executed_in_1): ... this.
15220 (fill_always_executed_in): Move contains_call computation to
15221 this new function from ...
15222 (tree_ssa_lim_initialize): ... here.
15223 (tree_ssa_lim): Call fill_always_executed_in.
15225 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15227 * postreload.c (reload_combine): Fix code detecting returns.
15229 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15231 * function.c (emit_use_return_register_into_block): On cc0 targets,
15232 do not emit the sequence between cc0 setter and user.
15234 2013-03-25 Kai Tietz <ktietz@redhat.com>
15236 * config/i386/predicates.md (local_symbolic_operand): Interpret
15237 dll-imported symbols as none-local.
15239 2013-03-25 Richard Biener <rguenther@suse.de>
15241 * tree-ssa-loop-im.c (struct depend): Remove.
15242 (struct lim_aux_data): Make depends a vec of gimples.
15243 (free_lim_aux_data): Adjust.
15244 (add_dependency): Likewise.
15245 (set_level): Likewise.
15247 2013-03-25 Richard Biener <rguenther@suse.de>
15249 PR middle-end/56434
15250 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
15251 the pointer returned by calls with ECF_MALLOC set.
15253 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15255 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
15257 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15259 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
15260 using MMXMODE mode iterator.
15261 (*move<mode>_internal): Merge with *movv2sf_internal and
15262 *movv2sf_internal_rex64 using MMXMODE mode iterator.
15264 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15266 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
15267 (record_last_mem_set_info): Likewise.
15269 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
15270 of XNEWVEC followed by memset.
15271 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
15273 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15275 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
15276 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
15277 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
15278 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
15279 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
15280 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
15281 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
15282 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
15283 BARRIER_P instead of GET_CODE.
15285 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15287 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
15288 inaccuracy in the probing code.
15290 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
15291 (ctrapdi4): Likewise.
15293 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15295 * calls.c (expand_call): Add missing guard to code handling return
15296 of non-BLKmode structures in MSB.
15297 * function.c (expand_function_end): Likewise.
15299 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15301 * combine.c (try_combine): Adjust comment. Do not add the set of
15302 insn #0 if the destination indirectly is set or dies in insn #2.
15303 Tidy up code to distribute a new note.
15305 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
15307 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
15308 also for alternatives 16 and 17.
15310 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
15312 * config/i386/sse.md (*mov<mode>_internal): Merge with
15313 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
15314 Emit insn template depending on type attribute. Use
15315 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15316 movd instead of movq mnemonic for interunit moves. Rewrite mode
15317 attribute calculation. Remove unit attribute calculation.
15318 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
15319 Set prefix_data16 attribute for DImode ssemov types.
15320 Use Ym instead of y for SSE-MMX conversion alternatives.
15321 Reorder operand constraints.
15323 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
15325 * df.h (df_insn_delete): Adjust prototype.
15326 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
15327 and let it decide whether mark the basic block dirty.
15328 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
15329 * df-scan.c (df_insn_info_delete): New helper function, split
15330 off from df_insn_delete.
15331 (df_scan_free_bb_info): Use it.
15332 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
15334 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
15335 that the insn is actually an insn and it has a non-NULL basic block.
15336 Do not mark basic block dirty if only deleting a DEBUG_INSN.
15338 2013-03-22 Richard Biener <rguenther@suse.de>
15340 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
15342 (mem_ref_alloc): Do not allocate them.
15343 (refs_independent_p): Do not query or maintain a cache.
15345 2013-03-22 Richard Biener <rguenther@suse.de>
15347 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
15348 (gather_mem_refs_in_loops): Do not compute it.
15349 (analyze_memory_references): Do not allocate it.
15350 (tree_ssa_lim_finalize): Do not free it.
15351 (for_all_locs_in_loop): Do not query all_refs_in_loop.
15353 2013-03-22 Richard Biener <rguenther@suse.de>
15355 * is-a.h (as_a): Use gcc_checking_assert.
15357 2013-03-22 Ian Bolton <ian.bolton@arm.com>
15359 * config/aarch64/aarch64.c (aarch64_print_operand): New
15360 format specifier for printing a constant in hex.
15361 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
15362 format specifier for printing second operand.
15364 2013-03-22 Richard Biener <rguenther@suse.de>
15366 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
15368 (gather_mem_refs_in_loops): Perform store accumulation here.
15369 (create_vop_ref_mapping_loop): Remove.
15370 (create_vop_ref_mapping): Likewise.
15371 (analyze_memory_references): Initialize refs_stored_in_loop.
15372 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
15373 (record_indep_loop): Remove.
15374 (record_dep_loop): New function.
15375 (ref_indep_loop_p_1): Adjust to only walk over references
15376 in the loop, not its subloops.
15377 (ref_indep_loop_p): Rename to ...
15378 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
15379 maintaining a more fine-grained cache.
15380 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
15381 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
15383 2013-03-22 Richard Biener <rguenther@suse.de>
15385 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
15386 (struct mem_ref): Make accesses_in_loop a vec of a vec of
15387 aggregate mem_ref_loc.
15388 (free_mem_ref_locs): Inline into ...
15389 (memref_free): ... this and adjust.
15390 (mem_ref_alloc): Adjust.
15391 (mem_ref_locs_alloc): Remove.
15392 (record_mem_ref_loc): Adjust.
15393 (get_all_locs_in_loop): Rewrite into ...
15394 (for_all_locs_in_loop): ... this iterator.
15395 (rewrite_mem_ref_loc): New functor.
15396 (rewrite_mem_refs): Use for_all_locs_in_loop.
15397 (sm_set_flag_if_changed): New functor.
15398 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
15399 (ref_always_accessed): New functor.
15400 (ref_always_accessed_p): Use for_all_locs_in_loop.
15402 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15404 * tree-pass.h (PROP_gimple_lvec): New.
15405 * passes.c (dump_properties): Handle PROP_gimple_lvec.
15406 (init_optimization_passes): Move pass_lower_vector.
15407 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
15409 (pass_lower_vector): Provide PROP_gimple_lvec.
15410 (pass_lower_vector_ssa): Likewise.
15411 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
15413 2013-03-21 Mark Wielaard <mjw@redhat.com>
15415 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
15417 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
15419 * config/i386/i386.md (*movdi_internal): Disparage slightly
15420 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
15421 conversion alternatives.
15423 2013-03-21 Jakub Jelinek <jakub@redhat.com>
15425 PR middle-end/48087
15426 * diagnostic.def (DK_WERROR): New kind.
15427 * diagnostic.h (werrorcount): Define.
15428 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
15429 promoted to DK_ERROR, increment DK_WERROR counter instead of
15431 * toplev.c (toplev_main): Call print_ignored_options even if
15432 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
15433 even if just werrorcount is non-zero.
15436 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
15438 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
15439 (string_cst_pool_decl): New function.
15440 (optimize_one_addr_into_implicit_ptr): New function.
15441 (resolve_addr_in_expr): Optimize DWARF location expression
15442 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
15443 which doesn't live in memory, but has DW_AT_location or
15444 DW_AT_const_value, or refers to a string literal, into
15445 DW_OP_GNU_implicit_pointer.
15446 (optimize_location_into_implicit_ptr): New function.
15447 (resolve_addr): If removing DW_AT_location of a variable because
15448 it was DW_OP_addr of address of the variable, but the variable doesn't
15449 live in memory, try to emit const value attribute for the initializer.
15451 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15453 * tree.h (VECTOR_TYPE_P): New macro.
15454 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
15455 TYPE_MODE): Use it.
15456 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
15457 VEC_COND_EXPR cannot be lvalues.
15458 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
15460 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15462 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
15463 Restrict the transformation to equal modes.
15465 2013-03-21 Richard Biener <rguenther@suse.de>
15467 PR tree-optimization/39326
15468 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
15469 (MEM_ANALYZABLE): Adjust.
15470 (record_mem_ref_loc): Move bitmap ops ...
15471 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
15472 unanalyzable refs, do not record locations for it.
15473 (analyze_memory_references): Allocate ref zero as shared
15475 (refs_independent_p): Do not test for unanalyzed mems here.
15476 (ref_indep_loop_p_1): Special-case disambiguation against
15477 the unanalyzed ref.
15478 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
15480 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
15482 * config/arm/arm-protos.h (tune_params): Add
15483 prefer_neon_for_64bits field.
15484 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
15485 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
15486 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
15487 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
15488 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
15489 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
15490 (arm_option_override): Handle -mneon-for-64bits new option.
15491 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
15492 (prefer_neon_for_64bits): Declare new variable.
15493 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
15494 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
15495 (arch_enabled): Handle new arch types. Remove support for onlya8
15497 (one_cmpldi2): Use new arch names.
15498 (zero_extend<mode>di2, extend<mode>di2): Ditto.
15499 * config/arm/arm.opt (mneon-for-64bits): Add option.
15500 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
15501 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
15502 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
15504 * doc/invoke.texi (-mneon-for-64bits): Document.
15506 2013-03-21 Richard Biener <rguenther@suse.de>
15508 PR tree-optimization/39326
15509 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
15510 (sort_bbs_in_loop_postorder_cmp): New function.
15511 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
15513 2013-03-21 Richard Biener <rguenther@suse.de>
15515 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
15516 (vect_insert_into_interleaving_chain): Likewise.
15517 (vect_drs_dependent_in_basic_block): Inline ...
15518 (vect_slp_analyze_data_ref_dependence): ... here. New function,
15520 (vect_analyze_data_ref_dependence): ... here. Simplify.
15521 (vect_check_interleaving): Simplify.
15522 (vect_analyze_data_ref_dependences): Likewise. Split out ...
15523 (vect_slp_analyze_data_ref_dependences): ... this new function.
15524 (dr_group_sort_cmp): New function.
15525 (vect_analyze_data_ref_accesses): Compute data-reference groups
15526 here instead of in vect_analyze_data_ref_dependence. Use
15527 a more efficient algorithm.
15528 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
15529 vect_slp_analyze_data_ref_dependences. Call
15530 vect_analyze_data_ref_accesses earlier.
15531 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
15532 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
15533 (vect_slp_analyze_data_ref_dependences): New prototype.
15535 2013-03-21 Richard Biener <rguenther@suse.de>
15537 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
15538 ref is stored in the loop.
15539 (find_refs_for_sm): Walk only over all stores.
15540 (store_motion_loop): Allocate from lim_bitmap_obstack.
15541 (store_motion): Likewise.
15543 2013-03-21 Richard Biener <rguenther@suse.de>
15545 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
15546 Update virtual SSA form.
15548 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15550 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
15551 * configure: Regenerate.
15552 * config.in: Regenerate.
15553 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
15554 if !HAVE_LD_EH_FRAME_CIEV3.
15556 2013-03-21 Richard Biener <rguenther@suse.de>
15558 * tree-cfg.c (verify_expr_no_block): New function.
15559 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
15560 nor DECL_VALUE_EXPR have locations with associated blocks.
15561 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
15562 (clear_unused_block_pointer): Remove code dealing with
15563 blocks in DECL_DEBUG_EXPR locations.
15565 2013-03-21 Richard Biener <rguenther@suse.de>
15567 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
15568 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
15569 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
15570 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
15571 instead of DECL_DEBUG_EXPR_IS_FROM.
15572 * gimplify.c (gimplify_modify_expr): Likewise.
15573 * tree-cfg.c (verify_expr_location_1): Likewise.
15574 * tree-complex.c (create_one_component_var): Likewise.
15575 * tree-sra.c (create_access_replacement): Likewise.
15576 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
15577 (clear_unused_block_pointer): Likewise.
15578 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
15579 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
15580 * var-tracking.c (var_debug_decl): Likewise.
15581 (track_expr_p): Likewise.
15582 * tree-inline.c (add_local_variables): Likewise. Set
15583 DECL_HAS_DEBUG_EXPR_P after copying it.
15584 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
15585 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
15587 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
15590 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
15591 * configure: Regenerate.
15592 * config.in: Regenerate.
15593 * config/i386/i386.md (*movdf_internal): Use
15594 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15595 movd instead of movq mnemonic for interunit moves.
15596 (*movdi_internal): Ditto.
15598 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15600 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
15601 (abd<mode>_3): New pattern.
15602 (aba<mode>_3): New pattern.
15603 (fabd<mode>_3): New pattern.
15605 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15607 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
15608 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
15609 occurrence of REGISTER_PREFIX as its empty string.
15611 2013-03-20 Jeff Law <law@redhat.com>
15613 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
15614 addititional equivalences for equality comparisons between an SSA_NAME
15615 and a constant where the SSA_NAME was set from a widening conversion.
15617 2013-03-20 Walter Lee <walt@tilera.com>
15619 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
15621 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15623 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
15624 depending on type attribute.
15625 (*movti_internal): Ditto.
15626 (*movtf_internal): Ditto.
15627 (*movxf_internal): Ditto.
15628 (*movdf_internal): Ditto.
15629 (*movsf_internal): Ditto.
15631 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15633 * config/i386/i386.md (*movti_internal): Set prefix attribute to
15634 maybe_vex for sselog1 and ssemov types.
15635 (*movdi_internal): Reorder operand constraints.
15636 (*movsi_internal): Ditto. Set prefix attribute to
15637 maybe_vex for sselog1 and ssemov types.
15638 (*movtf_internal): Set prefix attribute to maybe_vex
15639 for sselog1 and ssemov types.
15640 (*movdf_internal): Ditto. Set prefix_data16 attribute for
15641 DImode ssemov types. Reorder operand constraints.
15642 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
15643 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
15644 attribute for SImode ssemov types. Reorder operand constraints.
15646 2013-03-20 Martin Jambor <mjambor@suse.cz>
15648 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
15649 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
15651 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
15653 * config/rs6000/predicates.md (indexed_address, update_address_mem
15654 update_indexed_address_mem): New predicates.
15655 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
15656 attribute for load/store instructions.
15657 * config/rs6000/dfp.md (movsd_store): Likewise.
15658 (movsd_load): Likewise.
15659 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15660 (unnamed HI->DI extend define_insn): Likewise.
15661 (unnamed SI->DI extend define_insn): Likewise.
15662 (unnamed QI->SI extend define_insn): Likewise.
15663 (unnamed QI->HI extend define_insn): Likewise.
15664 (unnamed HI->SI extend define_insn): Likewise.
15665 (unnamed HI->SI extend define_insn): Likewise.
15666 (extendsfdf2_fpr): Likewise.
15667 (movsi_internal1): Likewise.
15668 (movsi_internal1_single): Likewise.
15669 (movhi_internal): Likewise.
15670 (movqi_internal): Likewise.
15671 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15672 attribute for load/store instructions.
15673 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15675 (mov<mode>_softfloat): Likewise.
15676 (mov<mode>_hardfloat32): Likewise.
15677 (mov<mode>_hardfloat64): Likewise.
15678 (mov<mode>_softfloat64): Likewise.
15679 (movdi_internal32): Likewise.
15680 (movdi_internal64): Likewise.
15681 (probe_stack_<mode>): Likewise.
15683 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15685 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15686 floating point, and decimal floating point to reload iterator.
15688 * config/rs6000/constraints.md (wl constraint): New constraints to
15689 return FLOAT_REGS if certain options are used to reduce the number
15690 of separate patterns that exist in the file.
15691 (wx constraint): Likewise.
15692 (wz constraint): Likewise.
15694 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15695 -mdebug=reg, print wg, wl, wx, and wz constraints.
15696 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15697 Initialize the reload functions for 64-bit binary/decimal floating
15699 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15700 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15701 create the buffer on the stack to overcome not having a 32-bit
15703 (rs6000_emit_move): Likewise.
15704 (rs6000_secondary_memory_needed_rtx): Likewise.
15705 (rs6000_alloc_sdmode_stack_slot): Likewise.
15706 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15707 via xxlxor, just like DFmode 0.0.
15709 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15710 define as 1 if we are running on a power7 or newer.
15711 (enum r6000_reg_class_enum): Add new constraints.
15713 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15714 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15715 with other moves by using conditional constraits (wg). Use LFIWZX
15716 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15717 (movsd splitter): Likewise.
15718 (movsd_hardfloat): Likewise.
15719 (movsd_softfloat): Likewise.
15721 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15722 binary and decimal floating point moves.
15723 (fmove_ok): New attributes to combine binary and decimal floating
15724 point moves, and to combine power6x (mfpgpr) moves along normal
15726 (real_value_to_target): Likewise.
15727 (f32_lr): Likewise.
15728 (f32_lm): Likewise.
15729 (f32_li): Likewise.
15730 (f32_sr): Likewise.
15731 (f32_sm): Likewise.
15732 (f32_si): Likewise.
15733 (movsf): Combine binary and decimal floating point moves. Combine
15734 power6x (mfpgpr) moves with other moves by using conditional
15735 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15736 (mov<mode> for SFmode/SDmode); Likewise.
15737 (SFmode/SDmode splitters): Likewise.
15738 (movsf_hardfloat): Likewise.
15739 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15740 (movsf_softfloat): Likewise.
15741 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15743 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15744 wx and wz constraints.
15746 * config/rs6000/constraints.md (wg constraint): New constraint to
15747 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15749 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15752 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15753 -mdebug=reg, print wg, wl, wx, and wz constraints.
15754 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15755 Initialize the reload functions for 64-bit binary/decimal floating
15757 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15758 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15759 create the buffer on the stack to overcome not having a 32-bit
15761 (rs6000_emit_move): Likewise.
15762 (rs6000_secondary_memory_needed_rtx): Likewise.
15763 (rs6000_alloc_sdmode_stack_slot): Likewise.
15764 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15765 via xxlxor, just like DFmode 0.0.
15767 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15768 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15769 with other moves by using conditional constraits (wg). Use LFIWZX
15770 and STFIWX for loading SDmode on power7.
15771 (movdd splitters): Likewise.
15772 (movdd_hardfloat32): Likewise.
15773 (movdd_softfloat32): Likewise.
15774 (movdd_hardfloat64_mfpgpr): Likewise.
15775 (movdd_hardfloat64): Likewise.
15776 (movdd_softfloat64): Likewise.
15778 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15779 64-bit binary and decimal floating point moves.
15780 (FMOVE64X): Likewise.
15781 (movdf): Combine 64-bit binary and decimal floating point moves.
15782 Combine power6x (mfpgpr) moves with other moves by using
15783 conditional constraits (wg).
15784 (mov<mode> for DFmode/DDmode): Likewise.
15785 (DFmode/DDmode splitters): Likewise.
15786 (movdf_hardfloat32): Likewise.
15787 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15788 (movdf_softfloat32): Likewise.
15789 (movdf_hardfloat64_mfpgpr): Likewise.
15790 (movdf_hardfloat64): Likewise.
15791 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15792 (movdf_softfloat64): Likewise.
15793 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15794 (reload_<mode>_load): Move to later in the file so they aren't in
15795 the middle of the floating point move insns.
15796 (reload_<mode>_store): Likewise.
15798 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15801 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15802 constraint if -mdebug=reg.
15803 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15804 Enable using dd reload support if needed.
15806 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15807 binary and decimal floating point moves in rs6000.md.
15808 (movtd_internal): Likewise.
15810 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15811 decimal floating point moves.
15813 (movtf_internal): Likewise.
15814 (mov<mode>_internal, TDmode/TFmode): Likewise.
15815 (movtf_softfloat): Likewise.
15816 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15818 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15819 movdi_internal64, using wg constraint for move direct operations.
15820 (movdi_internal64): Likewise.
15822 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15823 MODES_TIEABLE_P for selected modes. Print the numerical value of
15824 the various virtual registers. Use GPR/FPR first/last values,
15825 instead of hard coding the register numbers. Print which modes
15826 have reload functions registered.
15827 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15828 settings before/after setting cpu, target and subtarget settings.
15829 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15830 and for secondary reload failures in rs6000_secondary_reload_inner.
15831 (rs6000_secondary_reload_fail): Likewise.
15832 (rs6000_secondary_reload_inner): Likewise.
15834 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15835 macros for first/last GPR and FPR registers.
15836 (LAST_GPR_REGNO): Likewise.
15837 (FIRST_FPR_REGNO): Likewise.
15838 (LAST_FPR_REGNO): Likewise.
15840 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
15841 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
15842 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
15843 (vcond<mode><mode>): Likewise.
15844 (vcondu<mode><mode>): Likewise.
15845 (vector_gtu<mode>): Likewise.
15846 (vector_gte<mode>): Likewise.
15847 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
15848 to prevent the compiler from converting DImode operations to TImode.
15849 (ior<mode>3): Likewise.
15850 (and<mode>3): Likewise.
15851 (one_cmpl<mode>2): Likewise.
15852 (nor<mode>3): Likewise.
15853 (andc<mode>3): Likewise.
15855 * config/rs6000/constraints.md (wt constraint): New constraint
15856 that returns VSX_REGS if TImode is allowed in VSX registers.
15858 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
15859 constant under VSX.
15861 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
15862 similar to TImode, but it is restricted to being in the GPRs.
15864 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
15865 TImode to occupy a single VSX register.
15867 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
15868 -mvsx-timode for power7/power8.
15869 (power7 cpu): Likewise.
15870 (power8 cpu): Likewise.
15872 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
15873 sure that TFmode/TDmode take up two registers if they are ever
15874 allowed in the upper VSX registers.
15875 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
15877 (rs6000_init_hard_regno_mode_ok): Likewise.
15878 (rs6000_debug_reg_global): Add debugging for PTImode and wt
15879 constraint. Print if LRA is turned on.
15880 (rs6000_option_override_internal): Give an error if -mvsx-timode
15881 and VSX is not enabled.
15882 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
15883 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
15884 to reg+offset addressing. Use PTImode when checking offset
15885 addresses for validity.
15886 (reg_offset_addressing_ok_p): Likewise.
15887 (rs6000_legitimate_offset_address_p): Likewise.
15888 (rs6000_legitimize_address): Likewise.
15889 (rs6000_legitimize_reload_address): Likewise.
15890 (rs6000_legitimate_address_p): Likewise.
15891 (rs6000_eliminate_indexed_memrefs): Likewise.
15892 (rs6000_emit_move): Likewise.
15893 (rs6000_secondary_reload): Likewise.
15894 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
15895 reloads to fpr registers to continue to use reg+offset addressing,
15896 but 64-bit reloads to altivec registers need reg+reg addressing.
15897 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
15898 it. Treat LO_SUM like a PLUS operation.
15899 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
15900 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
15901 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
15902 registers to share a register with a smaller sized type, since VSX
15903 puts scalars in the upper 64-bits.
15904 (print_operand): Add support for PTImode.
15905 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
15906 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
15907 registers, but don't have arithmetic support.
15908 (rs6000_memory_move_cost): Add test for VSX.
15909 (rs6000_opt_masks): Add -mvsx-timode.
15911 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
15914 (VSr): Use wt constraint for TImode.
15915 (VSv): Drop TImode support.
15916 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
15917 (vsx_movti_64bit): Likewise.
15918 (vsx_movti_32bit): Likewise.
15919 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
15920 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
15921 one '?' on the appropriate output constraint. Do not allow TImode
15922 logical operations on 32-bit systems.
15923 (vsx_ior<mode>3): Likewise.
15924 (vsx_xor<mode>3): Likewise.
15925 (vsx_one_cmpl<mode>2): Likewise.
15926 (vsx_nor<mode>3): Likewise.
15927 (vsx_andc<mode>3): Likewise.
15928 (vsx_concat_<mode>): Likewise.
15929 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
15931 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
15932 OPTION_MASK_VSX_TIMODE.
15933 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
15934 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
15936 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
15937 (TI2 iterator): New iterator for TImode, PTImode.
15938 (wd mode attribute): Add values for vector types.
15939 (movti_string): Replace TI move operations with operations for TImode
15940 and PTImode. Add support for TImode being allowed in VSX registers.
15941 (mov<mode>_string, TImode/PTImode): Likewise.
15942 (movti_ppc64): Likewise.
15943 (mov<mode>_ppc64, TImode/PTImode): Likewise.
15944 (TI mode splitters): Likewise.
15946 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
15949 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
15951 PR tree-optimization/56355
15952 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
15953 Also handle integers with undefined overflow.
15955 2013-03-20 Catherine Moore <clm@codesourcery.com>
15956 Maciej W. Rozycki <macro@codesourcery.com>
15957 Tom de Vries <tom@codesourcery.com>
15958 Nathan Sidwell <nathan@codesourcery.com>
15959 Iain Sandoe <iain@codesourcery.com>
15960 Nathan Froyd <froydnj@codesourcery.com>
15961 Chao-ying Fu <fu@mips.com>
15963 * doc/extend.texi: (micromips, nomicromips, nocompression):
15964 Document new function attributes.
15965 * doc/invoke.texi (minterlink-compressed, mmicromips,
15966 m14k, m14ke, m14kec): Document new options.
15967 (minterlink-mips16): Update documentation.
15968 * doc/md.texi (ZC, ZD): Document new constraints.
15969 * configure.ac (gcc_cv_as_micromips): Check if linker
15970 supports the .set micromips directive.
15971 * configure: Regenerate.
15972 * config.in: Regenerate.
15973 * config/mips/mips-tables.opt: Regenerate.
15974 * config/mips/micromips.md: New file.
15975 * constraints.md (ZC, ZD): New constraints.
15976 * config/mips/predicates.md (movep_src_register): New predicate.
15977 (movep_src_operand): New predicate.
15978 (non_volatile_mem_operand): New predicate.
15979 * config/mips/mips.md (multimem): New type.
15980 (length): Differentiate between 17-bit and 18-bit branch offsets.
15981 (MOVEP1, MOVEP2): New mode iterator.
15982 (mov_<load>l): Use ZC constraint.
15983 (mov_<load>r): Likewise.
15984 (mov_<store>l): Likewise.
15985 (mov_<store>r): Likewise.
15986 (*branch_equality<mode>_inverted): Add microMIPS support.
15987 (*branch_equality<mode>): Likewise.
15988 (*jump_absolute): Likewise.
15989 (indirect_jump_<mode>): Likewise.
15990 (tablejump_<mode>): Likewise.
15991 (<optab>_internal): Likewise.
15992 (sibcall_internal): Likewise.
15993 (sibcall_value_internal): Likewise.
15994 (prefetch): Use constraint ZD.
15995 * config/mips/mips.opt (minterlink-compressed): New option.
15996 (minterlink-mips16): Now an alias for minterlink-compressed.
15997 (mmicromips): New option.
15998 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15999 (compare_and_swap_12): Likewise.
16000 (sync_add<mode>): Likewise.
16001 (sync_<optab>_12): Likewise.
16002 (sync_old_<optab>_12): Likewise.
16003 (sync_new_<optab>_12): Likewise.
16004 (sync_nand_12): Likewise.
16005 (sync_old_nand_12): Likewise.
16006 (sync_new_nand_12): Likewise.
16007 (sync_sub<mode>): Likewise.
16008 (sync_old_add<mode>): Likewise.
16009 (sync_old_sub<mode>): Likewise.
16010 (sync_new_add<mode>): Likewise.
16011 (sync_new_sub<mode>): Likewise.
16012 (sync_<optab><mode>): Likewise.
16013 (sync_old_<optab><mode>): Likewise.
16014 (sync_new_<optab><mode>): Likewise.
16015 (sync_nand<mode>): Likewise.
16016 (sync_old_nand<mode>): Likewise.
16017 (sync_new_nand<mode>): Likewise.
16018 (sync_lock_test_and_set<mode>): Likewise.
16019 (test_and_set_12): Likewise.
16020 (atomic_compare_and_swap<mode>): Likewise.
16021 (atomic_exchange<mode>_llsc): Likewise.
16022 (atomic_fetch_add<mode>_llsc): Likewise.
16023 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16024 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16025 (umips_save_restore_pattern_p): Likewise.
16026 (umips_load_store_pair_p): Likewise.
16027 (umips_output_load_store_pair): Likewise.
16028 (umips_movep_target_p): Likewise.
16029 (umips_12bit_offset_address_p): Likewise.
16030 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16031 (mips_base_mips16): Rename this...
16032 (mips_base_compression_flags): ...to this. Update all uses.
16033 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16034 (mips_mips16_decl_p): Delete.
16035 (mips_nomips16_decl_p): Delete.
16036 (mips_get_compress_on_flags): New function.
16037 (mips_get_compress_off_flags): New function.
16038 (mips_get_compress_mode): New function.
16039 (mips_get_compress_on_name): New function.
16040 (mips_get_compress_off_name): New function.
16041 (mips_insert_attributes): Support multiple compression types.
16042 (mips_merge_decl_attributes): Likewise.
16043 (umips_12bit_offset_address_p): New function.
16044 (mips_start_function_definition): Emit .set micromips directive.
16045 (mips_call_may_need_jalx_p): New function.
16046 (mips_function_ok_for_sibcall): Add microMIPS support.
16047 (mips_print_operand_punctuation): Support short delay slots and
16049 (umips_swm_mask, umips_swm_encoding): New.
16050 (umips_build_save_restore): New function.
16051 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16052 (was_mips16_p): Remove.
16053 (old_compression_mode): New.
16054 (mips_set_compression_mode): New function.
16055 (mips_set_current_function): Add microMIPS support.
16056 (mips_option_override): Likewise.
16057 (umips_save_restore_pattern_p): New function.
16058 (umips_output_save_restore): New function.
16059 (umips_load_store_pair_p_1): New function.
16060 (umips_load_store_pair_p): New function.
16061 (umips_output_load_store_pair_1): New function.
16062 (umips_output_load_store_pair): New function.
16063 (umips_movep_target_p) New function.
16064 (mips_prepare_pch_save): Add microMIPS support.
16065 * config/mips/mips.h (TARGET_COMPRESSION): New.
16066 (TARGET_CPU_CPP_BUILTINS): Update macro
16067 to use new compression flags and to support microMIPS.
16068 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16069 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16070 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16071 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16072 (ASM_SPEC): Support mmicromips and mno-micromips.
16073 (M16STORE_REG_P): New macro.
16074 (MIPS_CALL): Support TARGET_MICROMIPS.
16075 (MICROMIPS_J): New macro.
16076 (mips_base_mips16): Rename this...
16077 (mips_base_compression_flags): ...to this.
16078 (UMIPS_12BIT_OFFSET_P): New macro.
16079 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16080 (MULTILIB_DIRNAMES): Likewise.
16081 2013-03-20 Richard Biener <rguenther@suse.de>
16083 PR tree-optimization/56661
16084 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
16085 the result does not have to be distinct.
16087 2013-03-20 Richard Biener <rguenther@suse.de>
16089 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
16092 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16093 Steven Bosscher <steven@gcc.gnu.org>
16095 PR rtl-optimization/56605
16096 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
16098 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16101 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
16102 that require movd instead of movq.
16104 2013-03-20 Richard Biener <rguenther@suse.de>
16106 * tree-ssa-structalias.c (struct variable_info): Add pointer
16107 to the first field of an aggregate with sub-vars. Make
16108 this and the pointer to the next subfield its ID.
16109 (vi_next): New function.
16110 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
16111 storedanything_id, integer_id): Increment by one.
16112 (new_var_info, get_call_vi, lookup_call_clobber_vi,
16113 get_call_clobber_vi): Adjust.
16114 (solution_set_expand): Simplify and speedup.
16115 (solution_set_add): Inline into ...
16116 (set_union_with_increment): ... this. Adjust accordingly.
16117 (do_sd_constraint): Likewise.
16118 (do_ds_constraint): Likewise.
16119 (do_complex_constraint): Simplify.
16120 (build_pred_graph): Adjust.
16121 (solve_graph): Likewise. Simplify and speedup.
16122 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
16123 get_constraint_for_component_ref, get_constraint_for_1,
16124 first_vi_for_offset, first_or_preceding_vi_for_offset,
16125 create_function_info_for, create_variable_info_for_1,
16126 create_variable_info_for, intra_create_variable_infos): Adjust.
16127 (init_base_vars): Push NULL for ID zero.
16128 (compute_points_to_sets): Adjust.
16130 2013-03-20 Richard Biener <rguenther@suse.de>
16132 * cfgloop.c (verify_loop_structure): Streamline and avoid
16133 ICEing on corrupt loop tree.
16134 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
16137 2013-03-20 Richard Biener <rguenther@suse.de>
16139 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
16140 check whether an SSA update is needed.
16142 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
16144 * config/mips/constraints.md (T): Rename to...
16148 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
16149 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
16151 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16153 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
16154 (*subsi3_carryin_uxtw): Likewise.
16156 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16158 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
16159 (*rorsi3_insn_uxtw): Likewise.
16161 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16163 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
16164 (*extrsi5_insn_uxtw): Likewise.
16166 2013-03-19 Richard Biener <rguenther@suse.de>
16168 PR tree-optimization/56273
16169 * passes.c (init_optimization_passes): Move second VRP after DOM.
16171 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
16173 * config/i386/i386.md (*movti_internal): Merge from
16174 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
16175 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
16176 nox64 isa attributes.
16178 2013-03-18 Richard Biener <rguenther@suse.de>
16180 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
16182 (merge_node_constraints): Likewise.
16183 (build_succ_graph): Likewise.
16184 (valid_graph_edge): Inline into single caller.
16185 (unify_nodes): Likewise. Use bitmap_set_bit return value
16187 (scc_visit): Fix formatting and variable use.
16188 (do_sd_constraint): Use gcc_checking_assert.
16189 (do_ds_constraint): Likewise.
16190 (do_complex_constraint): Likewise.
16191 (condense_visit): Likewise. Cleanup.
16192 (dump_pred_graph): New function.
16193 (perform_var_substitution): Dump the pred-graph before
16194 variable substitution.
16195 (find_equivalent_node): Use gcc_checking_assert.
16196 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
16198 2013-03-18 Richard Biener <rguenther@suse.de>
16200 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
16201 Remove cond_expr_stmt_list argument and do not gimplify the
16203 (vect_loop_versioning): Adjust.
16204 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
16205 Cleanup to use less temporaries.
16206 (vect_create_data_ref_ptr): Cleanup.
16208 2013-03-18 Jakub Jelinek <jakub@redhat.com>
16210 PR tree-optimization/56635
16211 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
16212 require types_compatible_p types.
16214 2013-03-18 Nick Clifton <nickc@redhat.com>
16216 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
16217 spurious backslash.
16219 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
16220 Add missing line to comment describing function.
16222 2013-03-18 Richard Biener <rguenther@suse.de>
16224 PR tree-optimization/56210
16225 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
16226 Handle string / character search functions.
16227 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
16229 2013-03-18 Richard Biener <rguenther@suse.de>
16231 PR middle-end/56483
16232 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
16233 and implement properly.
16234 * gimple.h (gimple_cond_single_var_p): Remove.
16236 2013-03-18 Richard Biener <rguenther@suse.de>
16238 * tree-data-ref.h (find_data_references_in_loop): Declare.
16239 * tree-data-ref.c (get_references_in_stmt): Use a stack
16240 vector pre-allocated in the callers.
16241 (find_data_references_in_stmt): Adjust.
16242 (graphite_find_data_references_in_stmt): Likewise.
16243 (create_rdg_vertices): Likewise.
16244 (find_data_references_in_loop): Export.
16245 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
16246 Compute dependences here...
16247 (vect_analyze_data_refs): ...not here. When we encounter
16248 a non-vectorizable data reference in basic-block vectorization
16249 truncate the data reference vector. Do not bother to
16250 fixup data-dependence information for gather loads.
16251 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
16252 of data references, as reported.
16254 2013-03-18 Richard Biener <rguenther@suse.de>
16256 PR tree-optimization/3713
16257 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
16258 has_constants and expr.
16259 (stmt_has_constants): Properly valueize SSA names when deciding
16260 whether the stmt has constants.
16262 2013-03-18 Richard Biener <rguenther@suse.de>
16264 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
16265 whole function when there is nothing to do.
16266 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
16267 * tree-vectorizer.c (vectorize_loops): Update virtual and
16268 loop-closed SSA once.
16269 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
16271 2013-03-18 Richard Biener <rguenther@suse.de>
16273 PR middle-end/56113
16274 * domwalk.c (bb_postorder): New global static.
16275 (cmp_bb_postorder): New function.
16276 (walk_dominator_tree): Replace scheme imposing an order for
16277 visiting dominator sons by one sorting them at the time they
16278 are pushed on the stack.
16280 2013-03-18 Richard Biener <rguenther@suse.de>
16282 PR tree-optimization/39326
16283 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
16284 (struct mem_ref): Replace mem member with ao_ref typed member.
16285 (MEM_ANALYZABLE): Adjust.
16286 (memref_eq): Likewise.
16287 (mem_ref_alloc): Likewise.
16288 (gather_mem_refs_stmt): Likewise.
16289 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
16290 (execute_sm_if_changed_flag_set): Adjust.
16291 (execute_sm): Likewise.
16292 (ref_always_accessed_p): Likewise.
16293 (refs_independent_p): Likewise.
16294 (can_sm_ref_p): Likewise.
16296 2013-03-18 Jakub Jelinek <jakub@redhat.com>
16299 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
16300 return 1 even for !unsignedp.
16302 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16304 * config/i386/i386.md (isa): Add x64 and nox64.
16305 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
16306 (*pushtf): Enable *roF alternative for x64 isa only.
16307 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
16308 mode attribute of integer alternatives to DImode for TARGET_64BIT.
16309 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
16310 (*movtf_internal): Merge from *movtf_internal_rex64 and
16311 *movtf_internal_sse. Use x64 and nox64 isa attributes.
16312 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
16313 nox64 isa attributes.
16314 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
16315 nox64 isa attributes.
16316 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
16318 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16320 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16322 2013-03-17 Jakub Jelinek <jakub@redhat.com>
16325 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
16326 class names. Remove trailing comma after "ALL_REGS".
16328 2013-03-16 Jan Hubicka <jh@suse.cz>
16330 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
16331 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
16332 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
16333 of cgraph_get_create_node.
16334 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
16336 2013-03-16 Jason Merrill <jason@redhat.com>
16339 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
16340 with DW_AT_default_value.
16342 2013-03-16 Jakub Jelinek <jakub@redhat.com>
16344 * BASE-VER: Set to 4.9.0.
16346 2013-03-14 Andi Kleen <ak@linux.intel.com>
16349 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
16350 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
16351 Document _x* TSX intrinsics.
16353 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
16354 David Holsgrove <david.holsgrove@xilinx.com>
16356 * configure.ac: Add MicroBlaze TLS support detection.
16357 * configure: Regenerate.
16358 * config/microblaze/microblaze-protos.h
16359 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
16360 symbol_mentioned_p, label_mentioned_p): Add prototypes.
16361 * config/microblaze/microblaze.c (microblaze_address_type): Add
16362 ADDRESS_TLS and tls_reloc address types.
16363 (microblaze_address_info): Add tls_reloc.
16364 (TARGET_HAVE_TLS): Define.
16365 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
16366 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
16367 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
16368 load_tls_operand, microblaze_call_tls_get_addr,
16369 microblaze_legitimize_tls_address): New functions.
16370 (microblaze_classify_unspec): Handle UNSPEC_TLS.
16371 (get_base_reg): Use microblaze_tls_symbol_p.
16372 (microblaze_classify_address): Handle TLS.
16373 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
16374 label_mentioned_p and microblaze_tls_referenced_p.
16375 (microblaze_legitimize_address): Handle TLS.
16376 (microblaze_address_insns): Handle ADDRESS_TLS.
16377 (pic_address_needs_scratch): Handle TLS.
16378 (print_operand_address): Handle TLS.
16379 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
16380 (microblaze_expand_move): Handle TLS.
16381 (microblaze_legitimate_constant_p): Check
16382 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
16383 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
16384 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
16385 (PIC_OFFSET_TABLE_REGNUM): Set.
16386 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
16387 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
16388 (addsi3, movsi_internal2, movdf_internal): Update constraints
16389 * config/microblaze/predicates.md (arith_plus_operand): Define
16390 (move_operand): Redefine as move_src_operand,
16391 check microblaze_tls_referenced_p.
16393 2013-03-14 Ian Bolton <ian.bolton@arm.com>
16395 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
16396 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
16398 2013-03-14 Ian Bolton <ian.bolton@arm.com>
16400 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
16403 2013-03-14 Jakub Jelinek <jakub@redhat.com>
16405 PR tree-optimization/53265
16406 * common.opt (Waggressive-loop-optimizations): New option.
16407 * tree-ssa-loop-niter.c: Include tree-pass.h.
16408 (do_warn_aggressive_loop_optimizations): New function.
16409 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
16410 if number_of_latch_executions returned constant.
16411 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
16412 early. If number_of_latch_executions returned constant, set
16413 nb_iterations_upper_bound back to it.
16414 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
16416 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
16417 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
16419 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
16420 (MULTILIB_OSDIRNAMES): Set.
16421 * genmultilib: If defaultosdirname doesn't start with :: , set
16422 defaultosdirname2 instead, clear it and emit two . multilib_raw
16423 entries instead of just one.
16425 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
16427 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16428 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
16429 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16430 (SUBTARGET_OVERRIDE_OPTIONS): New.
16432 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
16435 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
16436 (musermode): Convert to Var(TARGET_USERMODE).
16437 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
16438 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
16439 * config/sh/sh.c (sh_option_override): Use
16440 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
16441 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
16443 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
16445 (udivsi3_i4_single, divsi3_i4_single): Use
16446 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
16448 2013-03-13 Dave Korn <dave.korn.cygwin@....>
16450 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
16453 2013-03-13 Richard Biener <rguenther@suse.de>
16455 PR tree-optimization/56608
16456 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
16457 calls when vectorizing basic-blocks.
16459 2013-03-13 Jakub Jelinek <jakub@redhat.com>
16462 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
16465 2013-03-12 Jakub Jelinek <jakub@redhat.com>
16467 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
16469 2013-03-11 Jan Hubicka <jh@suse.cz>
16472 * lto-streamer-out.c (output_symbol_p): Skip references from
16473 constructors of external variables.
16475 2013-03-11 Jan Hubicka <jh@suse.cz>
16477 PR middle-end/56571
16478 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
16480 * emit-rtl.c (verify_rtx_sharing): Likewise.
16481 (copy_insn_1): Likewise.
16482 * rtl.c (copy_rtx): Likewise.
16484 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
16487 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
16488 output_operand_lossage message.
16490 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
16493 * arm.c (shift_op): Validate RTL pattern on the fly.
16494 (arm_print_operand, case 'S'): Don't use shift_operator to validate
16497 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16500 * config/pa/pa.md (call_value): Check for calls to powf and direct to
16501 new call patterns that clobber %fr12.
16502 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
16503 split and postreload patterns.
16504 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
16505 registers %fr12 and %fr12R as call used.
16507 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
16509 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
16510 (canon_address, record_store, replace_read, check_mem_read_rtx,
16511 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
16512 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
16513 rest_of_handle_dse): Likewise.
16515 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
16517 PR middle-end/56524
16518 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
16520 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
16521 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
16522 (save_optabs_if_changed): Replace with...
16523 (init_tree_optimization_optabs): ...this.
16524 * optabs.c (save_optabs_if_changed): Rename to...
16525 (init_tree_optimization_optabs): ...this. Take the optimization node
16526 as argument. Do nothing if the base optabs are already correct.
16527 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
16528 to recompute optabs.
16529 * function.h (function): Remove optabs field.
16530 * function.c (invoke_set_current_function_hook): Call
16531 init_tree_optimization_optabs. Use the result to initialize
16534 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
16536 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
16537 if GTMA_HAS_NO_INSTRUMENTATION.
16538 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
16539 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
16540 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
16541 * gimple-pretty-print.c (dump_gimple_transaction): Handle
16542 GTMA_HAS_NO_INSTRUMENTATION.
16544 2013-03-08 Jakub Jelinek <jakub@redhat.com>
16546 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
16549 2013-03-08 Marek Polacek <polacek@redhat.com>
16550 Jakub Jelinek <jakub@redhat.com>
16552 PR tree-optimization/56478
16553 * predict.c (is_comparison_with_loop_invariant_p): Change the
16554 type of loop_step to tree.
16555 (predict_loops): Adjust.
16556 (predict_iv_comparison): Perform the computations on double_ints.
16558 2013-03-08 Richard Biener <rguenther@suse.de>
16560 PR tree-optimization/56570
16561 * tree-cfg.c (verify_expr_location_1): Verify locations for
16563 * tree-sra.c (create_access_replacement): Strip locations
16564 from DECL_DEBUG_EXPRs.
16566 2013-03-08 Richard Biener <rguenther@suse.de>
16568 * tree-inline.c (expand_call_inline): Do not associate
16569 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
16570 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
16572 2013-03-08 Richard Biener <rguenther@suse.de>
16574 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
16575 or block changes with -Og. Fix for location / block encoding
16576 changes and PHI arguments with locations.
16578 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
16580 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
16582 (struct output_info): Likewise.
16583 (register_overhead): Remove bad gcc_assert.
16584 (bitmap_find_bit): If there is only a single bitmap element, do not
16585 count a miss as a search.
16586 (print_statistics): Update for counter type changes.
16587 (dump_bitmap_statistics): Likewise. Print headers such that they
16588 are properly lined up with the printed counters.
16590 2013-03-07 Jakub Jelinek <jakub@redhat.com>
16592 PR tree-optimization/56559
16593 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
16594 check that it has only a single use.
16596 2013-03-07 Richard Biener <rguenther@suse.de>
16598 * doc/invoke.texi (fwhole-program): Discourage use in combination
16601 2013-03-06 Jakub Jelinek <jakub@redhat.com>
16603 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
16605 PR tree-optimization/56539
16606 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
16607 instead of GSI_CONTINUE_LINKING as last argument to
16608 force_gimple_operand_gsi. Adjust function comment.
16610 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
16613 PR middle-end/56548
16614 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
16615 promoted mode, convert the result back to the original mode.
16617 2013-03-06 Richard Biener <rguenther@suse.de>
16619 PR middle-end/56294
16620 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
16621 (insert_updated_phi_nodes_compare_uids): New function.
16622 (update_ssa): Sort symbols_to_rename after UID before
16623 traversing it to insert PHI nodes.
16625 2013-03-06 Richard Biener <rguenther@suse.de>
16627 PR middle-end/50494
16628 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
16629 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
16632 2013-02-13 Richard Biener <rguenther@suse.de>
16635 * varasm.c (output_constant_def_1): Get the decl representing
16636 the constant as argument.
16637 (output_constant_def): Wrap output_constant_def_1.
16638 (make_decl_rtl): Use output_constant_def_1 with the decl
16639 representing the constant.
16640 (build_constant_desc): Optionally re-use a decl already
16641 representing the constant.
16642 (tree_output_constant_def): Adjust.
16644 2013-03-06 Joey Ye <joey.ye@arm.com>
16647 * gcc.c (convert_white_space): New function.
16648 (main): Handles white space in function name.
16650 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
16653 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
16654 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
16655 to SH_DIV_CALL_TABLE for TARGET_SH2.
16656 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16658 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16659 call-table options.
16661 2013-03-05 Sterling Augustine <saugustine@google.com>
16662 Cary Coutant <ccoutant@google.com>
16665 * dwarf2out.c (resolve_addr): Don't call
16666 remove_loc_list_addr_table_entries a second time for the same
16669 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16672 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16673 (avoid_complex_debug_insns): New function.
16674 (expand_debug_locations): Call it.
16676 PR rtl-optimization/56484
16677 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16678 lifetimes of hard registers on small register class machines.
16680 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16682 * config/microblaze/microblaze-protos.h: Rename
16683 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16684 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16686 (microblaze_fast_interrupt_function_p): New function.
16687 (microblaze_is_interrupt_handler): Rename to
16688 microblaze_is_interrupt_variant and add fast_interrupt check.
16689 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16690 (save_restore_insns): Likewise.
16691 (compute_frame_size): Likewise.
16692 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16693 (microblaze_globalize_label): Likewise.
16694 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16695 * config/microblaze/microblaze.md: Use wrapper
16696 microblaze_is_interrupt_variant.
16698 2013-03-05 Kai Tietz <ktietz@redhat.com>
16700 * sdbout.c (sdbout_one_type): Switch to current function's section
16701 supporting cold/hot.
16703 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16705 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16708 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16710 PR middle-end/56461
16711 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16712 if VALGRIND_GET_VBITS is defined, temporarily make object
16713 memory all defined, and restore previous valgrind addressability
16714 and definability afterwards. Free this_object at the end.
16716 PR middle-end/56461
16717 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16718 right before calling lra_create_live_ranges, also call it
16719 when clearing live_p. Only call lra_clear_live_ranges
16720 at the end if live_p.
16722 PR middle-end/56461
16723 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16725 2013-03-05 Richard Biener <rguenther@suse.de>
16727 PR tree-optimization/56521
16728 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16731 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16734 * except.h (remove_unreachable_eh_regions): New prototype.
16735 * except.c (remove_eh_handler_splicer): New function, split out
16736 of remove_eh_handler.
16737 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16738 warning about running it on many EH regions one at a time.
16739 (remove_unreachable_eh_regions_worker): New function, walk the
16740 EH tree in depth-first order and remove non-marked regions.
16741 (remove_unreachable_eh_regions): New function.
16742 * tree-eh.c (mark_reachable_handlers): New function, split out
16743 from remove_unreachable_handlers.
16744 (remove_unreachable_handlers): Use mark_reachable_handlers and
16745 remove_unreachable_eh_regions.
16746 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16747 and remove_unreachable_eh_regions.
16749 2013-03-05 Richard Biener <rguenther@suse.de>
16751 PR middle-end/56525
16752 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16753 not freeing them until the end.
16755 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16757 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16759 2013-03-05 Richard Biener <rguenther@suse.de>
16761 PR tree-optimization/56270
16762 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16763 of loads after scheduling an SLP instance.
16765 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16767 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16769 (check_gcc_parallelize): Run guality.exp as a separate job from
16770 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16771 struct-layout-1.exp with stackalign.exp.
16773 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16775 PR middle-end/56461
16776 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16777 load_index sbitmap even if some bit in it isn't set.
16779 PR middle-end/56461
16780 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16781 (discover_iteration_bound_by_body_walk): Change queues to
16782 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16783 spelling in comment. Call safe_push on queues[bound_index] directly.
16784 Release queues[queue_index] in every iteration unconditionally.
16785 Release bounds vector.
16787 PR middle-end/56461
16788 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16789 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16790 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16793 2013-03-05 Richard Biener <rguenther@suse.de>
16796 * tree-inline.c (remap_blocks_to_null): New function.
16797 (expand_call_inline): When expanding a call stmt without
16798 an associated block inline remap all callee blocks to NULL.
16800 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16802 PR rtl-optimization/56494
16803 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16804 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16805 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16807 PR middle-end/56461
16808 * sel-sched-ir.c (free_sched_pools): Release
16809 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16810 if succs_info_pool.max_top isn't -1.
16813 * opts.c (opts_obstack, opts_concat): Moved to...
16814 * opts-common.c (opts_obstack, opts_concat): ... here.
16816 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16818 PR middle-end/56461
16819 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16821 2013-03-04 Martin Jambor <mjambor@suse.cz>
16823 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16824 all appropriate places.
16826 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16828 PR tree-optimization/56424
16829 * ipa-split.c (split_function): Do not set the RSO flag if result is
16830 not by reference and its type is a register type.
16832 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16834 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16835 (microblaze_legitimate_pic_operand): Likewise
16836 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16837 new function microblaze_legitimate_pic_operand
16838 * config/microblaze/microblaze-protos.h
16839 (microblaze_legitimate_pic_operand): Declare.
16841 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16843 * config/microblaze/predicates.md (call_insn_simple_operand):
16844 New predicate for supported rtx code types.
16845 * config/microblaze/microblaze.md (call_internal1): Use
16846 call_insn_simple_operand predicate.
16848 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16850 PR middle-end/56461
16851 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
16852 partitions.ordered_remove.
16854 PR middle-end/56461
16855 * tree-vect-stmts.c (vectorizable_conversion): Don't call
16856 vec_oprnds0.create (1) for modifier == NONE.
16858 PR middle-end/56461
16859 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
16860 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
16861 vec_oprnds1 right before pushing anything to it for
16864 PR middle-end/56461
16865 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
16866 set nbbs to 0 instead of having separate code path.
16867 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
16868 instead of false as last argument if returning NULL.
16870 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
16872 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
16873 the attribute is now called "target" instead of "option".
16874 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
16875 * doc/tm.texi.in (Target Attributes): Likewise document the correct
16876 attribute/pragma name for TARGET_OPTION_VALID_P and
16877 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
16878 * doc/tm.texi: Regenerated.
16880 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
16882 * config/microblaze/microblaze.c:
16883 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
16884 * config/microblaze/microblaze.h: Add -mxl-reorder to
16886 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
16887 instructions emitted if TARGET_REORDER.
16888 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
16889 or 0 for -m/-mno case, but initialises as 2 to detect default use case
16892 2013-03-01 Xinliang David Li <davidxl@google.com>
16894 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
16897 2013-03-01 Jakub Jelinek <jakub@redhat.com>
16899 PR middle-end/56461
16900 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
16901 vector even when returning true. Fix up function comment formatting.
16903 PR middle-end/56461
16904 * ira-build.c (ira_loop_nodes_count): New variable.
16905 (create_loop_tree_nodes): Initialize it.
16906 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
16908 PR middle-end/56461
16909 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
16910 method on dr_chain and result_chain.
16911 * tree-vect-stmts.c (vectorizable_store): Only call
16912 result_chain.create if j == 0.
16914 PR middle-end/56461
16915 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
16916 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
16917 before overwriting it.
16919 2013-03-01 Tobias Burnus <burnus@net-b.de>
16921 * doc/extended.texi (C Extensions): Change order in @menu
16923 (Other MIPS Built-in Functions): Move last MIPS entry before
16924 "picoChip Built-in Functions".
16925 (SH Built-in Functions): Move after RX Built-in Functions.
16926 * doc/gcc.texi (Introduction): Change order in @menu
16928 * doc/md.texi (Constraints): Ditto.
16929 * gty.texi (Type Information): Ditto.
16930 (User-provided marking routines for template types): Make
16932 * doc/invoke.texi (AArch64 Options): Move before
16933 "Adapteva Epiphany Options".
16935 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
16936 Jakub Jelinek <jakub@redhat.com>
16939 * asan.c (gate_asan): Lookup no_sanitize_address instead of
16940 no_address_safety_analysis attribute.
16941 * doc/extend.texi (no_address_safety_attribute): Rename to
16942 no_sanitize_address attribute, mention no_address_safety_analysis
16943 attribute as deprecated alias.
16945 2013-02-28 Jakub Jelinek <jakub@redhat.com>
16947 PR middle-end/56461
16948 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
16949 type to vec<vec<tree> > *.
16950 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
16951 to be vec<tree> instead of vec<tree> *, set vec_defs
16952 to vNULL and call vec_defs.create (number_of_vects), adjust other
16954 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
16955 vectorizable_condition): Adjust vect_get_slp_defs callers.
16957 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16959 * config/aarch64/aarch64.c
16960 (aarch64_float_const_representable): Remove unused variable.
16962 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16964 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
16966 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16968 * config/aarch64/aarch64-builtins.c
16969 (aarch64_init_simd_builtins): Make static.
16971 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16973 * config/aarch64/aarch64.c
16974 (aarch64_simd_make_constant): Make static.
16976 2013-02-28 Martin Jambor <mjambor@suse.cz>
16978 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
16979 with no initialization to the RHS of debug statements.
16981 2013-02-28 Martin Jambor <mjambor@suse.cz>
16983 PR tree-optimization/56294
16984 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
16986 (get_access_replacement): Do not call create_access_replacement.
16987 Assert a replacement exists.
16988 (get_repl_default_def_ssa_name): Create the replacement declaration
16991 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16993 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
16994 final_end_function.
16996 2013-02-28 Marek Polacek <polacek@redhat.com>
16998 PR rtl-optimization/56466
16999 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
17000 if we're changing a loop.
17001 (peel_loops_completely): Likewise.
17003 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
17006 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
17008 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17011 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
17012 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
17013 INTX_FTYPE_FX, FX_FTYPE_INTX.
17014 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
17016 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17018 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17019 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
17020 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
17021 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
17022 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
17023 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
17024 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
17025 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
17026 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
17027 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
17028 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
17029 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
17030 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
17031 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
17032 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
17033 (avrxmega6): Increase max flash segments from 5 to 6.
17034 * config/avr/t-multilib: Regenerate.
17035 * config/avr/avr-tables.opt: Regenerate.
17036 * doc/avr-mmcu.texi: Regenerate.
17038 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17040 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
17041 (avr_device_to_arch): Rename to avr_device_to_ld.
17042 (avr_device_to_as): New prototype.
17043 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
17044 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
17045 * config/avr/driver-avr.c (avr_device_to_as): New.
17046 (avr_device_to_arch): Rename to avr_device_to_ld.
17048 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17050 PR middle-end/56461
17051 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
17052 method on dr_chain and result_chain.
17054 PR middle-end/56461
17055 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
17056 pointer_set_destroy on not_executed_last_iteration.
17058 PR middle-end/56461
17059 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
17061 PR middle-end/56461
17062 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
17063 FOR_EACH_DEFINED_FUNCTION when freeing state.
17065 PR middle-end/56461
17066 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
17068 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
17071 PR middle-end/56461
17072 * ipa-cp.c (decide_whether_version_node): Call vec_free on
17073 known_aggs[i].items and release known_aggs vector.
17075 PR middle-end/56461
17076 * ipa-reference.c (propagate): Free node_info even for alias nodes.
17078 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17080 * config/microblaze/microblaze.c (microblaze_emit_compare):
17081 Use xor for EQ/NE comparisions.
17082 * config/microblaze/microblaze.md (cstoresf4): Add constraints
17083 (cbranchsf4): Adjust operator to comparison_operator.
17085 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17087 PR middle-end/56461
17088 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
17090 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
17091 vec_safe_push, always update *slot.
17092 (redirect_edge_var_map_clear): Use vec_free.
17093 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
17094 (free_var_map_entry): Use vec_free.
17095 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
17096 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
17098 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
17100 PR middle-end/45472
17101 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
17102 when the may_trap_p bit of the exprs being merged differs.
17103 Reorder tests for speculativeness in the logical and operator.
17105 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17107 * incpath.c (add_standard_paths): Use reconcat instead of concat
17108 where appropriate and avoid leaking memory.
17110 * opts.h: Include obstack.h.
17111 (opts_concat): New prototype.
17112 (opts_obstack): New declaration.
17113 * opts.c (opts_concat): New function.
17114 (opts_obstack): New variable.
17115 (init_options_struct): Call gcc_init_obstack on opts_obstack.
17116 (finish_options): Use opts_concat instead of concat
17117 and XOBNEWVEC instead of XNEWVEC.
17118 * opts-common.c (generate_canonical_option, decode_cmdline_option,
17119 generate_option): Likewise.
17120 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
17121 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
17124 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
17125 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
17127 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17129 PR middle-end/56461
17130 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
17132 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17134 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
17135 (arm_block_move_unaligned_straight): Likewise.
17136 (arm_adjust_block_mem): Likewise.
17138 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17141 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
17142 temp, cond and label.
17143 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
17146 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
17147 * config/c6x/c6x.h (dbx_register_map): Update declaration.
17150 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
17151 of prologue/epilogue functions.
17154 * config/tilegx/tilegx.c (tilegx_expand_prologue):
17155 Remove unused variable cfa_offset.
17156 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
17159 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
17160 type promotion to unsigned.
17163 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
17164 for HOST_WIDE_INT of 32 bit / same size as int.
17165 (arm_block_move_unaligned_straight): Likewise.
17166 (arm_adjust_block_mem): Likewise.
17169 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
17172 2013-02-26 Marek Polacek <polacek@redhat.com>
17174 PR tree-optimization/56426
17175 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
17177 2013-02-26 Richard Biener <rguenther@suse.de>
17180 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
17181 unused variable loops.
17183 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17185 PR tree-optimization/56448
17186 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
17187 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
17188 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
17189 later operands of the references, or even first operand for
17190 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
17192 PR tree-optimization/56443
17193 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
17194 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
17195 to type_for_mode langhook.
17197 2013-02-25 Matt Turner <mattst88@gmail.com>
17199 * doc/invoke.texi: Document r4700.
17201 2013-02-25 Richard Biener <rguenther@suse.de>
17203 PR tree-optimization/56175
17204 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
17206 (simplify_bitwise_binary): ... here. Also guard the conversion
17207 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
17209 2013-02-25 Catherine Moore <clm@codesourcery.com>
17212 2013-02-24 Catherine Moore <clm@codesourcery.com>
17213 Maciej W. Rozycki <macro@codesourcery.com>
17214 Tom de Vries <tom@codesourcery.com>
17215 Nathan Sidwell <nathan@codesourcery.com>
17216 Iain Sandoe <iain@codesourcery.com>
17217 Nathan Froyd <froydnj@codesourcery.com>
17218 Chao-ying Fu <fu@mips.com>
17220 * doc/extend.texi: (micromips, nomicromips, nocompression):
17221 Document new function attributes.
17222 * doc/invoke.texi (minterlink-compressed, mmicromips,
17223 m14k, m14ke, m14kec): Document new options.
17224 (minterlink-mips16): Update documentation.
17225 * doc/md.texi (ZC, ZD): Document new constraints.
17226 * configure.ac (gcc_cv_as_micromips): Check if linker
17227 supports the .set micromips directive.
17228 * configure: Regenerate.
17229 * config.in: Regenerate.
17230 * config/mips/mips-tables.opt: Regenerate.
17231 * config/mips/micromips.md: New file.
17232 * constraints.md (ZC, AD): New constraints.
17233 * config/mips/predicates.md (movep_src_register): New predicate.
17234 (movep_src_operand): New predicate.
17235 (non_volatile_mem_operand): New predicate.
17236 * config/mips/mips.md (multimem): New type.
17237 (length): Differentiate between 17-bit and 18-bit branch offsets.
17238 (MOVEP1, MOVEP2): New mode iterator.
17239 (mov_<load>l): Use ZC constraint.
17240 (mov_<load>r): Likewise.
17241 (mov_<store>l): Likewise.
17242 (mov_<store>r): Likewise.
17243 (*branch_equality<mode>_inverted): Add microMIPS support.
17244 (*branch_equality<mode>): Likewise.
17245 (*jump_absolute): Likewise.
17246 (indirect_jump_<mode>): Likewise.
17247 (tablejump_<mode>): Likewise.
17248 (<optab>_internal): Likewise.
17249 (sibcall_internal): Likewise.
17250 (sibcall_value_internal): Likewise.
17251 (prefetch): Use constraint ZD.
17252 * config/mips/mips.opt (minterlink-compressed): New option.
17253 (minterlink-mips16): Now an alias for minterlink-compressed.
17254 (mmicromips): New option.
17255 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17256 (compare_and_swap_12): Likewise.
17257 (sync_add<mode>): Likewise.
17258 (sync_<optab>_12): Likewise.
17259 (sync_old_<optab>_12): Likewise.
17260 (sync_new_<optab>_12): Likewise.
17261 (sync_nand_12): Likewise.
17262 (sync_old_nand_12): Likewise.
17263 (sync_new_nand_12): Likewise.
17264 (sync_sub<mode>): Likewise.
17265 (sync_old_add<mode>): Likewise.
17266 (sync_old_sub<mode>): Likewise.
17267 (sync_new_add<mode>): Likewise.
17268 (sync_new_sub<mode>): Likewise.
17269 (sync_<optab><mode>): Likewise.
17270 (sync_old_<optab><mode>): Likewise.
17271 (sync_new_<optab><mode>): Likewise.
17272 (sync_nand<mode>): Likewise.
17273 (sync_old_nand<mode>): Likewise.
17274 (sync_new_nand<mode>): Likewise.
17275 (sync_lock_test_and_set<mode>): Likewise.
17276 (test_and_set_12): Likewise.
17277 (atomic_compare_and_swap<mode>): Likewise.
17278 (atomic_exchange<mode>_llsc): Likewise.
17279 (atomic_fetch_add<mode>_llsc): Likewise.
17280 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17281 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17282 (umips_save_restore_pattern_p): Likewise.
17283 (umips_load_store_pair_p): Likewise.
17284 (umips_output_load_store_pair): Likewise.
17285 (umips_movep_target_p): Likewise.
17286 (umips_12bit_offset_address_p): Likewise.
17287 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17288 (mips_base_mips16): Rename this...
17289 (mips_base_compression_flags): ...to this. Update all uses.
17290 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17291 (mips_mips16_decl_p): Delete.
17292 (mips_nomips16_decl_p): Delete.
17293 (mips_get_compress_on_flags): New function.
17294 (mips_get_compress_off_flags): New function.
17295 (mips_get_compress_mode): New function.
17296 (mips_get_compress_on_name): New function.
17297 (mips_get_compress_off_name): New function.
17298 (mips_insert_attributes): Support multiple compression types.
17299 (mips_merge_decl_attributes): Likewise.
17300 (umips_12bit_offset_address_p): New function.
17301 (mips_start_function_definition): Emit .set micromips directive.
17302 (mips_call_may_need_jalx_p): New function.
17303 (mips_function_ok_for_sibcall): Add microMIPS support.
17304 (mips_print_operand_punctuation): Support short delay slots and
17306 (umips_swm_mask, umips_swm_encoding): New.
17307 (umips_build_save_restore): New function.
17308 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17309 (was_mips16_p): Remove.
17310 (old_compression_mode): New.
17311 (mips_set_compression_mode): New function.
17312 (mips_set_current_function): Add microMIPS support.
17313 (mips_option_override): Likewise.
17314 (umips_save_restore_pattern_p): New function.
17315 (umips_output_save_restore): New function.
17316 (umips_load_store_pair_p_1): New function.
17317 (umips_load_store_pair_p): New function.
17318 (umips_output_load_store_pair_1): New function.
17319 (umips_output_load_store_pair): New function.
17320 (umips_movep_target_p) New function.
17321 (mips_prepare_pch_save): Add microMIPS support.
17322 * config/mips/mips.h (TARGET_COMPRESSION): New.
17323 (TARGET_CPU_CPP_BUILTINS): Update macro
17324 to use new compression flags and to support microMIPS.
17325 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17326 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17327 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17328 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17329 (ASM_SPEC): Support mmicromips and mno-micromips.
17330 (M16STORE_REG_P): New macro.
17331 (MIPS_CALL): Support TARGET_MICROMIPS.
17332 (MICROMIPS_J): New macro.
17333 (mips_base_mips16): Rename this...
17334 (mips_base_compression_flags): ...to this.
17335 (UMIPS_12BIT_OFFSET_P): New macro.
17336 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17337 (MULTILIB_DIRNAMES): Likewise.
17339 2013-02-25 Tom de Vries <tom@codesourcery.com>
17341 PR rtl-optimization/56131
17342 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
17343 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
17344 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
17346 2013-02-25 Tobias Burnus <burnus@net-b.de>
17348 * doc/invoke.texi (-fsanitize=): Move from optimization
17349 to debugging options.
17351 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
17353 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
17355 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
17356 Alexander Monakov <amonakov@ispras.ru>
17358 PR middle-end/56077
17359 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
17360 flush pending lists also on non-jumps. Adjust comment.
17362 2013-02-24 Catherine Moore <clm@codesourcery.com>
17363 Maciej W. Rozycki <macro@codesourcery.com>
17364 Tom de Vries <tom@codesourcery.com>
17365 Nathan Sidwell <nathan@codesourcery.com>
17366 Iain Sandoe <iain@codesourcery.com>
17367 Nathan Froyd <froydnj@codesourcery.com>
17368 Chao-ying Fu <fu@mips.com>
17370 * doc/extend.texi: (micromips, nomicromips, nocompression):
17371 Document new function attributes.
17372 * doc/invoke.texi (minterlink-compressed, mmicromips,
17373 m14k, m14ke, m14kec): Document new options.
17374 (minterlink-mips16): Update documentation.
17375 * doc/md.texi (ZC, ZD): Document new constraints.
17376 * configure.ac (gcc_cv_as_micromips): Check if linker
17377 supports the .set micromips directive.
17378 * configure: Regenerate.
17379 * config.in: Regenerate.
17380 * config/mips/mips-tables.opt: Regenerate.
17381 * config/mips/micromips.md: New file.
17382 * constraints.md (ZC, AD): New constraints.
17383 * config/mips/predicates.md (movep_src_register): New predicate.
17384 (movep_src_operand): New predicate.
17385 (non_volatile_mem_operand): New predicate.
17386 * config/mips/mips.md (multimem): New type.
17387 (length): Differentiate between 17-bit and 18-bit branch offsets.
17388 (MOVEP1, MOVEP2): New mode iterator.
17389 (mov_<load>l): Use ZC constraint.
17390 (mov_<load>r): Likewise.
17391 (mov_<store>l): Likewise.
17392 (mov_<store>r): Likewise.
17393 (*branch_equality<mode>_inverted): Add microMIPS support.
17394 (*branch_equality<mode>): Likewise.
17395 (*jump_absolute): Likewise.
17396 (indirect_jump_<mode>): Likewise.
17397 (tablejump_<mode>): Likewise.
17398 (<optab>_internal): Likewise.
17399 (sibcall_internal): Likewise.
17400 (sibcall_value_internal): Likewise.
17401 (prefetch): Use constraint ZD.
17402 * config/mips/mips.opt (minterlink-compressed): New option.
17403 (minterlink-mips16): Now an alias for minterlink-compressed.
17404 (mmicromips): New option.
17405 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17406 (compare_and_swap_12): Likewise.
17407 (sync_add<mode>): Likewise.
17408 (sync_<optab>_12): Likewise.
17409 (sync_old_<optab>_12): Likewise.
17410 (sync_new_<optab>_12): Likewise.
17411 (sync_nand_12): Likewise.
17412 (sync_old_nand_12): Likewise.
17413 (sync_new_nand_12): Likewise.
17414 (sync_sub<mode>): Likewise.
17415 (sync_old_add<mode>): Likewise.
17416 (sync_old_sub<mode>): Likewise.
17417 (sync_new_add<mode>): Likewise.
17418 (sync_new_sub<mode>): Likewise.
17419 (sync_<optab><mode>): Likewise.
17420 (sync_old_<optab><mode>): Likewise.
17421 (sync_new_<optab><mode>): Likewise.
17422 (sync_nand<mode>): Likewise.
17423 (sync_old_nand<mode>): Likewise.
17424 (sync_new_nand<mode>): Likewise.
17425 (sync_lock_test_and_set<mode>): Likewise.
17426 (test_and_set_12): Likewise.
17427 (atomic_compare_and_swap<mode>): Likewise.
17428 (atomic_exchange<mode>_llsc): Likewise.
17429 (atomic_fetch_add<mode>_llsc): Likewise.
17430 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17431 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17432 (umips_save_restore_pattern_p): Likewise.
17433 (umips_load_store_pair_p): Likewise.
17434 (umips_output_load_store_pair): Likewise.
17435 (umips_movep_target_p): Likewise.
17436 (umips_12bit_offset_address_p): Likewise.
17437 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17438 (mips_base_mips16): Rename this...
17439 (mips_base_compression_flags): ...to this. Update all uses.
17440 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17441 (mips_mips16_decl_p): Delete.
17442 (mips_nomips16_decl_p): Delete.
17443 (mips_get_compress_on_flags): New function.
17444 (mips_get_compress_off_flags): New function.
17445 (mips_get_compress_mode): New function.
17446 (mips_get_compress_on_name): New function.
17447 (mips_get_compress_off_name): New function.
17448 (mips_insert_attributes): Support multiple compression types.
17449 (mips_merge_decl_attributes): Likewise.
17450 (umips_12bit_offset_address_p): New function.
17451 (mips_start_function_definition): Emit .set micromips directive.
17452 (mips_call_may_need_jalx_p): New function.
17453 (mips_function_ok_for_sibcall): Add microMIPS support.
17454 (mips_print_operand_punctuation): Support short delay slots and
17456 (umips_swm_mask, umips_swm_encoding): New.
17457 (umips_build_save_restore): New function.
17458 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17459 (was_mips16_p): Remove.
17460 (old_compression_mode): New.
17461 (mips_set_compression_mode): New function.
17462 (mips_set_current_function): Add microMIPS support.
17463 (mips_option_override): Likewise.
17464 (umips_save_restore_pattern_p): New function.
17465 (umips_output_save_restore): New function.
17466 (umips_load_store_pair_p_1): New function.
17467 (umips_load_store_pair_p): New function.
17468 (umips_output_load_store_pair_1): New function.
17469 (umips_output_load_store_pair): New function.
17470 (umips_movep_target_p) New function.
17471 (mips_prepare_pch_save): Add microMIPS support.
17472 * config/mips/mips.h (TARGET_COMPRESSION): New.
17473 (TARGET_CPU_CPP_BUILTINS): Update macro
17474 to use new compression flags and to support microMIPS.
17475 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17476 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17477 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17478 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17479 (ASM_SPEC): Support mmicromips and mno-micromips.
17480 (M16STORE_REG_P): New macro.
17481 (MIPS_CALL): Support TARGET_MICROMIPS.
17482 (MICROMIPS_J): New macro.
17483 (mips_base_mips16): Rename this...
17484 (mips_base_compression_flags): ...to this.
17485 (UMIPS_12BIT_OFFSET_P): New macro.
17486 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17487 (MULTILIB_DIRNAMES): Likewise.
17489 2013-02-24 Jakub Jelinek <jakub@redhat.com>
17492 * target-globals.c (save_target_globals): For init_reg_sets and
17493 target_reinit remporarily set this_fn_optabs to this_target_optabs.
17495 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
17497 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
17498 * config/aarch64/t-aarch64
17499 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
17501 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
17503 PR inline-asm/56148
17504 * lra-constraints.c (process_alt_operands): Reload operand
17505 conflicting with earlier clobber only if no more other conflicting
17508 2013-02-22 Jakub Jelinek <jakub@redhat.com>
17511 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
17512 if not linking a shared library.
17514 2013-02-22 Seth LaForge <sethml@google.com>
17516 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
17518 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
17520 * config/arm/arm.md (split for extendsidi): Update condition.
17521 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
17522 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
17523 (qhs_zextenddi_cstr): Likewise.
17525 2013-02-21 Jakub Jelinek <jakub@redhat.com>
17527 PR middle-end/56420
17528 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
17529 avoid signed wrapping.
17530 (expand_mult): Handle properly multiplication by
17531 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
17532 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
17533 in the compiler if coeff is HOST_WIDE_INT_MIN.
17534 (expand_divmod): Don't make ext_op1 static, change it's type to
17535 uhwi. Avoid undefined behavior in -INTVAL (op1).
17537 PR rtl-optimization/50339
17538 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
17540 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
17541 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
17542 into splitting_ashiftrt field.
17543 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
17545 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
17548 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
17550 PR middle-end/56108
17551 * trans-mem.c (execute_tm_mark): Do not expand transactions that
17552 are sure to go irrevocable.
17554 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
17556 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
17557 scalars are valid operands.
17559 2013-02-21 Martin Jambor <mjambor@suse.cz>
17561 PR tree-optimization/56310
17562 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
17563 only matching indices and non-negative final offsets.
17564 (intersect_aggregates_with_edge): Pass src_idx to
17565 agg_replacements_to_vector. Pass src_idx insstead of index to
17566 intersect_with_agg_replacements.
17568 2013-02-21 Martin Jambor <mjambor@suse.cz>
17570 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
17571 instead of hard-wired defaults.
17573 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
17575 * doc/invoke.texi (MIPS Options): Update documentation of the
17576 floating-point multiply-accumulate instruction restrictions.
17578 2013-02-21 Kostya Serebryany <kcc@google.com>
17580 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
17581 asan_shadow_offset on x86_64 linux.
17583 2013-02-21 Richard Biener <rguenther@suse.de>
17585 PR tree-optimization/56415
17587 2013-02-11 Richard Biener <rguenther@suse.de>
17589 PR tree-optimization/56273
17590 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17593 2013-02-21 Jakub Jelinek <jakub@redhat.com>
17596 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
17599 PR inline-asm/56405
17600 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
17601 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
17603 2013-02-20 Jan Hubicka <jh@suse.cz>
17605 PR tree-optimization/56265
17606 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
17607 when target is referenced for first time.
17609 2013-02-20 Richard Biener <rguenther@suse.de>
17611 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
17612 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
17613 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
17614 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
17615 not return anything.
17616 (rename_ssa_copies): Do not remove unused locals.
17617 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
17618 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
17619 * passes.c (execute_function_todo): Do not schedule unused locals
17620 removal if cleanup_tree_cfg did something.
17621 * tree-ssa-live.c (remove_unused_locals): Dump statistics
17622 about the number of removed locals.
17624 2013-02-20 Richard Biener <rguenther@suse.de>
17626 PR tree-optimization/56398
17627 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
17629 2013-02-20 Martin Jambor <mjambor@suse.cz>
17631 PR tree-optimization/55334
17632 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
17633 restricted pointers to arrays.
17635 2013-02-20 Richard Biener <rguenther@suse.de>
17636 Jakub Jelinek <jakub@redhat.com>
17638 PR tree-optimization/56396
17639 * tree-ssa-ccp.c (n_const_val): New static variable.
17640 (get_value): Return NULL for SSA names we don't have a lattice
17642 (ccp_initialize): Initialize n_const_val.
17643 * tree-ssa-copy.c (n_copy_of): New static variable.
17644 (init_copy_prop): Initialize n_copy_of.
17645 (get_value): Return NULL_TREE for SSA names we don't have a
17648 2013-02-20 Martin Jambor <mjambor@suse.cz>
17650 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
17652 2013-02-20 Richard Biener <rguenther@suse.de>
17654 * genpreds.c (write_lookup_constraint): Do not compare first
17655 letter of the constraint again.
17657 2013-02-20 Richard Biener <rguenther@suse.de>
17659 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17661 (get_use_iv_cost): Terminate hashtable walk when coming across
17664 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17666 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17667 reassociation for avx2 targets.
17669 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17671 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17672 Add version check for v8.10.a to enable microblaze_has_clz
17673 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
17674 version and TARGET_PATTERN_COMPARE check
17675 * config/microblaze/microblaze.md: New clzsi2 instruction
17677 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17679 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17680 function before branching.
17682 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17684 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17686 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17688 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17690 PR middle-end/55889
17691 * sel-sched.c: Include ira.h.
17692 (implicit_clobber_conflict_p): New function.
17693 (moveup_expr): Use it.
17694 * Makefile.in (sel-sched.o): Depend on ira.h.
17696 2013-02-19 Richard Biener <rguenther@suse.de>
17698 PR tree-optimization/56384
17699 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17700 (vn_hash_type): Split out from ...
17701 (vn_hash_constant_with_type): ... here.
17702 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17703 (vn_phi_eq): Compare types from vn_phi_s structure.
17704 (vn_phi_lookup): Populate vn_phi_s type.
17705 (vn_phi_insert): Likewise.
17707 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17709 PR tree-optimization/56350
17710 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17711 if haven't found reduction or nested cycle operand, rather than
17712 asserting we must find it.
17714 PR tree-optimization/56381
17715 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17718 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17719 Jakub Jelinek <jakub@redhat.com>
17722 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17723 (swap_optab_enable): Same.
17724 (init_all_optabs): Use argument instead of global.
17725 * tree.h (struct tree_optimization_option): New field target_optabs.
17726 * expr.h (init_all_optabs): Add argument to prototype.
17727 (TREE_OPTIMIZATION_OPTABS): New.
17728 (save_optabs_if_changed): Protoize.
17729 * optabs.h: Declare this_fn_optabs.
17730 * optabs.c (save_optabs_if_changed): New.
17731 Declare this_fn_optabs.
17732 (init_optabs): Add argument to init_all_optabs() call.
17733 * function.c (invoke_set_current_function_hook): Handle per
17735 * function.h (struct function): New field optabs.
17736 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17737 optimization_current_node has changed.
17738 * target-globals.h (save_target_globals_default_opts): Protoize.
17739 * target-globals.c (save_target_globals_default_opts): New.
17741 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17744 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17745 registers %fr12 and %fr12R as call used.
17748 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17749 and HImode, require all displacements to be an integer multiple of
17751 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17752 only allow QImode and HImode when reload is in progress and strict is
17753 true. Likewise for symbolic addresses. Use base14_operand to check
17754 displacements in REG+BASE addresses.
17756 2013-02-18 Richard Biener <rguenther@suse.de>
17758 PR tree-optimization/56366
17759 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17760 handle sign-conversion of outer-loop initial induction value.
17762 2013-02-18 Richard Biener <rguenther@suse.de>
17764 PR middle-end/56349
17765 * cfghooks.c (merge_blocks): If we merge a latch into another
17766 block adjust references to it.
17767 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17768 (verify_loop_structure): Verify that a recorded latch is in fact
17771 2013-02-18 Richard Biener <rguenther@suse.de>
17773 PR tree-optimization/56321
17774 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17775 order SSA name release and virtual operand unlinking.
17777 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17779 * config/microblaze/microblaze.md (save_stack_block): Define.
17780 (restore_stack_block): Likewise.
17782 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17784 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17785 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17786 * config/microblaze/microblaze.c (microblaze_option_override):
17787 Bail out early for PIC modes when target does not support PIC.
17789 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17791 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17792 Replace with a microblaze version.
17793 (microblaze_trampoline_init): Adapt for microblaze.
17794 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17797 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17798 Dodji Seketeli <dodji@redhat.com>
17801 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17802 (instrument_mem_region_access): Do not forget to always put
17803 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17804 0) statement, even for cases where either 'base' or 'base + len'
17805 are not instrumented -- because they have been previously
17806 instrumented. Simplify the logic by putting all the statements
17807 instrument 'base + len' inside a sequence, and then insert that
17808 sequence right before the current insertion point. Then, to
17809 instrument 'base + len', just get an iterator on that statement.
17810 And do not forget to update the pointer to iterator the function
17811 received as argument.
17813 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17815 PR rtl-optimization/56348
17816 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17818 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17820 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17821 (clean_graph_dump_file): Pass base to start_graph_dump.
17823 2013-02-14 Richard Henderson <rth@redhat.com>
17826 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17828 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17830 * collect2-aix.h: Define F_LOADONLY.
17832 2013-02-14 Richard Biener <rguenther@suse.de>
17835 * varasm.c (output_constant_def_1): Get the decl representing
17836 the constant as argument.
17837 (output_constant_def): Wrap output_constant_def_1.
17838 (make_decl_rtl): Use output_constant_def_1 with the decl
17839 representing the constant.
17840 (build_constant_desc): Optionally re-use a decl already
17841 representing the constant.
17842 (tree_output_constant_def): Adjust.
17844 2013-02-14 Dodji Seketeli <dodji@redhat.com>
17847 * asan.c (instrument_builtin_call): Really put the length of the
17848 second source argument into src1_len.
17850 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17852 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
17853 argument. If it is false, don't create edge from then_bb to
17855 (insert_if_then_before_iter): Pass true to it.
17856 (build_check_stmt): Pass false to it.
17857 (transform_statements): Flush hash table only on extended basic
17858 block boundaries, rather than at the beginning of every bb.
17859 Don't flush hash table on nonfreeing_call_p calls.
17860 * tree-flow.h (nonfreeing_call_p): New prototype.
17861 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
17863 2013-02-13 David S. Miller <davem@davemloft.net>
17865 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
17867 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
17870 * ira.c (max_regno_before_ira): Move from ...
17872 (fix_reg_equiv_init): Use max_regno_before_ira instead of
17875 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17877 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
17879 2013-02-13 Richard Biener <rguenther@suse.de>
17882 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
17883 globals in MEM_REFs.
17885 2013-02-13 Richard Biener <rguenther@suse.de>
17887 * loop-init.c (loop_optimizer_init): Clear loop state when
17888 re-initializing preserved loops.
17889 * loop-unswitch.c (unswitch_single_loop): Return whether
17890 we unswitched the loop. Do not verify loop state here.
17891 (unswitch_loops): When we unswitched a loop discover new loops.
17893 2013-02-13 Kostya Serebryany <kcc@google.com>
17895 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
17897 * sanitizer.def: Rename __asan_init to __asan_init_v1.
17899 2013-02-12 Dodji Seketeli <dodji@redhat.com>
17901 Avoid instrumenting duplicated memory access in the same basic block
17902 * Makefile.in (asan.o): Add new dependency on hash-table.h
17903 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
17904 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
17905 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
17906 (free_mem_ref_resources, has_mem_ref_been_instrumented)
17907 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
17908 (get_mem_ref_of_assignment): New functions.
17909 (get_mem_refs_of_builtin_call): Extract from
17910 instrument_builtin_call and tweak a little bit to make it fit with
17912 (instrument_builtin_call): Use the new
17913 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
17914 of is_gimple_builtin_call.
17915 (instrument_derefs, instrument_mem_region_access): Insert the
17916 instrumented memory reference into the hash table.
17917 (maybe_instrument_assignment): Renamed instrument_assignment into
17918 this, and change it to advance the iterator when instrumentation
17919 actually happened and return true in that case. This makes it
17920 homogeneous with maybe_instrument_assignment, and thus give a
17921 chance to callers to be more 'regular'.
17922 (transform_statements): Clear the memory reference hash table
17923 whenever we enter a new BB, when we cross a function call, or when
17924 we are done transforming statements. Use
17925 maybe_instrument_assignment instead of instrumentation. No more
17926 need to special case maybe_instrument_assignment and advance the
17927 iterator after calling it; it's now handled just like
17928 maybe_instrument_call. Update comment.
17930 2013-02-13 Richard Biener <rguenther@suse.de>
17932 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
17933 Fix loop discovery code.
17935 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
17937 PR inline-asm/56148
17938 * lra-constraints.c (process_alt_operands): Match early clobber
17939 operand with itself. Check conflicts with earlyclobber only if
17940 the operand is not reloaded. Prefer to reload conflicting operand
17941 if earlyclobber and matching operands are the same.
17943 2013-02-12 Richard Biener <rguenther@suse.de>
17946 * lto-streamer-out.c (write_symbol): Do not output symbols
17947 for hard register variables.
17949 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
17952 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
17953 (umulsidi3_insn, mulsidi3_insn): New insns.
17955 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
17957 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
17958 (struct tune_params): Add vec_costs field.
17959 * config/arm/arm.c (arm_builtin_vectorization_cost)
17960 (arm_add_stmt_cost): New functions.
17961 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
17962 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
17963 (arm_default_vec_cost): New struct of type cpu_vec_costs.
17964 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
17965 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
17966 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
17967 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
17969 2013-02-12 Richard Biener <rguenther@suse.de>
17972 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
17973 decls again if possible.
17975 2013-02-12 Richard Biener <rguenther@suse.de>
17977 PR middle-end/56288
17978 * tree-ssa.c (verify_ssa_name): Fix check, move
17979 SSA_NAME_IN_FREE_LIST check up.
17981 2013-02-12 Jakub Jelinek <jakub@redhat.com>
17982 Steven Bosscher <steven@gcc.gnu.org>
17984 PR rtl-optimization/56151
17985 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
17986 equal to op0 or op1, and last_insn pattern is CODE operation
17987 with MEM dest and one of the operands matches that MEM.
17989 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
17991 * doc/extend.texi: Document Function Multiversioning and "default"
17992 parameter string to target attribute.
17993 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
17994 target attribute parameter is "default".
17995 (ix86_compare_version_priority): Remove checks for target attribute.
17996 (ix86_mangle_function_version_assembler_name): Change error to sorry.
17997 Remove check for target attribute equal to NULL. Add assert.
17998 (ix86_generate_version_dispatcher_body): Change error to sorry.
18000 2013-02-11 Iain Sandoe <iain@codesourcery.com>
18001 Jack Howarth <howarth@bromo.med.uc.edu>
18002 Patrick Marlier <patrick.marlier@gmail.com>
18005 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
18006 define ENDFILE_SPEC as TM_DESTRUCTOR.
18007 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
18009 2013-02-11 Alexander Potapenko <glider@google.com>
18010 Jack Howarth <howarth@bromo.med.uc.edu>
18011 Jakub Jelinek <jakub@redhat.com>
18014 * config/darwin.c (cdtor_record): Rename ctor_record.
18015 (sort_cdtor_records): Rename sort_ctor_records.
18016 (finalize_dtors): New routine to sort destructors by
18017 priority before use in assemble_integer.
18018 (machopic_asm_out_destructor): Use finalize_dtors if needed.
18020 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
18022 PR rtl-optimization/56275
18023 * simplify-rtx.c (avoid_constant_pool_reference): Check that
18024 offset is non-negative and less than cmode size before
18025 calling simplify_subreg.
18027 2013-02-11 Richard Biener <rguenther@suse.de>
18029 PR tree-optimization/56264
18030 * cfgloop.h (fix_loop_structure): Adjust prototype.
18031 * loop-init.c (fix_loop_structure): Return the number of
18032 newly discovered loops.
18033 * tree-cfgcleanup.c (repair_loop_structures): When new loops
18034 are discovered, do a full loop-closed SSA rewrite.
18036 2013-02-11 Richard Biener <rguenther@suse.de>
18038 PR tree-optimization/56273
18039 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18041 (check_array_ref): Fix missing newline in dumps.
18042 (search_for_addr_array): Likewise.
18044 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
18046 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
18048 2013-02-09 Jakub Jelinek <jakub@redhat.com>
18051 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
18053 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
18055 PR rtl-optimization/56246
18056 * lra-constraints.c (simplify_operand_subreg): Try to reuse
18058 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
18059 constraints are satisfied.
18061 2013-02-08 Jeff Law <law@redhat.com>
18064 * emit-rtl.c (reg_is_parm_p): New function.
18065 * regs.h (reg_is_parm_p): New prototype.
18066 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
18067 callee-clobbered registers.
18069 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
18072 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
18073 If there is no implicit builtin declaration, just return NULL.
18075 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18077 * config/i386/sse.md (FMAMODEM): New mode iterator.
18078 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
18079 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
18081 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18083 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
18084 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
18085 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
18087 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18089 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
18090 (microblaze*-*-elf): Likewise.
18091 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
18093 * config/microblaze/microblaze-c.c: Add builtin defines for
18094 _LITTLE_ENDIAN and _BIG_ENDIAN.
18095 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
18096 add to TARGET_DEFAULT flags.
18097 Expand ASM_SPEC and LINK_SPEC.
18098 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
18099 * config/microblaze/microblaze.md: Update extendsidi2 and
18100 movdi_internal instructions to use low-order / high-order reg
18102 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
18103 options and inversemask / mask of LITTLE_ENDIAN.
18104 * config/microblaze/t-microblaze: Expand multilib options to
18105 include mlittle-endian (le) and update exceptions patterns.
18107 2013-02-08 Jakub Jelinek <jakub@redhat.com>
18109 PR rtl-optimization/56195
18110 * lra-constraints.c (get_reload_reg): Don't reuse regs
18111 if they have smaller mode than requested, if they have
18112 wider mode than requested, try to return a SUBREG.
18114 PR tree-optimization/56250
18115 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
18116 if type is unsigned and code isn't MULT_EXPR.
18118 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18120 PR tree-optimization/56064
18121 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
18122 bits according to mode.
18123 * fixed-value.h (fixed_from_double_int)
18124 (const_fixed_from_double_int): Adjust comments.
18126 2013-02-08 Richard Biener <rguenther@suse.de>
18129 * lto-streamer.h (struct data_in): Remove current_file, current_line
18130 and current_col members.
18131 * lto-streamer-out.c (lto_output_location): Stream changed bits
18132 en-block for efficiency.
18133 * lto-streamer-in.c (clear_line_info): Remove.
18134 (lto_input_location): Cache current file, line and column
18135 globally via local statics. Read changed bits en-block.
18136 (input_function): Do not call clear_line_info.
18137 (lto_read_body): Likewise.
18138 (lto_input_toplevel_asms): Likewise.
18140 2013-02-08 Michael Matz <matz@suse.de>
18142 PR tree-optimization/52448
18143 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
18144 (nt_call_phase): New static.
18145 (add_or_mark_expr): Only mark accesses with newer phase than any
18147 (nonfreeing_call_p): New.
18148 (nt_init_block): Update nt_call_phase, mark blocks as visited.
18149 (nt_fini_block): Keep blocks marked as visited.
18150 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
18152 2013-02-08 Richard Biener <rguenther@suse.de>
18154 * ira.c (ira): Free broken dominator information.
18156 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18158 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
18160 2013-02-08 Marek Polacek <polacek@redhat.com>
18162 * cfgloop.c (verify_loop_structure): Add more checking of headers.
18164 2013-02-08 Richard Biener <rguenther@suse.de>
18166 PR middle-end/56181
18167 * cfgloop.h (flow_loops_find): Adjust.
18168 (bb_loop_header_p): Declare.
18169 * cfgloop.c (bb_loop_header_p): New function split out from ...
18170 (flow_loops_find): ... here. Adjust function signature,
18171 support incremental loop structure update.
18172 (verify_loop_structure): Cleanup. Verify a loop is a loop.
18173 * cfgloopmanip.c (fix_loop_structure): Move ...
18174 * loop-init.c (fix_loop_structure): ... here.
18175 (apply_loop_flags): Split out from ...
18176 (loop_optimizer_init): ... here.
18177 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
18178 in incremental mode, only remove dead loops here.
18180 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18183 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
18184 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
18185 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
18186 (*round<mode>3.libgcc): New insns for fixed-modes.
18187 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
18188 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
18189 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
18190 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
18191 implementations. Define to __builtin_avr_absFX,
18192 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
18193 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
18194 __builtin_avr_countlsFX, respectively.
18195 * config/avr/avr-c.c (target.h): Include it.
18196 (enum avr_builtin_id): New enum.
18197 (avr_resolve_overloaded_builtin): New static function.
18198 (avr_register_target_pragmas): Use it to set
18199 targetm.resolve_overloaded_builtin.
18200 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
18201 tree nodes used by DEF_BUILTIN.
18202 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
18203 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
18204 <AVR_BUILTIN_xxBITS>: Same.
18206 2013-02-08 Richard Biener <rguenther@suse.de>
18208 * cfgloop.c (verify_loop_structure): Properly handle
18209 a loop exiting to another loop header.
18210 * ira-int.h (ira_loops): Remove.
18211 * ira.c (ira_loops): Remove.
18212 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
18213 (do_reload): Use loop_optimizer_finalize.
18214 * ira-build.c (create_loop_tree_nodes): Use get_loops and
18215 number_of_loops to access the loop tree.
18216 (more_one_region_p): Likewise.
18217 (finish_loop_tree_nodes): Likewise.
18218 (rebuild_regno_allocno_maps): Likewise.
18219 (mark_loops_for_removal): Likewise.
18220 (mark_all_loops_for_removal): Likewise.
18221 (remove_unnecessary_regions): Likewise.
18222 (ira_build): Likewise.
18223 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18225 2013-02-08 Richard Biener <rguenther@suse.de>
18227 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
18228 * ipa-pure-const.c (analyze_function): Avoid calling
18229 mark_irreducible_loops twice.
18230 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
18232 2013-02-07 David S. Miller <davem@davemloft.net>
18234 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
18236 * var-tracking.c (vt_add_function_parameter): Test the presence of
18237 HAVE_window_save properly and do not remap argument registers when
18238 we have a leaf function.
18240 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
18243 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
18245 * config/i386/i386.c (ix86_print_operand): Ditto.
18247 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18249 * lra-constraints.c (process_alt_operands): Fix recently added comment.
18251 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18253 PR rtl-optimization/56225
18254 * lra-constraints.c (process_alt_operands): Check that reload hard
18255 reg can hold value for strict_low_part.
18257 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18260 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
18261 dwarf2out_end_function.
18262 (in_first_function_p, maybe_at_text_label_p,
18263 first_loclabel_num_not_at_text_label): New variables.
18264 (dwarf2out_var_location): In the first function find out
18265 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
18266 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
18269 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
18271 PR rtl-optimization/56178
18272 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
18273 SUBREG of a register. Tidy up related block of code.
18274 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
18275 note if the source is a register or a SUBREG of a register.
18277 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18280 * config/rs6000/rs6000.md (ptrm): New mode attr.
18281 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
18282 call_value_indirect_aix<pttrsize>,
18283 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
18286 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18288 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
18289 if -bnortl. Convert to strcmp and strncmp.
18291 2013-02-07 Alan Modra <amodra@gmail.com>
18294 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
18295 addresses won't wrap when offsetting.
18296 (rs6000_secondary_reload): Provide secondary reloads needed for
18297 wrapping LO_SUM addresses.
18299 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
18301 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
18302 MACH, just __MACH__.
18304 2013-02-06 Richard Biener <rguenther@suse.de>
18306 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
18307 instead of calling fix_loop_structure.
18309 2013-02-06 Jakub Jelinek <jakub@redhat.com>
18311 PR middle-end/56217
18312 * omp-low.c (use_pointer_for_field): Return false if
18313 lower_send_shared_vars doesn't generate any copy-out code.
18315 2013-02-06 Tom de Vries <tom@codesourcery.com>
18317 PR rtl-optimization/56131
18318 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
18319 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
18320 of the label is NULL. Add comment.
18322 2013-02-05 Jakub Jelinek <jakub@redhat.com>
18324 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
18327 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
18328 (STATIC_LIBTSAN_LIBS): Likewise.
18329 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
18330 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
18331 is defined, don't add anything else beyond that.
18332 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
18333 (LINK_COMMAND_SPEC): Use them.
18335 PR tree-optimization/56205
18336 * tree-stdarg.c (check_all_va_list_escapes): Return true if
18337 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
18338 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
18340 2013-02-05 Richard Biener <rguenther@suse.de>
18342 PR tree-optimization/53342
18343 PR tree-optimization/53185
18344 * tree-vectorizer.h (vect_check_strided_load): Remove.
18345 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
18346 not disallow peeling for vectorized strided loads.
18347 (vect_check_strided_load): Make static and simplify.
18348 (vect_analyze_data_refs): Adjust.
18349 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
18350 correctly when vectorizing strided loads.
18352 2013-02-05 Richard Biener <rguenther@suse.de>
18354 * doc/install.texi: Refer to ISL, not PPL.
18356 2013-02-05 Jan Hubicka <jh@suse.cz>
18358 PR tree-optimization/55789
18359 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
18361 2013-02-05 Jan Hubicka <jh@suse.cz>
18363 PR tree-optimization/55789
18364 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
18365 the dead call anyway.
18367 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
18370 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
18372 2013-02-04 Alexander Potapenko <glider@google.com>
18373 Jack Howarth <howarth@bromo.med.uc.edu>
18374 Jakub Jelinek <jakub@redhat.com>
18377 * config/darwin.c (sort_ctor_records): Stabilized qsort
18378 on constructor priority by using original position.
18379 (finalize_ctors): New routine to sort constructors by
18380 priority before use in assemble_integer.
18381 (machopic_asm_out_constructor): Use finalize_ctors if needed.
18383 2013-02-04 Jakub Jelinek <jakub@redhat.com>
18386 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
18387 about visibility on artificial decls.
18388 * config/sol2.c (solaris_assemble_visibility): Likewise.
18390 2013-02-04 Kai Tietz <ktietz@redhat.com>
18393 * config/i386/i386.c (function_value_ms_64): Add additional valtype
18394 argument and improve checking of return-argument types for 16-byte
18396 (ix86_function_value_1): Add additional valtype argument on call
18397 of function_value_64.
18398 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
18399 handling infunction_value_64 function.
18401 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
18403 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
18405 2013-02-04 Richard Biener <rguenther@suse.de>
18407 PR tree-optimization/56188
18408 * tree-ssa-structalias.c (label_visit): Consider case with
18409 initially non-empty points-to set.
18410 (perform_var_substitution): Dump node mapping and clean up.
18412 2013-02-04 Richard Guenther <rguenther@suse.de>
18415 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
18416 node prevail as last resort.
18417 (lto_symtab_merge_decls): Remove guard on LTRANS here.
18418 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
18420 2013-02-04 Richard Biener <rguenther@suse.de>
18422 PR tree-optimization/56113
18423 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
18425 (equiv_class_lookup_or_add): ... this.
18426 (label_visit): Adjust and fix error in previous patch.
18427 (perform_var_substitution): Adjust.
18429 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
18431 * config/sh/divtab.c: Fix formatting and comments throughout the file.
18432 * config/sh/sh4-300.md: Likewise.
18433 * config/sh/sh4a.md: Likewise.
18434 * config/sh/constraints.md: Likewise.
18435 * config/sh/sh.md: Likewise.
18436 * config/sh/netbsd-elf.h: Likewise.
18437 * config/sh/predicates.md: Likewise.
18438 * config/sh/sh-protos.h: Likewise.
18439 * config/sh/ushmedia.h: Likewise.
18440 * config/sh/linux.h: Likewise.
18441 * config/sh/sh.c: Likewise.
18442 * config/sh/superh.h: Likewise.
18443 * config/sh/elf.h: Likewise.
18444 * config/sh/sh4.md: Likewise.
18445 * config/sh/sh.h: Likewise.
18447 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18449 * config/pa/constraints.md: Adjust unused letters. Change "T"
18450 constraint to match_test floating_point_store_memory_operand().
18451 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
18452 (base14_operand): New.
18453 (floating_point_store_memory_operand): New.
18454 (integer_store_memory_operand): Revise to use base14_operand and
18455 reg_plus_base_memory_operand.
18456 (move_dest_operand): Allow symbolic_memory_operands.
18457 (symbolic_memory_operand): Check for LO_SOM.
18458 (symbolic_operand): Change default case to break.
18459 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
18460 CONST_DOUBLE values to be reloaded by putting them into memory when
18461 the destination is a floating point register.
18462 (movdf): Remove code to handle CONST_DOUBLE.
18464 (reload_indf_r1): New.
18465 (reload_insf_r1): New.
18466 Consistently use "Q" and "T" constraints with integer and floating
18467 point move instructions, respectively.
18468 (movdi): Remove FAIL.
18469 Change predicate for source operand unamed DImode move from
18470 general_operand to move_src_operand.
18471 (umulsidi3): Change predicate for destination operand to
18473 Likewise for similar unamed patterns.
18474 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
18475 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
18476 (hppa_legitimize_address): Simplify mask calculation.
18477 (pa_emit_move_sequence): Revised handling of secondary reloads from
18478 REG+D addresses for floating point loads and stores. Directly handle
18479 loading CONST0_RTX (mode) to a floating point register.
18480 (pa_secondary_reload): Handle reloading DF and SFmode constant values
18481 to floating point registers. Don't restrict secondary reloads to
18482 floating point registers to integer modes. Revise some comments and
18484 (TARGET_LEGITIMATE_ADDRESS_P): Define.
18485 (pa_legitimate_address_p): New.
18486 (pa_legitimize_reload_address): New.
18487 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
18488 (STRICT_REG_OK_FOR_BASE_P): New.
18489 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
18490 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
18492 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
18493 Andrew Dixie <andrewd@gentrack.com>
18495 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
18498 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
18500 * expmed.c (extract_bit_field_1): Pass the full width of the
18501 structure to get_best_reg_extraction_insn.
18503 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
18506 * configure.ac (use_cxa_atexit): Add AIX.
18507 * configure: Regenerate.
18509 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
18511 2013-02-01 Jakub Jelinek <jakub@redhat.com>
18514 * final.c (need_profile_function): New variable.
18515 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
18516 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
18517 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
18518 notes, targetm.asm_out.function_prologue doesn't emit anything,
18519 HAVE_prologue and profiler should be emitted before prologue,
18520 set need_profile_function instead of emitting it.
18521 (final_scan_insn): If need_profile_function, emit
18522 profile_function on the first NOTE_INSN_BASIC_BLOCK or
18523 NOTE_INSN_FUNCTION_BEG note.
18525 2013-02-01 Richard Henderson <rth@redhat.com>
18527 * config/rs6000/rs6000.md (smulditi3): New.
18530 * config/alpha/alpha.md (umulditi3): New.
18532 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
18534 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
18535 (ASM_OUTPUT_ALIGNED_LOCAL): New.
18537 2013-02-01 Richard Biener <rguenther@suse.de>
18539 PR tree-optimization/56113
18540 * tree-ssa-structalias.c (label_visit): Reduce work for
18541 single-predecessor nodes.
18543 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
18545 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
18546 range isn't testing for zero.
18548 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
18550 PR middle-end/56113
18551 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
18553 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
18554 Nick Clifton <nickc@redhat.com>
18556 * config/v850/constraints.md (Q): Define as a memory constraint.
18557 * config/v850/predicates.md (label_ref_operand): New predicate.
18558 (e3v5_shift_operand): New predicate.
18559 (ior_operator): New predicate.
18560 * config/v850/t-v850: Add e3v5 multilib.
18561 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
18562 (v850_gen_movdi): Prototype.
18563 * config/v850/v850.c: Add support for e3v5 architecture.
18564 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
18566 (construct_save_jarl): Add e3v5 long JARL support.
18567 (v850_adjust_insn_length): New function. Adjust length of call
18568 insns when using e3v5 instructions.
18569 (v850_gen_movdi): New function: Generate instructions to move a
18571 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
18572 (CPP_SPEC): Define __v850e3v5__ as appropriate.
18573 (TARGET_USE_FPU): Enable for e3v5.
18574 (CONST_OK_FOR_W): New macro.
18575 (ADJUST_INSN_LENGTH): Define.
18576 * config/v850/v850.md (UNSPEC_LOOP): Define.
18577 (attr cpu): Add v850e3v5.
18578 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
18579 (movdi): New pattern.
18580 (movdi_internal): New pattern.
18581 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
18582 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
18583 (cstoresf4): Likewise.
18584 (cstoredf4): Likewise.
18585 (insv): New pattern.
18586 (rotlso3_a): New pattern.
18587 (rotlsi3_b): New pattern
18588 (rotlsi3_v850e3v5): New pattern.
18589 (doloop_begin): New pattern.
18590 (fix_loop_counter): New pattern.
18591 (doloop_end): New pattern.
18592 (branch_normal): Add e3v5 long branch support.
18593 (branch_invert): Likewise.
18594 (branch_z_normal): Likewise.
18595 (branch_z_invert): Likewise.
18596 (branch_nz_normal): Likewise.
18597 (branch_nz_invert): Likewise.
18598 (call_internal_short): Add e3v5 register-indirect JARL support.
18599 (call_internal_long): Likewise.
18600 (call_value_internal_short): Likewise.
18601 (call_value_internal_long): Likewise.
18602 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
18603 (mloop): New option.
18604 * config.gcc: Add support for configuring v840e3v5 target.
18605 * doc/invoke.texi: Document new v850 specific command line options.
18607 2013-01-31 Paul Koning <ni1d@arrl.net>
18611 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
18612 children if parent is a class.
18613 (prune_unused_types_prune): Don't add DW_AT_declaration.
18615 2013-01-31 Richard Biener <rguenther@suse.de>
18617 PR tree-optimization/56157
18618 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
18619 match up operand with SLP child.
18621 2013-01-31 Jason Merrill <jason@redhat.com>
18624 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
18625 parameters the first time.
18626 (gen_scheduled_generic_parms_dies): Check completeness here.
18628 2013-01-31 Richard Biener <rguenther@suse.de>
18630 PR middle-end/53073
18631 * common.opt (faggressive-loop-optimizations): New flag,
18632 enabled by default.
18633 * doc/invoke.texi (faggressive-loop-optimizations): Document.
18634 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
18635 infer_loop_bounds_from_undefined by it.
18637 2013-01-31 Richard Biener <rguenther@suse.de>
18639 PR tree-optimization/56150
18640 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
18641 visit virtual operands.
18642 (find_uses_to_rename_bb): Likewise.
18644 2013-01-31 Richard Biener <rguenther@suse.de>
18646 PR tree-optimization/56150
18647 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
18648 mixed store non-store stmts.
18650 2013-01-30 Jakub Jelinek <jakub@redhat.com>
18653 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
18654 LIBASAN_EARLY_SPEC is defined.
18655 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
18656 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18658 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18661 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18662 invalid args instead of ICEing on it.
18663 (ix86_valid_target_attribute_tree): Return error_mark_node if
18664 ix86_valid_target_attribute_inner_p failed.
18665 (ix86_valid_target_attribute_p): Return false only if
18666 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18667 target("default") attribute.
18668 (sorted_attr_string): Change argument from const char * to tree,
18669 merge in all target attribute arguments rather than just one.
18670 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18671 instead of free. Avoid using strcat.
18672 (ix86_mangle_function_version_assembler_name): Mangle
18673 target("default") as if no target attribute is present. Adjust
18674 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18675 instead of xmalloc and XDELETEVEC instead of free.
18676 (ix86_function_versions): Don't return true if one of the decls
18677 doesn't have target attribute. If they don't and one of the decls
18678 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18679 sorted_attr_string caller. Use XDELETEVEC instead of free.
18680 (ix86_supports_function_versions): Remove.
18681 (make_name): Fix up formatting.
18682 (make_dispatcher_decl): Remove resolver_name and its initialization.
18683 Avoid leaking memory.
18684 (is_function_default_version): Return true if there is
18685 target("default") attribute rather than no target attribute at all.
18686 (make_resolver_func): Avoid leaking memory.
18687 (ix86_generate_version_dispatcher_body): Likewise.
18688 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18689 * target.def (supports_function_versions): Remove.
18690 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18691 * doc/tm.texi: Regenerated.
18693 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18695 PR rtl-optimization/56144
18696 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18697 for values with side effects.
18699 2013-01-30 Richard Biener <rguenther@suse.de>
18701 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18702 (sparseset_pop): Likewise.
18703 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18704 to be able to use quick_push in the worker loop.
18706 2013-01-30 Marek Polacek <polacek@redhat.com>
18708 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18710 2013-01-30 Richard Biener <rguenther@suse.de>
18713 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18715 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18717 PR tree-optimization/56064
18718 * fixed-value.c (fixed_from_double_int): New function.
18719 * fixed-value.h (fixed_from_double_int): New prototype.
18720 (const_fixed_from_double_int): New static inline function.
18721 * fold-const.c (native_interpret_fixed): New static function.
18722 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18723 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18724 (native_encode_fixed): New static function.
18725 (native_encode_expr) <FIXED_CST>: Use it.
18726 (native_interpret_int): Move double_int worker code to...
18727 * double-int.c (double_int::from_buffer): ...this new static method.
18728 * double-int.h (double_int::from_buffer): Prototype it.
18730 2013-01-30 Richard Biener <rguenther@suse.de>
18732 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18733 New pointer-map and obstack.
18734 (init_alias_vars): Allocate pointer-map and obstack.
18735 (delete_points_to_sets): Free them.
18736 (find_what_var_points_to): Cache result.
18737 (find_what_p_points_to): Adjust for changed interface of
18738 find_what_var_points_to.
18739 (compute_points_to_sets): Likewise.
18740 (ipa_pta_execute): Likewise.
18742 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18744 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18745 * configure: Regenerate.
18746 * config.in: Regenerate.
18747 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18748 #nobits/#progbits if supported.
18750 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18753 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18754 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18755 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18757 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18759 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18760 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18762 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18764 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18765 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18767 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18769 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18771 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18772 * config/arm/cortex-a7.md: New bypasses using
18773 arm_mac_accumulator_is_result.
18775 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18777 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18778 (cortex_a7_neon_mla): Likewise.
18779 (cortex_a7_fpfmad): New reservation.
18780 (cortex_a7_fpmacs): Use ffmas and update required units.
18781 (cortex_a7_fpmuld): Update required units and latency.
18782 (cortex_a7_fpmacd): Likewise.
18783 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18784 (cortex_a7_neon). Likewise.
18785 (bypass) Update participating units.
18787 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18789 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18790 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18792 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18793 (vfp_fmul): Use ffmad.
18794 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18795 (cortex_r4_fmacd): Use ffmad.
18796 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18797 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18798 (cortex_a9_fmacd): Use ffmad.
18799 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18800 (cortex_a8_vfp_macd): Use ffmad.
18801 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18802 (cortex_a5_fpmacd): Use ffmad.
18803 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18804 (cortex_a15_vfp_macd): Use ffmad.
18805 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18807 2013-01-29 Jason Merrill <jason@redhat.com>
18810 * varasm.c (default_assemble_visibility): Don't warn about
18811 visibility on artificial decls.
18813 2013-01-29 Richard Biener <rguenther@suse.de>
18815 PR tree-optimization/56113
18816 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18818 (label_visit): Free duplicate bitmaps and record the leader instead.
18819 (perform_var_substitution): Adjust.
18821 2013-01-29 Richard Biener <rguenther@suse.de>
18823 PR tree-optimization/55270
18824 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18825 the CFG, schedule loops for fixup.
18827 2013-01-29 Nick Clifton <nickc@redhat.com>
18829 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18832 2013-01-28 Leif Ekblad <leif@rdos.net>
18834 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18835 * config/i386/i386.h (TARGET_RDOS): New macro.
18836 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18837 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18838 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
18839 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
18840 DEFAULT_LARGE_SECTION_THRESHOLD.
18841 * config/i386/i386.md (R14_REG, R15_REG): New constants.
18842 * config/i386/rdos.h: New file.
18843 * config/i386/rdos64.h: New file.
18845 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
18848 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
18851 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18853 PR rtl-optimization/56117
18854 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
18855 call cselib_lookup_from_insn on the MEM before calling
18856 add_insn_mem_dependence.
18858 2013-01-28 Richard Biener <rguenther@suse.de>
18860 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
18861 to a stmt that didn't have one.
18862 (copy_phis_for_bb): Likewise for PHI arguments.
18863 (copy_debug_stmt): Likewise for debug stmts.
18865 2013-01-28 Richard Biener <rguenther@suse.de>
18867 PR tree-optimization/56034
18868 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
18869 (partition_builtin_p): Adjust.
18870 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
18871 it is the last partition.
18872 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
18873 up the vertex for the definition.
18874 (classify_partition): Classify whether a partition is a
18875 PKIND_REDUCTION, thus has uses outside of the loop.
18876 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
18877 Merge all PKIND_REDUCTION partitions into the last partition.
18878 (tree_loop_distribution): Seed partitions from reductions as well.
18880 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18882 PR tree-optimization/56125
18883 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
18884 pow(x,c) into sqrt(x) * powi(x, n/2) or
18885 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
18886 optimizing for size.
18887 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
18888 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
18891 PR tree-optimization/56094
18892 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
18893 to UNKNOWN_LOCATION while gimplifying expr.
18895 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
18898 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
18899 operand 0 in movabs insn template for -masm=intel asm alternative.
18900 (*movabs<mode>_2): Ditto for operand 1.
18902 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
18905 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
18908 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18910 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
18913 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18915 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
18916 Undef to avoid warning.
18918 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18920 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
18921 * configure: Regenerate.
18923 2013-01-25 Jakub Jelinek <jakub@redhat.com>
18925 PR tree-optimization/56098
18926 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
18927 for stmts with volatile ops.
18928 (cond_store_replacement): Don't optimize if assign has volatile ops.
18929 (cond_if_else_store_replacement_1): Don't optimize if either
18930 then_assign or else_assign have volatile ops.
18931 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
18934 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18936 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
18938 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18940 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
18941 missing ':' in asm example.
18943 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
18945 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
18946 entries into lane and laneq entries.
18947 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
18948 Remove AdvSIMD scalar modes.
18949 (aarch64_sq<r>dmulh_laneq<mode>): New.
18950 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
18952 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
18953 builtin implementations to relfect changes in RTL in aarch64-simd.md.
18954 * config/aarch64/iterators.md (VCOND): New.
18957 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18960 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
18961 Add NULL LIBNAME argument to existing definitions.
18962 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
18963 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
18964 * config/avr/avr.c (DEF_BUILTIN): Same.
18965 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
18966 (avr_expand_builtin): Expand to a vanilla call if a libgcc
18967 implementation is available (DECL_ASSEMBLER_NAME is set).
18968 (avr_fold_absfx): New static function.
18969 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
18970 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
18971 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
18972 AVR_BUILTIN_ABSLLK.
18973 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
18974 (abshk, absk, abslk, absllk): Provide as static inline functions.
18976 2013-01-25 Marek Polacek <polacek@redhat.com>
18978 PR tree-optimization/56035
18979 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
18981 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
18983 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
18984 (*movtf_internal_rex64): Add (!o,C) alternative
18985 (*movxf_internal_rex64): Ditto.
18986 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
18988 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
18990 * doc/invoke.texi: fix typo.
18991 * doc/objc.texi: fix typo.
18993 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
18995 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
18996 for the first two alternatives.
18998 2013-01-24 Diego Novillo <dnovillo@google.com>
19000 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
19001 (ggc-zone.o): Remove.
19002 * configure.ac: Remove option --with-gc.
19003 * configure: Re-generate.
19004 * doc/install.texi: Remove documentation for --with-gc.
19005 * gengtype.c (write_enum_defn): Remove. Update all users.
19006 (write_Types_process_field): Remove generation of gt_e_* argument.
19007 (output_type_enum): Remove. Update all users.
19008 (write_enum_defn): Remove. Update all users.
19009 (enum alloc_zone): Remove. Update all users.
19010 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
19011 * ggc-common.c (ggc_splay_alloc): Remove first argument.
19012 Update all callers.
19013 (struct ptr_data): Remove field TYPE. Update all users.
19014 (gt_pch_note_object): Remove argument TYPE. Update all users.
19015 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
19017 * ggc-none.c (ggc_alloc_typed_stat): Remove.
19018 (struct alloc_zone): Remove.
19019 (ggc_internal_alloc_zone_stat): Remove.
19020 (ggc_internal_cleared_alloc_zone_stat): Remove.
19021 * ggc-page.c (ggc_alloc_typed_stat): Remove.
19022 (ggc_pch_count_object): Remove last argument. Update all users.
19023 (ggc_pch_alloc_object): Remove last argument. Update all users.
19024 (struct alloc_zone): Remove.
19025 * ggc-zone.c: Remove.
19026 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
19027 (struct alloc_zone): Remove.
19028 (ggc_alloc_typed_stat): Remove.
19029 (ggc_alloc_typed): Remove.
19030 (ggc_splay_alloc): Remove first argument.
19031 (rtl_zone): Remove. Update all users.
19032 (tree_zone): Remove. Update all users.
19033 (tree_id_zone): Remove. Update all users.
19034 (ggc_internal_zone_alloc_stat): Remove. Update all users.
19035 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
19036 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
19037 * tree-ssanames.c: Remove references to zone allocator in comments.
19039 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
19041 * config/avr/avr.c (avr_out_fract): Make register numbers that
19042 might be outside of source operand signed.
19044 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
19046 * config/i386/constraints.md (Yf): New constraint.
19047 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
19048 of f constraint to conditionaly disable x87 register preferences.
19049 (*movdf_internal): Ditto.
19050 (*movsf_internal): Ditto.
19052 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
19054 PR inline-asm/55934
19055 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
19056 that have operands with impossible constraints.
19057 Add a FIXME for a speed-up opportunity.
19058 * lra-constraints.c (process_alt_operands): Verify that a class
19059 selected from constraints on asms is valid for the operand mode.
19060 (curr_insn_transform): Remove incorrect comment.
19062 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
19064 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
19065 TOC operand is a valid symbol ref in the constant pool.
19067 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19069 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
19071 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
19074 * config/avr/stdfix.h: New file.
19075 * t-avr (stdfix-gcc.h): New rule to build it.
19076 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
19078 2013-01-23 Kostya Serebryany <kcc@google.com>
19080 * config/darwin.h: remove dependency on
19081 CoreFoundation (asan on Mac OS).
19083 2013-01-23 Jakub Jelinek <jakub@redhat.com>
19086 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
19087 instead of cmpdi_operand for first comparison operand.
19088 Don't assert that comparison operands aren't both constants.
19090 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19092 * doc/install.texi (Downloading the Source): Update references to
19093 downloading separate components.
19095 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19097 * doc/extend.texi (__int128): Improve grammar.
19099 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
19102 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
19103 alternative to (o,r).
19104 (*movdi_internal_rex64): Remove (!o,n) alternative.
19105 (DImode immediate->memory splitter): Remove.
19106 (DImode immediate->memory peephole2): Remove.
19107 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
19108 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
19109 alternative to (!o,*r).
19110 (*movtf_internal_sse): New pattern.
19111 (*movxf_internal_rex64): New pattern.
19112 (*movxf_internal): Disable for TARGET_64BIT.
19113 (*movdf_internal_rex64): Remove (!o,F) alternative.
19115 2013-01-22 Jakub Jelinek <jakub@redhat.com>
19117 PR middle-end/56074
19118 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
19119 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
19120 * tree-vect-loop-manip.c (find_loop_location): Also ignore
19121 stmt locations where LOCATION_LOCUS of the stmt location is
19122 UNKNOWN_LOCATION or BUILTINS_LOCATION.
19125 * config/i386/i386.md (UNSPEC_STOS): New.
19126 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
19127 *strsetqi_1): Add UNSPEC_STOS.
19129 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
19132 * doc/invoke.texi: Remove left over -Wsynth example.
19134 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19136 PR tree-optimization/56051
19137 * fold-const.c (fold_binary_loc): Don't fold
19138 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
19139 a narrowing conversion, or widening conversion from signed
19142 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19144 PR rtl-optimization/56023
19145 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
19146 dependent on debug instruction.
19148 2013-01-21 Martin Jambor <mjambor@suse.cz>
19150 PR middle-end/56022
19151 * function.c (allocate_struct_function): Call
19152 invoke_set_current_function_hook earlier.
19154 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19156 * reload1.c (init_reload): Only initialize reload_obstack
19157 during the first call.
19159 2013-01-21 Marek Polacek <polacek@redhat.com>
19161 * cfgloop.c (verify_loop_structure): Fix up grammar.
19163 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
19165 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
19166 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
19168 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19171 * config/arm/marvell-pj4.md: Update copyright year.
19172 Fix up use of alu to alu_reg and simple_alu_imm.
19174 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19176 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
19178 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
19181 * lra-constraints.c (curr_insn_transform): Don't reuse original
19182 insn for secondary memory move when memory mode should be different.
19184 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19186 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
19187 atomic_storedi_1): New patterns.
19189 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19191 btver2 pipeline descriptions.
19192 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
19194 * config/i386/i386.md (btver2_decode): New type attributes.
19195 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
19197 * config/i386/btver2.md: New file describing btver2 pipelines.
19199 2013-01-19 Andrew Pinski <apinski@cavium.com>
19201 PR tree-optimization/52631
19202 * tree-ssa-sccvn (visit_use): Before looking up the original
19203 statement, try looking up the simplified expression.
19205 2013-01-19 Anthony Green <green@moxielogic.com>
19207 * config/moxie/moxie.c (moxie_expand_prologue): Set
19208 current_function_static_stack_size.
19210 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19212 PR tree-optimization/56029
19213 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
19214 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
19216 2013-01-18 Sharad Singhai <singhai@google.com>
19218 PR tree-optimization/55995
19219 * dumpfile.c (dump_loc): Print location only if available.
19220 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
19222 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
19225 * lra-constraints.c (curr_insn_transform): Reuse original insn for
19226 secondary memory move.
19227 (inherit_reload_reg): Use rclass instead of cl for
19228 check_secondary_memory_needed_p.
19230 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19232 PR middle-end/56015
19233 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
19234 the case where writing real complex part of target modifies op1.
19236 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
19238 * config/aarch64/aarch64-simd.md
19239 (aarch64_vcond_internal<mode>): Handle unordered cases.
19240 * config/aarch64/iterators.md (v_cmp_result): New.
19242 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
19243 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19245 * config/arm/marvell-pj4.md: New file.
19246 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
19247 * config/arm/arm.md (generic_sched): Add marvell_pj4.
19248 (generic_vfp): Likewise.
19249 * config/arm/arm-cores.def: Add marvell-pj4.
19250 * config/arm/arm-tune.md: Regenerate.
19251 * config/arm/arm-tables.opt: Regenerate.
19252 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
19253 * doc/invoke.texi: Document marvell-pj4.
19255 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
19257 * config/aarch64/arm_neon.h: Map scalar types to standard types.
19259 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19264 * var-tracking.c (negative_power_of_two_p): New.
19265 (global_get_addr_cache, local_get_addr_cache): New.
19266 (get_addr_from_global_cache, get_addr_from_local_cache): New.
19267 (vt_canonicalize_addr): Rewrite using the above. Adjust the
19269 (vt_stack_offset_p): Remove.
19270 (vt_canon_true_dep): Always canonicalize loc's address.
19271 (clobber_overlapping_mems): Make sure we have a MEM.
19272 (local_get_addr_clear_given_value): New.
19273 (val_reset): Clear local cached entries.
19274 (compute_bb_dataflow): Create and release the local cache.
19275 Disable duplicate MEMs clobbering.
19276 (emit_notes_in_bb): Clobber MEMs likewise.
19277 (vt_emit_notes): Create and release the local cache.
19278 (vt_initialize, vt_finalize): Create and release the global
19279 cache, respectively.
19280 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
19282 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19284 PR libmudflap/53359
19285 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
19286 not found in the symtab.
19288 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19291 PR rtl-optimization/55547
19292 PR rtl-optimization/53827
19295 * alias.c (offset_overlap_p): New, factored out of...
19296 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
19297 the conservative special case for symbolic constants. Don't
19298 adjust zero sizes on alignment.
19300 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
19302 PR rtl-optimization/52573
19303 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
19304 REG_UNUSED for the same register.
19306 2013-01-17 Richard Biener <rguenther@suse.de>
19307 Marek Polacek <polacek@redhat.com>
19309 PR rtl-optimization/55833
19310 * loop-unswitch.c (unswitch_loops): Move loop verification...
19311 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
19312 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
19313 Set it to true when we're removing a loop from hierarchy tree in
19314 an irreducible region.
19315 (fix_bb_placements): Adjust caller.
19316 (fix_loop_placements): Likewise.
19318 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
19320 * config/avr/builtins.def (DEF_BUILTIN): Factor out
19321 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
19322 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
19323 Remove ID. Adjust comments.
19324 * config/avr/avr-c.c (avr_builtin_name): Remove.
19325 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
19326 * config/avr/avr.c (avr_tolower): New static function.
19327 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
19328 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
19329 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
19332 2013-01-17 Jan Hubicka <jh@suse.cz>
19334 PR tree-optimization/55273
19335 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
19337 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
19340 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
19341 store through atomic_store<mode>_1.
19342 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
19344 2013-01-17 Martin Jambor <mjambor@suse.cz>
19346 PR tree-optimizations/55264
19347 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
19348 for virtual methods.
19349 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
19350 virtual methods before inlining is over.
19351 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
19353 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
19356 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
19358 PR rtl-optimization/56005
19359 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
19360 pending reads for prefetch.
19362 2013-01-16 Ian Bolton <ian.bolton@arm.com>
19364 * config/aarch64/aarch64.md
19365 (*cstoresi_neg_uxtw): New pattern.
19366 (*cmovsi_insn_uxtw): New pattern.
19367 (*<optab>si3_uxtw): New pattern.
19368 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
19369 (*<optab>si3_insn_uxtw): New pattern.
19370 (*bswapsi2_uxtw): New pattern.
19372 2013-01-16 Richard Biener <rguenther@suse.de>
19374 * tree-inline.c (tree_function_versioning): Remove set but
19375 never used variable.
19377 2013-01-16 Richard Biener <rguenther@suse.de>
19379 PR tree-optimization/55964
19380 * tree-flow.h (rename_variables_in_loop): Remove.
19381 (rename_variables_in_bb): Likewise.
19382 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
19383 (copy_loop_before): Adjust and delete update-ssa status.
19384 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
19385 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
19386 (rename_variables_in_loop): Remove.
19387 (slpeel_update_phis_for_duplicate_loop): Likewise.
19388 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
19389 use available cfg machinery instead of duplicating it.
19390 Update PHI nodes and perform poor-mans SSA update here.
19391 (slpeel_tree_peel_loop_to_edge): Adjust.
19393 2013-01-16 Richard Biener <rguenther@suse.de>
19395 PR tree-optimization/54767
19396 PR tree-optimization/53465
19397 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
19398 (vrp_visit_phi_node): For PHI arguments coming via backedges
19399 drop all symbolical range information.
19400 (execute_vrp): Compute backedges.
19402 2013-01-16 Richard Biener <rguenther@suse.de>
19404 * doc/install.texi: Update CLooG and ISL requirements to
19407 2013-01-16 Christian Bruel <christian.bruel@st.com>
19410 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
19411 (broken_move): Handle UNSPECV_SP_SWITCH_B.
19412 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
19414 2013-01-16 DJ Delorie <dj@redhat.com>
19416 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
19417 (UNSPECV_SP_SWITCH_E): New.
19418 (sp_switch_1): Change to an unspec.
19419 (sp_switch_2): Change to an unspec. Don't use post-inc when we
19422 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
19424 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
19425 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
19426 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
19427 (expand_mem_thread_fence): Ditto.
19428 (expand_mem_signal_fence): Ditto.
19429 (expand_atomic_load): Ditto.
19430 (expand_atomic_store): Ditto.
19432 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
19434 PR rtl-optimization/55547
19435 PR rtl-optimization/53827
19438 * alias.c (memrefs_conflict_p): Set sizes to negative after
19441 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19444 * function.c (thread_prologue_and_epilogue_insns): Always
19445 add crtl->drap_reg to set_up_by_prologue.set, even if
19446 stack_realign_drap is false.
19448 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
19450 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
19451 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
19452 *call): Fix indention.
19454 2013-01-15 Tom de Vries <tom@codesourcery.com>
19457 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
19460 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
19462 PR rtl-optimization/55153
19463 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
19465 2013-01-15 Martin Jambor <mjambor@suse.cz>
19467 PR tree-optimization/55920
19468 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
19469 accesses as grp_to_be_debug_replaced.
19471 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19473 PR tree-optimization/55920
19474 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
19475 there is non-useless type conversion needed from debug rhs to lhs,
19476 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
19478 2013-01-15 Joseph Myers <joseph@codesourcery.com>
19479 Mikael Pettersson <mikpe@it.uu.se>
19482 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
19484 (ASM_OUTPUT_CASE_LABEL): Remove.
19485 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
19486 * final.c (shorten_branches): Update alignment of labels before
19487 jump tables if CASE_VECTOR_SHORTEN_MODE.
19489 2013-01-15 Richard Biener <rguenther@suse.de>
19492 * system.h: Do not include gmp.h for building host tools.
19494 2013-01-15 Richard Biener <rguenther@suse.de>
19496 PR middle-end/55882
19497 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
19498 account for bitpos when computing alignment.
19500 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
19502 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
19503 (ix86_target_macros_internal): Likewise.
19505 * config/i386/i386.c (m_CORE2I7): Removed.
19506 (m_CORE_HASWELL): New macro.
19507 (m_CORE_ALL): Likewise.
19508 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
19509 (initial_ix86_arch_features): Likewise.
19510 (processor_target_table): Initializations for Core avx2.
19511 (cpu_names): New names "core-avx2".
19512 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
19513 PROCESSOR_CORE_HASWELL.
19514 (ix86_issue_rate): New case.
19515 (ia32_multipass_dfa_lookahead): Likewise.
19516 (ix86_sched_init_global): Likewise.
19518 * config/i386/i386.h (TARGET_HASWELL): New macro.
19519 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
19520 (processor_type): New PROCESSOR_HASWELL.
19522 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19524 PR tree-optimization/55955
19525 * tree-vect-loop.c (vectorizable_reduction): Give up early on
19526 *SHIFT_EXPR and *ROTATE_EXPR codes.
19528 PR tree-optimization/48766
19529 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
19530 -ftrapv disable -fwrapv.
19532 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19535 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
19536 etc. to 1 and not to __flash.
19537 Use LL suffix for __INT24_MAX__ with -mint8.
19538 Use ULL suffix for __UINT24_MAX__ with -mint8.
19540 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19542 * config/avr/avr-arch.h
19543 (struct base_arch_s): Use typedef avr_arch_t instead.
19544 (struct arch_info_s): Use typedef avr_arch_info_t instead.
19545 (struct mcu_type_s): Use typedef avr_mcu_t instead.
19546 * config/avr/avr.c: Same.
19547 * config/avr/avr-devices.c: Same.
19548 * config/avr/driver-avr.c: Same.
19549 * config/avr/gen-avr-mmcu-texi.c: Same.
19550 * config/avr/avr-mcus.def: Adjust comment.
19552 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
19554 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
19555 * config/aarch64/iterators.md (VALLDI): New.
19557 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
19558 Andi Kleen <ak@linux.intel.com>
19561 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
19562 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
19565 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19567 * config/avr/avr-stdint.h: Remove trailing blanks.
19568 * config/avr/avr-log.h: Same.
19569 * config/avr/avr-arch.h: Same.
19570 * config/avr/avr-devices.c: Same.
19571 * config/avr/avr-dimode.md: Same.
19572 * config/avr/predicates.md: Same.
19573 * config/avr/avr-c.c: Same. And fix typo.
19575 * config/avr/avr-protos.h: Same. And:
19576 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
19577 (init_cumulative_args): Rename to avr_init_cumulative_args.
19578 (expand_prologue): Rename to avr_expand_prologue.
19579 (expand_epilogue): Rename to avr_expand_epilogue.
19580 (adjust_insn_length): Rename to avr_adjust_insn_length.
19581 (notice_update_cc): Rename to avr_notice_update_cc.
19582 (final_prescan_insn): Rename to avr_final_prescan_insn.
19583 * config/avr/avr.c: Same.
19584 * config/avr/avr.h: Same.
19585 * config/avr/avr.md: Remove trailing blanks.
19586 (prologue): Use avr_expand_prologue.
19587 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
19589 2013-01-14 Richard Biener <rguenther@suse.de>
19591 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
19592 verify_location, collect_subblocks): New functions.
19593 (verify_gimple_in_cfg): Verify that locations only reference
19594 BLOCKs in the functions BLOCK tree.
19596 2013-01-14 Richard Biener <rguenther@suse.de>
19598 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
19600 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
19602 (insert_out_of_ssa_copy_on_edge): Likewise.
19603 (rewrite_close_phi_out_of_ssa): Likewise.
19604 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
19606 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
19607 propagated constants.
19608 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
19611 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19613 * config/avr/avr-modes.def: Add GPL copyright notice.
19615 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
19617 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
19618 MEMMODEL_MASK to determine memory model.
19619 (atomic_store<mode>): Ditto from operands[2].
19620 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
19622 2013-01-13 Jakub Jelinek <jakub@redhat.com>
19625 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
19626 (fold_gimple_assign): Don't call unshare_expr here.
19627 (fold_ctor_reference): Call unshare_expr.
19629 2013-01-13 Terry Guo <terry.guo@arm.com>
19631 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
19632 * doc/fragments.texi: Document MULTILIB_REUSE.
19633 * gcc.c (multilib_reuse): New internal spec.
19634 (set_multilib_dir): Also search multilib from multilib_reuse.
19635 * genmultilib (tmpmultilib3): Refactor code.
19636 (tmpmultilib4): Ditto.
19637 (multilib_reuse): New multilib argument.
19639 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
19641 * Makefile.in: Update copyright.
19643 2013-01-12 Tom de Vries <tom@codesourcery.com>
19645 PR middle-end/55890
19646 * calls.c (expand_call): Check if arg_nr is valid.
19648 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19650 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
19651 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
19652 documentation. Add missing '__' in front of
19653 __builtin_ia32_packssdw256.
19655 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19658 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19659 NO_REGS for larl operands.
19660 (s390_reload_larl_operand): Use s390_load_address instead of
19663 2013-01-11 Richard Biener <rguenther@suse.de>
19665 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19666 (verify_node_sharing): ... here.
19667 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19669 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19671 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19672 Substitute TREECHECKING.
19673 * configure: Regenerate.
19674 * Makefile.in (TREECHECKING): New.
19676 2013-01-11 Richard Guenther <rguenther@suse.de>
19678 PR tree-optimization/44061
19679 * tree-vrp.c (extract_range_basic): Compute zero as
19680 value-range for __builtin_constant_p of function parameters.
19682 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19684 Update copyright years.
19686 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19688 PR rtl-optimization/55672
19689 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19690 const to be eliminable.
19692 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19694 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19695 * configure: Regenerate.
19697 2013-01-10 Richard Biener <rguenther@suse.de>
19699 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19701 2013-01-10 Richard Biener <rguenther@suse.de>
19704 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19705 locations for virtual PHI arguments.
19706 (rewrite_update_phi_arguments): Likewise.
19708 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19710 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19713 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19715 PR tree-optimization/55921
19716 * tree-complex.c (expand_complex_asm): New function.
19717 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19719 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19722 * config/s390/s390.c (s390_symref_operand_p)
19723 (s390_loadrelative_operand_p): Merge the two functions.
19724 (s390_check_qrst_address, print_operand_address): Add parameters
19725 to s390_loadrelative_operand_p invokation.
19726 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19727 (s390_reload_larl_operand, s390_secondary_reload): Use
19728 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19729 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19731 2013-01-09 Mike Stump <mikestump@comcast.net>
19733 * dse.c (record_store): Remove unnecessary assert.
19735 2013-01-09 Jan Hubicka <jh@suse.cz>
19737 PR tree-optimization/55569
19738 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19739 * cfgloop.h (scale_loop_profile): Likewise.
19741 2013-01-09 Jan Hubicka <jh@suse.cz>
19744 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19746 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19748 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19750 PR middle-end/55114
19751 * expr.h (maybe_emit_group_store): Declare.
19752 * expr.c (maybe_emit_group_store): New function.
19753 * builtins.c (expand_builtin_int_roundingfn): Call it.
19754 (expand_builtin_int_roundingfn_2): Likewise.
19756 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19758 PR rtl-optimization/55829
19759 * lra-constraints.c (match_reload): Add code for absent output.
19760 (curr_insn_transform): Add code for reloads of matched inputs
19763 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19765 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19766 attribute of movddup insn to DF.
19767 (*vec_interleave_lowv2df): Ditto.
19768 (vec_dupv2df): Ditto.
19770 2013-01-09 Jan Hubicka <jh@suse.cz>
19772 PR tree-optimiation/55875
19773 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19774 EVERY_ITERATION parameter.
19775 (number_of_iterations_exit): Check if exit is executed every iteration.
19776 (idx_infer_loop_bounds): Similarly here.
19777 (n_of_executions_at_most): Simplify
19778 to only test for cases where statement is dominated by the
19779 particular bound; handle correctly the "postdominance" test.
19780 (scev_probably_wraps_p): Use max loop iterations info
19781 as a global bound first.
19783 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19784 Nick Clifton <nickc@redhat.com>
19786 * config/v850/v850.md (cbranchsf4): New pattern.
19787 (cstoresf4): New pattern.
19788 (cbranchdf4): New pattern.
19789 (cstoredf4): New pattern.
19790 (movsicc): Disallow floating point comparisons.
19791 (cmpsf_le_insn): Fix order of operators.
19792 (cmpsf_lt_insn): Likewise.
19793 (cmpsf_eq_insn): Likewise.
19794 (cmpdf_le_insn): Likewise.
19795 (cmpdf_lt_insn): Likewise.
19796 (cmpdf_eq_insn): Likewise.
19797 (cmpsf_ge_insn): Use LE comparison.
19798 (cmpdf_ge_insn): Likewise.
19799 (cmpsf_gt_insn): Use LT comparison.
19800 (cmpdf_gt_insn): Likewise.
19801 (cmpsf_ne_insn): Delete pattern.
19802 (cmpdf_ne_insn): Delete pattern.
19803 * config/v850/v850.c (v850_gen_float_compare): Use
19804 gen_cmpdf_eq_insn for NE comparison.
19805 (v850_float_z_comparison_operator)
19806 (v850_float_nz_comparison_operator): Move from here ...
19807 * config/v850/predicates.md: ... to here. Move GT and GE
19808 comparisons into v850_float_z_comparison_operator.
19809 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19811 (v850_float_nz_comparison_operator): Likewise.
19813 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19815 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19816 with calls to gen_insvsi/gen_insvdi.
19818 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19820 * config/i386/i386.c (initial_ix86_tune_features): Set up
19821 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19823 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19824 Jakub Jelinek <jakub@redhat.com>
19826 PR tree-optimization/48189
19827 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19828 If nitercst is 0, don't predict the exit edge.
19830 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19832 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19833 in asm_fprintf with reg_names.
19834 (aarch64_print_operand_address): Likewise.
19835 (aarch64_return_addr): Likewise.
19836 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19838 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19840 * config/pa/pa.h (VAL_U6_BITS_P): Define.
19841 (INT_U6_BITS): Likewise.
19842 * config/pa/predicates.md (uint6_operand): New predicate.
19843 (shift5_operand, shift6_operand): Likewise.
19844 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
19846 (lshrdi3): Use shift6_operand.
19847 (shrpsi4, shrpdi4): New insn patterns.
19848 (extzv): Delete expander.
19849 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
19850 predicates in unamed zero extract patterns. Tighten common constraint.
19851 (extv): Delete expander.
19852 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
19853 predicates in unamed sign extract patterns. Tighten common constraint.
19854 (insv): Delete expander.
19855 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
19856 predicates in unamed insert patterns. Tighten common constraint.
19857 Change uint32_operand predicate to uint6_operand predicate in unamed
19858 DImode pattern to insert constant values of type 1...1xxxx.
19860 2013-01-04 Jan Hubicka <jh@suse.cz>
19862 PR tree-optimization/55823
19863 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
19866 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19867 Uros Bizjak <ubizjak@gmail.com>
19869 PR rtl-optimization/55845
19870 * df-problems.c (can_move_insns_across): Stop scanning at
19871 volatile_insn_p source instruction or give up if
19872 across_from .. across_to range contains any volatile_insn_p
19875 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19877 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
19878 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
19880 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
19881 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
19883 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19886 * asan.c (asan_clear_shadow): New function.
19887 (asan_emit_stack_protection): Use it.
19889 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19891 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
19892 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
19893 with tab instead of space.
19895 2013-01-08 Nick Clifton <nickc@redhat.com>
19897 * config/rl78/rl78.c (rl78_expand_prologue): Always select
19898 register bank 0 at the start of an interrupt handler.
19899 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
19902 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19904 * config/aarch64/aarch64-simd.md
19905 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
19906 (aarch64_simd_bsl): Likewise.
19907 (aarch64_vcond_internal<mode>): Likewise.
19908 (vcond<mode><mode>): Likewise.
19909 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
19910 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
19912 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19914 * config/aarch64/aarch64-builtins.c
19915 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
19917 2013-01-08 Martin Jambor <mjambor@suse.cz>
19920 * tree-sra.c (analyze_access_subtree): Return true also after
19921 potentially creating a debug-only replacement.
19923 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19925 PR middle-end/55890
19926 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
19928 PR tree-optimization/54120
19929 * tree-vrp.c (range_fits_type_p): Don't allow
19930 src_precision < precision from signed vr to unsigned_p
19931 if vr->min or vr->max is negative.
19932 (simplify_float_conversion_using_ranges): Test can_float_p
19933 against CODE_FOR_nothing.
19935 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19936 Richard Biener <rguenther@suse.de>
19938 PR middle-end/55851
19939 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
19940 types instead of just INTEGER_TYPE types.
19942 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
19944 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
19947 2013-01-07 Steve Ellcey <sellcey@mips.com>
19950 * config/mips/mips.opt: Change mad to mmad to match documentation.
19952 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19955 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
19956 .progmemx.data now.
19958 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19961 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
19962 (avr_addrspace_t): Add .section_name field.
19963 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
19965 (avr_addrspace): Same. Initialize .section_name. Remove last
19966 NULL entry. Put __memx into .progmemx.data.
19967 (progmem_section_prefix): Remove.
19968 (avr_asm_init_sections): No need to initialize progmem_section.
19969 (avr_asm_named_section): Use avr_addrspace[].section_name to get
19970 section name prefix.
19971 (avr_asm_select_section): Ditto. And use get_unnamed_section to
19972 retrieve the progmem section.
19973 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
19974 boundary to run over avr_addrspace[].
19975 (avr_register_target_pragmas): Ditto.
19977 2013-01-06 Jakub Jelinek <jakub@redhat.com>
19979 * varasm.c (output_constant_def_contents): For asan_protect_global
19980 protected strings, adjust DECL_ALIGN if needed, before testing for
19982 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
19983 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
19985 (output_object_block): For asan protected decls, emit asan padding
19986 after their contents.
19987 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
19988 (asan_finish_file): Test it here instead.
19990 2013-01-07 Nick Clifton <nickc@redhat.com>
19991 Matthias Klose <doko@debian.org>
19992 Doug Kwan <dougkwan@google.com>
19993 H.J. Lu <hongjiu.lu@intel.com>
19996 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
19998 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
20000 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
20002 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
20004 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
20006 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20009 * doc/install.texi (Cross-Compiler-Specific Options): Document
20012 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
20014 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
20015 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
20016 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
20017 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
20018 vqmovun_high_s64): Fix source operand number and update copyright.
20020 2013-01-07 Richard Biener <rguenther@suse.de>
20022 PR middle-end/55890
20023 * gimple.h (gimple_call_builtin_p): New overload.
20024 * gimple.c (validate_call): New function.
20025 (gimple_call_builtin_p): Likewise.
20026 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20027 Use gimple_call_builtin_p.
20028 (find_func_clobbers): Likewise.
20029 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
20030 (strlen_optimize_stmt): Likewise.
20032 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20034 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
20035 (vld1q_dup_*): Likewise.
20036 (vld1_*): Likewise.
20037 (vld1q_*): Likewise.
20038 (vld1_lane_*): Likewise.
20039 (vld1q_lane_*): Likewise.
20041 2013-01-07 Richard Biener <rguenther@suse.de>
20043 * lto-streamer.h (LTO_minor_version): Bump to 2.
20045 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20047 * config/aarch64/aarch64-protos.h
20048 (aarch64_const_double_zero_rtx_p): Rename to...
20049 (aarch64_float_const_zero_rtx_p): ...this.
20050 (aarch64_float_const_representable_p): New.
20051 (aarch64_output_simd_mov_immediate): Likewise.
20052 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
20053 move immediate case.
20054 * config/aarch64/aarch64.c
20055 (aarch64_const_double_zero_rtx_p): Rename to...
20056 (aarch64_float_const_zero_rtx_p): ...this.
20057 (aarch64_print_operand): Allow printing of new constants.
20058 (aarch64_valid_floating_const): New.
20059 (aarch64_legitimate_constant_p): Check for valid floating-point
20061 (aarch64_simd_valid_immediate): Likewise.
20062 (aarch64_vect_float_const_representable_p): New.
20063 (aarch64_float_const_representable_p): Likewise.
20064 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
20065 (aarch64_output_simd_mov_immediate): New.
20066 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
20067 (*movdf_aarch64): Likewise.
20068 * config/aarch64/constraints.md (Ufc): New.
20069 (Y): call aarch64_float_const_zero_rtx.
20070 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
20072 2013-01-07 Richard Biener <rguenther@suse.de>
20074 PR tree-optimization/55888
20075 PR tree-optimization/55862
20076 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
20077 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
20078 not if it is contained therein.
20080 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20082 * config/avr/t-avr: Typo.
20084 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20087 * config/avr/t-avr: Don't automatically rebuild
20088 $(srcdir)/config/avr/t-multilib
20089 $(srcdir)/config/avr/avr-tables.opt
20090 $(srcdir)/doc/avr-mmcu.texi
20091 (avr-mcus): New phony target to build them on request.
20092 (s-avr-mlib, s-avr-mmcu-texi): Remove.
20093 * avr/avr-mcus.def: Adjust comments.
20095 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
20097 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
20099 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20101 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
20103 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20105 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
20107 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
20109 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
20110 to generate profiling.
20111 * config/rs6000/aix64.h (LIB_SPEC): Same.
20113 2013-01-04 Andrew Pinski <apinski@cavium.com>
20115 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
20117 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
20119 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
20121 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
20123 (ix86_expand_move): Ditto.
20124 (ix86_zero_extend_to_Pmode): Ditto.
20125 (ix86_expand_call): Ditto.
20126 (ix86_expand_special_args_builtin): Ditto.
20127 (ix86_expand_builtin): Ditto.
20129 2013-01-04 Richard Biener <rguenther@suse.de>
20131 PR tree-optimization/55862
20132 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
20133 translating them through PHI nodes.
20135 2013-01-04 Martin Jambor <mjambor@suse.cz>
20137 PR tree-optimization/55755
20138 * tree-sra.c (sra_modify_assign): Do not check that an access has no
20139 children when trying to avoid producing a VIEW_CONVERT_EXPR.
20141 2013-01-04 Marek Polacek <polacek@redhat.com>
20143 PR middle-end/55859
20144 * opts.c (default_options_optimization): Clarify error message.
20146 2013-01-04 Richard Biener <rguenther@suse.de>
20148 PR middle-end/55863
20149 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
20152 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20155 * config/pa/pa.md (movsi): Revert previous change.
20156 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
20159 2013-01-03 Richard Henderson <rth@redhat.com>
20161 * config/i386/i386.c (ix86_expand_move): Always assign to op1
20162 after eliminating TLS symbols.
20164 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
20167 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
20168 * graphite-poly.c (debug_gmp_value): Likewise.
20170 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
20173 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
20174 selected code model, define __code_mode_small__, __code_model_medium__,
20175 __code_model_large__, __code_model_32__ or __code_model_kernel__.
20176 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
20177 xchg temporary register with %k. Declare temporary register as
20179 [__x86_64__]: For medium and large code models, preserve %rbx register.
20181 2013-01-03 Richard Biener <rguenther@suse.de>
20183 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
20184 (dump_subscript): Adjust.
20185 (finalize_ddr_dependent): Do not dump redundant info.
20186 (analyze_siv_subscript): Adjust.
20187 (subscript_dependence_tester): Likewise.
20188 (compute_affine_dependence): Likewise.
20190 2013-01-03 Richard Biener <rguenther@suse.de>
20193 2013-01-03 Richard Biener <rguenther@suse.de>
20195 PR tree-optimization/55857
20196 * tree-vect-stmts.c (vectorizable_load): Do not setup
20197 re-alignment for invariant loads.
20199 2013-01-02 Richard Biener <rguenther@suse.de>
20201 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
20202 invariant load do not generate a vector load from the scalar location.
20204 2013-01-03 Richard Biener <rguenther@suse.de>
20206 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
20207 for not vectorizing.
20208 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
20209 not build INDIRECT_REFs, call get_name once only.
20210 (vect_create_data_ref_ptr): Likewise. Dump base object kind
20211 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
20213 2013-01-03 Richard Biener <rguenther@suse.de>
20215 PR tree-optimization/55857
20216 * tree-vect-stmts.c (vectorizable_load): Do not setup
20217 re-alignment for invariant loads.
20219 2013-01-03 Richard Biener <rguenther@suse.de>
20222 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
20223 prefer a built-in decl.
20225 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20227 * gcc.c (process_command): Update copyright notice dates.
20228 * gcov.c (print_version): Likewise.
20229 * gcov-dump.c (print_version): Likewise.
20231 PR rtl-optimization/55838
20232 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
20233 iv0.step, iv1.step and step.
20235 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20236 Marc Glisse <marc.glisse@inria.fr>
20238 PR tree-optimization/55832
20239 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
20240 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
20241 integer_{one,zero}_node.
20243 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20246 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
20247 * var-tracking.c (reverse_op): Don't add reverse ops to
20248 VALUEs that have already
20249 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
20251 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
20253 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
20255 2013-01-02 Teresa Johnson <tejohnson@google.com>
20257 * dumpfile.c (dump_loc): Print filename with location.
20258 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
20259 new location_t parameter to emit complete unroll message with
20260 new dump framework.
20261 (canonicalize_loop_induction_variables): Compute loops location
20262 and pass to try_unroll_loop_completely.
20263 * loop-unroll.c (report_unroll_peel): New function.
20264 (peel_loops_completely): Use new dump format with location
20265 for main dumpfile message, and invoke report_unroll_peel on success.
20266 (decide_unrolling_and_peeling): Ditto.
20267 (decide_peel_once_rolling): Remove old dumpfile message subsumed
20268 by report_unroll_peel.
20269 (decide_peel_completely): Ditto.
20270 (decide_unroll_constant_iterations): Ditto.
20271 (decide_unroll_runtime_iterations): Ditto.
20272 (decide_peel_simple): Ditto.
20273 (decide_unroll_stupid): Ditto.
20274 * cfgloop.c (get_loop_location): New function.
20275 * cfgloop.h (get_loop_location): Declare.
20277 2013-01-02 Sriraman Tallam <tmsriram@google.com>
20279 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
20282 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20284 PR middle-end/55198
20285 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
20286 BLKmode objects when EXPAND_MEMORY is specified.
20288 2013-01-02 Sriraman Tallam <tmsriram@google.com>
20290 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
20292 (fold_builtin_cpu): Do not share cpu model decls across statements.
20294 2013-01-02 Jason Merrill <jason@redhat.com>
20297 * tree.c (build_array_type_1): Revert earlier change.
20299 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
20301 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
20303 * config/aarch64/aarch64-tune.md: Re-generate.
20305 2013-01-02 Richard Biener <rguenther@suse.de>
20307 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
20308 invariant load do not generate a vector load from the scalar location.
20310 2013-01-02 Richard Biener <rguenther@suse.de>
20313 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
20314 * configure: Regenerate.
20316 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20318 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
20319 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
20320 (expand_builtin_int_roundingfn_2): Keep the original target around
20321 for the fallback case.
20323 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20325 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
20326 to be clear for sign changes.
20328 2013-01-01 Jan Hubicka <jh@suse.cz>
20330 * ipa-inline-analysis.c: Fix formatting.
20332 2013-01-01 Jakub Jelinek <jakub@redhat.com>
20334 PR tree-optimization/55831
20335 * tree-vect-loop.c (get_initial_def_for_induction): Use
20336 gsi_after_labels instead of gsi_start_bb.
20338 Copyright (C) 2013 Free Software Foundation, Inc.
20340 Copying and distribution of this file, with or without modification,
20341 are permitted in any medium without royalty provided the copyright
20342 notice and this notice are preserved.