1 2014-05-09 Florian Weimer <fweimer@redhat.com>
3 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
5 (stack_protect_return_slot_p): New function.
6 (expand_used_vars): Call stack_protect_decl_p and
7 stack_protect_return_slot_p for -fstack-protector-strong.
9 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
12 * fold-const.c (fold_binary_loc): Changed width of mask.
14 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
16 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
17 unsigned int initializers for regno_in, regno_out.
19 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
22 * config/avr/avr.md (cc): Add new attribute set_vzn.
23 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
24 Set cc insn attribute to set_vzn instead of set_zn for alternatives
26 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
27 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
28 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
30 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
33 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
35 * wide-int.cc (UTItype): Define.
36 (UDWtype): Define for appropriate W_TYPE_SIZE.
38 2014-05-09 Richard Biener <rguenther@suse.de>
40 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
41 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
42 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
43 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
44 ssa_propagate): Adjust.
46 2014-05-08 Jeff Law <law@redhat.com>
48 PR tree-optimization/61009
49 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
50 tri-state rather than a boolean. When a block is too big to
51 thread through, inform caller via negative return value.
52 (thread_across_edge): If a block was too big for normal threading,
53 then it's too big for a joiner too, so remove temporary equivalences
54 and return immediately.
56 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
57 Matthias Klose <doko@ubuntu.com>
60 * optc-gen.awk: Fix option handling for -Wunused-parameter.
62 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
65 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
67 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
70 * config/alpha/alpha.c: Include gimple-iterator.h.
71 (alpha_gimple_fold_builtin): New function. Move
72 ALPHA_BUILTIN_UMULH folding from ...
73 (alpha_fold_builtin): ... here.
74 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
76 2014-05-08 Wei Mi <wmi@google.com>
79 * config/i386/i386.c (ix86_compute_frame_layout): Update
80 preferred_stack_boundary for call, expanded from tls descriptor.
81 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
82 to depend on SP register.
83 (*tls_local_dynamic_base_32_gnu): Ditto.
84 (*tls_local_dynamic_32_once): Ditto.
85 (tls_global_dynamic_64_<mode>): Set
86 ix86_tls_descriptor_calls_expanded_in_cfun.
87 (tls_local_dynamic_base_64_<mode>): Ditto.
88 (tls_global_dynamic_32): Set
89 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
90 to depend on SP register.
91 (tls_local_dynamic_base_32): Ditto.
93 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
95 * config/arm/arm_neon.h: Update comment.
96 * config/arm/neon-docgen.ml: Delete.
97 * config/arm/neon-gen.ml: Delete.
98 * doc/arm-neon-intrinsics.texi: Update comment.
100 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
102 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
104 (vand, vorr, veor, vorn, vbic): Remove.
105 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
107 (neon_vsub_unspec): Likewise.
108 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
110 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
112 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
113 (vadd_s16): Likewise.
114 (vadd_s32): Likewise.
115 (vadd_f32): Likewise.
117 (vadd_u16): Likewise.
118 (vadd_u32): Likewise.
119 (vadd_s64): Likewise.
120 (vadd_u64): Likewise.
121 (vaddq_s8): Likewise.
122 (vaddq_s16): Likewise.
123 (vaddq_s32): Likewise.
124 (vaddq_s64): Likewise.
125 (vaddq_f32): Likewise.
126 (vaddq_u8): Likewise.
127 (vaddq_u16): Likewise.
128 (vaddq_u32): Likewise.
129 (vaddq_u64): Likewise.
131 (vmul_s16): Likewise.
132 (vmul_s32): Likewise.
133 (vmul_f32): Likewise.
135 (vmul_u16): Likewise.
136 (vmul_u32): Likewise.
138 (vmulq_s8): Likewise.
139 (vmulq_s16): Likewise.
140 (vmulq_s32): Likewise.
141 (vmulq_f32): Likewise.
142 (vmulq_u8): Likewise.
143 (vmulq_u16): Likewise.
144 (vmulq_u32): Likewise.
146 (vsub_s16): Likewise.
147 (vsub_s32): Likewise.
148 (vsub_f32): Likewise.
150 (vsub_u16): Likewise.
151 (vsub_u32): Likewise.
152 (vsub_s64): Likewise.
153 (vsub_u64): Likewise.
154 (vsubq_s8): Likewise.
155 (vsubq_s16): Likewise.
156 (vsubq_s32): Likewise.
157 (vsubq_s64): Likewise.
158 (vsubq_f32): Likewise.
159 (vsubq_u8): Likewise.
160 (vsubq_u16): Likewise.
161 (vsubq_u32): Likewise.
162 (vsubq_u64): Likewise.
164 (vand_s16): Likewise.
165 (vand_s32): Likewise.
167 (vand_u16): Likewise.
168 (vand_u32): Likewise.
169 (vand_s64): Likewise.
170 (vand_u64): Likewise.
171 (vandq_s8): Likewise.
172 (vandq_s16): Likewise.
173 (vandq_s32): Likewise.
174 (vandq_s64): Likewise.
175 (vandq_u8): Likewise.
176 (vandq_u16): Likewise.
177 (vandq_u32): Likewise.
178 (vandq_u64): Likewise.
180 (vorr_s16): Likewise.
181 (vorr_s32): Likewise.
183 (vorr_u16): Likewise.
184 (vorr_u32): Likewise.
185 (vorr_s64): Likewise.
186 (vorr_u64): Likewise.
187 (vorrq_s8): Likewise.
188 (vorrq_s16): Likewise.
189 (vorrq_s32): Likewise.
190 (vorrq_s64): Likewise.
191 (vorrq_u8): Likewise.
192 (vorrq_u16): Likewise.
193 (vorrq_u32): Likewise.
194 (vorrq_u64): Likewise.
196 (veor_s16): Likewise.
197 (veor_s32): Likewise.
199 (veor_u16): Likewise.
200 (veor_u32): Likewise.
201 (veor_s64): Likewise.
202 (veor_u64): Likewise.
203 (veorq_s8): Likewise.
204 (veorq_s16): Likewise.
205 (veorq_s32): Likewise.
206 (veorq_s64): Likewise.
207 (veorq_u8): Likewise.
208 (veorq_u16): Likewise.
209 (veorq_u32): Likewise.
210 (veorq_u64): Likewise.
212 (vbic_s16): Likewise.
213 (vbic_s32): Likewise.
215 (vbic_u16): Likewise.
216 (vbic_u32): Likewise.
217 (vbic_s64): Likewise.
218 (vbic_u64): Likewise.
219 (vbicq_s8): Likewise.
220 (vbicq_s16): Likewise.
221 (vbicq_s32): Likewise.
222 (vbicq_s64): Likewise.
223 (vbicq_u8): Likewise.
224 (vbicq_u16): Likewise.
225 (vbicq_u32): Likewise.
226 (vbicq_u64): Likewise.
228 (vorn_s16): Likewise.
229 (vorn_s32): Likewise.
231 (vorn_u16): Likewise.
232 (vorn_u32): Likewise.
233 (vorn_s64): Likewise.
234 (vorn_u64): Likewise.
235 (vornq_s8): Likewise.
236 (vornq_s16): Likewise.
237 (vornq_s32): Likewise.
238 (vornq_s64): Likewise.
239 (vornq_u8): Likewise.
240 (vornq_u16): Likewise.
241 (vornq_u32): Likewise.
242 (vornq_u64): Likewise.
244 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
246 * wide-int.cc (UTItype): Define.
247 (UDWtype): Define for appropriate W_TYPE_SIZE.
249 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
251 PR tree-optimization/59100
252 * tree-ssa-phiopt.c: Include tree-inline.h.
253 (neutral_element_p, absorbing_element_p): New functions.
254 (value_replacement): Handle conditional binary operations with a
255 neutral or absorbing element.
257 2014-05-08 Richard Biener <rguenther@suse.de>
259 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
260 folding the expression.
261 (valueize_expr): Remove.
262 (visit_reference_op_load): Do not valueize the result of
264 (simplify_binary_expression): Likewise.
265 (simplify_unary_expression): Likewise.
267 2014-05-08 Richard Biener <rguenther@suse.de>
269 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
270 looking at TYPE_ARG_TYPES.
272 2014-05-08 Richard Biener <rguenther@suse.de>
274 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
275 pointer propagation special-case.
277 2014-05-08 Bin Cheng <bin.cheng@arm.com>
279 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
280 core part of address expressions.
282 2014-05-08 Alan Modra <amodra@gmail.com>
285 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
286 loads and stores when -mno-strict-align at any alignment.
287 (expand_block_clear): Similarly. Also correct calculation of
290 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
293 * tree-complex.c (expand_complex_move): Keep line info when expanding
295 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
296 of complex expression. Use new argument to display correct location
297 for values coming from phi statement.
298 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
299 (warn_uninitialized_phi): Pass location of phi argument to
301 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
302 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
304 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
306 * config/rs6000/predicates.md (indexed_address_mem): New.
307 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
308 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
309 fpstore_ux, fpstore_u.
310 (sign_extend, indexed, update): New.
311 (cell_micro): Adjust.
312 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
313 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
314 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
315 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
316 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
317 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
318 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
319 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
320 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
321 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
322 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
323 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
324 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
325 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
326 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
328 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
329 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
330 *vsx_extract_<mode>_store): Adjust.
331 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
332 is_cracked_insn, insn_must_be_first_in_group,
333 insn_must_be_last_in_group): Adjust.
335 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
337 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
338 ppc440-fpstore): Adjust.
339 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
340 ppc476-fpstore): Adjust.
341 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
342 ppc601-fpstore): Adjust.
343 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
345 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
347 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
348 ppc7450-fpstore): Adjust.
349 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
350 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
351 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
353 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
354 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
355 cell-fpstore, cell-fpstore-update): Adjust.
356 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
357 ppce300c3_store, ppce300c3_fpstore): Adjust.
358 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
359 e500mc_fpstore): Adjust.
360 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
361 e500mc64_store, e500mc64_fpstore): Adjust.
362 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
363 e5500_fpstore): Adjust.
364 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
365 e6500_fpstore): Adjust.
366 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
368 * config/rs6000/power4.md (power4-load, power4-load-ext,
369 power4-load-ext-update, power4-load-ext-update-indexed,
370 power4-load-update-indexed, power4-load-update, power4-fpload,
371 power4-fpload-update, power4-store, power4-store-update,
372 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
374 * config/rs6000/power5.md (power5-load, power5-load-ext,
375 power5-load-ext-update, power5-load-ext-update-indexed,
376 power5-load-update-indexed, power5-load-update, power5-fpload,
377 power5-fpload-update, power5-store, power5-store-update,
378 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
380 * config/rs6000/power6.md (power6-load, power6-load-ext,
381 power6-load-update, power6-load-update-indexed,
382 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
383 power6-fpload-update, power6-store, power6-store-update,
384 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
386 * config/rs6000/power7.md (power7-load, power7-load-ext,
387 power7-load-update, power7-load-update-indexed,
388 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
389 power7-fpload-update, power7-store, power7-store-update,
390 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
392 * config/rs6000/power8.md (power8-load, power8-load-update,
393 power8-load-ext, power8-load-ext-update, power8-fpload,
394 power8-fpload-update, power8-store, power8-store-update-indexed,
395 power8-fpstore, power8-fpstore-update): Adjust.
396 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
398 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
399 titan_lsu_store, titan_lsu_fpstore): Adjust.
400 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
402 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
405 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
406 unrolled byte insns. Emit address increments after move insns.
408 2014-05-07 David Malcolm <dmalcolm@redhat.com>
410 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
411 const_gimple, rather than a gimple.
412 (gimple_call_builtin_p): Likewise, for the three variants.
414 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
415 (gimple_call_builtin_p): Likewise, for the three variants.
417 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
419 PR tree-optimization/61095
420 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
422 2014-05-07 Richard Biener <rguenther@suse.de>
424 PR tree-optimization/61034
425 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
426 (maybe_skip_until): Use translate to take into account
427 lattices when trying to do disambiguations.
428 (get_continuation_for_phi_1): Likewise.
429 (get_continuation_for_phi): Adjust for added translate arguments.
430 (walk_non_aliased_vuses): Likewise.
431 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
432 (walk_non_aliased_vuses): Likewise.
433 (call_may_clobber_ref_p_1): Declare.
434 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
435 calls. Stop early if we are only supposed to disambiguate.
436 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
438 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
440 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
441 Emit an error when the function has arguments.
443 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
445 * cfgloop.h (unswitch_loops): Remove.
446 * doc/passes.texi: Remove references to loop-unswitch.c
447 * timevar.def (TV_LOOP_UNSWITCH): Remove.
449 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
451 * tree-vect-data-refs.c (vect_grouped_load_supported): New
452 check for loads group of length 3.
453 (vect_permute_load_chain): New permutations for loads group of
455 * tree-vect-stmts.c (vect_model_load_cost): Change cost
456 of vec_perm_shuffle for the new permutations.
458 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
460 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
461 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
462 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
463 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
464 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
465 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
466 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
467 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
469 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
471 * loop-unswitch.c: Delete.
473 2014-05-07 Richard Biener <rguenther@suse.de>
475 * config.gcc: Always set need_64bit_hwint to yes.
477 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
479 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
480 of using optimize_size.
482 2014-05-06 Mike Stump <mikestump@comcast.net>
484 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
486 2014-05-06 Joseph Myers <joseph@codesourcery.com>
488 * config/i386/sse.md (*mov<mode>_internal)
489 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
490 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
491 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
492 (*<code><mode>3, *andnot<mode>3<mask_name>)
493 (<mask_codefor><code><mode>3<mask_name>): Only consider
494 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
496 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
499 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
501 * lra-constraints.c (valid_address_p): Move earlier in file.
502 Add a constraint argument to the address_info version.
503 (satisfies_memory_constraint_p): New function.
504 (satisfies_address_constraint_p): Likewise.
505 (process_alt_operands, curr_insn_transform): Use them.
506 (process_address): Pass the constraint to valid_address_p when
507 checking address operands.
509 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
511 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
512 to their respective blocks. Fix inadvertent use of "node".
514 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
516 * emit-rtl.c (init_derived_machine_modes): New functionm, split
518 (init_emit_once): ...here.
519 * rtl.h (init_derived_machine_modes): Declare.
520 * toplev.c (do_compile): Call it even if no_backend.
522 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
523 Mike Stump <mikestump@comcast.net>
524 Richard Sandiford <rdsandiford@googlemail.com>
525 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
527 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
528 (rtx_equal_for_memref_p): Update comment.
529 (adjust_offset_for_component_ref): Use wide-int interfaces.
530 * builtins.c (get_object_alignment_2): Likewise.
531 (c_readstr): Likewise.
532 (target_char_cast): Add comment.
533 (determine_block_size): Use wide-int interfaces.
534 (expand_builtin_signbit): Likewise.
535 (fold_builtin_int_roundingfn): Likewise.
536 (fold_builtin_bitop): Likewise.
537 (fold_builtin_bswap): Likewise.
538 (fold_builtin_logarithm): Use signop.
539 (fold_builtin_pow): Likewise.
540 (fold_builtin_memory_op): Use wide-int interfaces.
541 (fold_builtin_object_size): Likewise.
542 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
543 nb_iterations_estimate.
544 (record_niter_bound): Use wide-int interfaces.
545 (get_estimated_loop_iterations_int): Likewise.
546 (get_estimated_loop_iterations): Likewise.
547 (get_max_loop_iterations): Likewise.
548 * cfgloop.h: Include wide-int.h.
549 (struct nb_iter_bound): Change bound to widest_int.
550 (struct loop): Change nb_iterations_upper_bound and
551 nb_iterations_estimate to widest_int.
552 (record_niter_bound): Switch to use widest_int.
553 (get_estimated_loop_iterations): Likewise.
554 (get_max_loop_iterations): Likewise.
555 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
557 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
558 * combine.c (try_combine): Likewise.
559 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
560 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
562 (aarch64_float_const_representable_p): Likewise.
563 * config/arc/arc.c: Include wide-int.h.
564 (arc_can_use_doloop_p): Use wide-int interfaces.
565 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
566 (vfp3_const_double_index): Likewise.
567 * config/avr/avr.c (avr_out_round): Likewise.
568 (avr_fold_builtin): Likewise.
569 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
570 (bfin_can_use_doloop_p): Likewise.
571 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
572 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
573 * config/i386/i386.c: Include wide-int.h.
574 (ix86_data_alignment): Use wide-int interfaces.
575 (ix86_local_alignment): Likewise.
576 (ix86_emit_swsqrtsf): Update real_from_integer.
577 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
578 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
579 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
580 (zero_constant): Likewise.
581 (input_operand): Likewise.
582 (splat_input_operand): Likewise.
583 (non_logical_cint_operand): Change const_double to const_wide_int.
584 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
585 (easy_altivec_constant): Remove comment.
586 (paired_expand_vector_init): Use CONSTANT_P.
587 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
588 (rs6000_emit_move): Update checks.
589 (rs6000_aggregate_candidate): Use wide-int interfaces.
590 (rs6000_expand_ternop_builtin): Likewise.
591 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
592 (rs6000_assemble_integer): Likewise.
593 (rs6000_hash_constant): Likewise.
594 (output_toc): Likewise.
595 (rs6000_rtx_costs): Likewise.
596 (rs6000_emit_swrsqrt); Update call to real_from_integer.
597 * config/rs6000/rs6000-c.c: Include wide-int.h.
598 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
599 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
600 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
601 Handle CONST_WIDE_INT.
602 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
603 Use tree_fits_uhwi_p.
604 * config/sparc/sparc.c: Include wide-int.h.
605 (sparc_fold_builtin): Use wide-int interfaces.
606 * config/vax/vax.c: Include wide-int.h.
607 (vax_float_literal): Use real_from_integer.
608 * coretypes.h (struct hwivec_def): New.
611 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
612 (equiv_constant): Handle CONST_WIDE_INT.
613 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
614 (cselib_hash_rtx): Handle CONST_WIDE_INT.
615 * dbxout.c (stabstr_U): Use wide-int interfaces.
616 (dbxout_type): Update to use cst_fits_shwi_p.
617 * defaults.h (LOG2_BITS_PER_UNIT): Define.
618 (TARGET_SUPPORTS_WIDE_INT): Add default.
619 * dfp.c: Include wide-int.h.
620 (decimal_real_to_integer2): Use wide-int interfaces and rename to
621 decimal_real_to_integer.
622 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
623 decimal_real_to_integer.
624 * doc/generic.texi (Constant expressions): Update for wide_int.
625 * doc/rtl.texi (const_double): Likewise.
626 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
627 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
628 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
629 (REAL_VALUE_FROM_INT): Remove.
630 (TARGET_SUPPORTS_WIDE_INT): New.
631 * doc/tm.texi: Regenerate.
632 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
633 * double-int.h: Include wide-int.h.
634 (struct wi::int_traits): New.
635 * dwarf2out.c (get_full_len): New.
636 (dw_val_equal_p): Add case dw_val_class_wide_int.
637 (size_of_loc_descr): Likewise.
638 (output_loc_operands): Likewise.
639 (insert_double): Remove.
640 (insert_wide_int): New.
642 (print_die): Add case dw_val_class_wide_int.
643 (attr_checksum): Likewise.
644 (attr_checksum_ordered): Likewise.
645 (same_dw_val_p): Likewise.
646 (size_of_die): Likewise.
647 (value_format): Likewise.
648 (output_die): Likewise.
649 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
651 (clz_loc_descriptor): Use wide-int interfaces.
652 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
653 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
654 (round_up_to_align): Use wide-int interfaces.
655 (field_byte_offset): Likewise.
656 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
657 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
658 CONST_DOUBLE handling. Use wide-int interfaces.
659 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
660 (gen_enumeration_type_die): Use add_AT_wide.
661 (hash_loc_operands): Add case dw_val_class_wide_int.
662 (compare_loc_operands): Likewise.
663 * dwarf2out.h: Include wide-int.h.
665 (enum dw_val_class): Add dw_val_class_wide_int.
666 (struct dw_val_struct): Add val_wide.
667 * emit-rtl.c (const_wide_int_htab): New.
668 (const_wide_int_htab_hash): New.
669 (const_wide_int_htab_eq): New.
670 (lookup_const_wide_int): New.
671 (const_double_htab_hash): Use wide-int interfaces.
672 (const_double_htab_eq): Likewise.
673 (rtx_to_double_int): Conditionally compile for wide-int.
674 (immed_double_int_const): Rename to immed_wide_int_const and
676 (immed_double_const): Conditionally compile for wide-int.
677 (init_emit_once): Use wide-int interfaces.
678 * explow.c (plus_constant): Likewise.
679 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
680 (lshift_value): Use wide-int interfaces.
681 (expand_mult): Likewise.
682 (choose_multiplier): Likewise.
683 (expand_smod_pow2): Likewise.
684 (make_tree): Likewise.
685 * expr.c (convert_modes): Consolidate handling of constants.
686 Use wide-int interfaces.
687 (emit_group_load_1): Add note.
688 (store_expr): Update comment.
689 (get_inner_reference): Use wide-int interfaces.
690 (expand_constructor): Update comment.
691 (expand_expr_real_2): Use wide-int interfaces.
692 (expand_expr_real_1): Likewise.
693 (reduce_to_bit_field_precision): Likewise.
694 (const_vector_from_tree): Likewise.
695 * final.c: Include wide-int-print.h.
696 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
697 * fixed-value.c: Include wide-int.h.
698 (fixed_from_string): Use wide-int interfaces.
699 (fixed_to_decimal): Likewise.
700 (fixed_convert_from_real): Likewise.
701 (real_convert_from_fixed): Likewise.
702 * fold-const.h (mem_ref_offset): Return an offset_int.
703 (div_if_zero_remainder): Remove code parameter.
704 * fold-const.c (div_if_zero_remainder): Remove code parameter.
705 Use wide-int interfaces.
706 (may_negate_without_overflow_p): Use wide-int interfaces.
707 (negate_expr_p): Likewise.
708 (fold_negate_expr): Likewise.
709 (int_const_binop_1): Likewise.
710 (const_binop): Likewise.
711 (fold_convert_const_int_from_int): Likewise.
712 (fold_convert_const_int_from_real): Likewise.
713 (fold_convert_const_int_from_fixed): Likewise.
714 (fold_convert_const_fixed_from_int): Likewise.
715 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
716 (sign_bit_p): Use wide-int interfaces.
717 (make_range_step): Likewise.
718 (build_range_check): Likewise. Pass an integer of the correct type
719 instead of using integer_one_node.
720 (range_predecessor): Pass an integer of the correct type instead
721 of using integer_one_node.
722 (range_successor): Likewise.
723 (merge_ranges): Likewise.
724 (unextend): Use wide-int interfaces.
725 (extract_muldiv_1): Likewise.
726 (fold_div_compare): Likewise.
727 (fold_single_bit_test): Likewise.
728 (fold_sign_changed_comparison): Likewise.
729 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
730 (fold_plusminus_mult_expr): Use wide-int interfaces.
731 (native_encode_int): Likewise.
732 (native_interpret_int): Likewise.
733 (fold_unary_loc): Likewise.
734 (pointer_may_wrap_p): Likewise.
735 (size_low_cst): Likewise.
736 (mask_with_tz): Likewise.
737 (fold_binary_loc): Likewise.
738 (fold_ternary_loc): Likewise.
739 (multiple_of_p): Likewise.
740 (tree_call_nonnegative_warnv_p): Update calls to
741 tree_int_cst_min_precision and real_from_integer.
742 (fold_negate_const): Use wide-int interfaces.
743 (fold_abs_const): Likewise.
744 (fold_relational_const): Use tree_int_cst_lt.
745 (round_up_loc): Use wide-int interfaces.
746 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
747 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
748 * gengtype.c: Remove include of double-int.h.
749 (do_typedef): Use wide-int interfaces.
750 (open_base_files): Add wide-int.h.
751 (main): Add offset_int and widest_int typedefs.
752 * gengtype-lex.l: Handle "^".
753 (CXX_KEYWORD): Add "static".
754 * gengtype-parse.c (require3): New.
755 (require_template_declaration): Handle constant template arguments
756 and nested templates.
757 * gengtype-state.c: Don't include "double-int.h".
758 * genpreds.c (write_one_predicate_function): Update comment.
759 (write_tm_constrs_h): Add check for hval and lval use in
761 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
762 (add_to_sequence): Likewise.
763 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
764 and const_double_operand.
765 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
767 * gimple-fold.c (get_base_constructor): Likewise.
768 (fold_array_ctor_reference): Likewise.
769 (fold_nonarray_ctor_reference): Likewise.
770 (fold_const_aggregate_ref_1): Likewise.
771 (gimple_val_nonnegative_real_p): Likewise.
772 (gimple_fold_indirect_ref): Likewise.
773 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
774 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
775 (struct slsr_cand_d): Change index to be widest_int.
776 (struct incr_info_d): Change incr to be widest_int.
777 (alloc_cand_and_find_basis): Use wide-int interfaces.
778 (slsr_process_phi): Likewise.
779 (backtrace_base_for_ref): Likewise. Return a widest_int.
780 (restructure_reference): Take a widest_int instead of a double_int.
781 (slsr_process_ref): Use wide-int interfaces.
782 (create_mul_ssa_cand): Likewise.
783 (create_mul_imm_cand): Likewise.
784 (create_add_ssa_cand): Likewise.
785 (create_add_imm_cand): Take a widest_int instead of a double_int.
786 (slsr_process_add): Use wide-int interfaces.
787 (slsr_process_cast): Likewise.
788 (slsr_process_copy): Likewise.
789 (dump_candidate): Likewise.
790 (dump_incr_vec): Likewise.
791 (replace_ref): Likewise.
792 (cand_increment): Likewise. Return a widest_int.
793 (cand_abs_increment): Likewise.
794 (replace_mult_candidate): Take a widest_int instead of a double_int.
795 (replace_unconditional_candidate): Use wide-int interfaces.
796 (incr_vec_index): Take a widest_int instead of a double_int.
797 (create_add_on_incoming_edge): Likewise.
798 (create_phi_basis): Use wide-int interfaces.
799 (replace_conditional_candidate): Likewise.
800 (record_increment): Take a widest_int instead of a double_int.
801 (record_phi_increments): Use wide-int interfaces.
802 (phi_incr_cost): Take a widest_int instead of a double_int.
803 (lowest_cost_path): Likewise.
804 (total_savings): Likewise.
805 (analyze_increments): Use wide-int interfaces.
806 (ncd_with_phi): Take a widest_int instead of a double_int.
807 (ncd_of_cand_and_phis): Likewise.
808 (nearest_common_dominator_for_cands): Likewise.
809 (insert_initializers): Use wide-int interfaces.
810 (all_phi_incrs_profitable): Likewise.
811 (replace_one_candidate): Likewise.
812 (replace_profitable_candidates): Likewise.
813 * godump.c: Include wide-int-print.h.
814 (go_output_typedef): Use wide-int interfaces.
815 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
816 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
817 (build_loop_iteration_domains): Likewise.
818 * hooks.h: Include wide-int.h rather than double-int.h.
819 (hook_bool_dint_dint_uint_bool_true): Delete.
820 (hook_bool_wint_wint_uint_bool_true): Declare.
821 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
822 (hook_bool_wint_wint_uint_bool_true): New.
823 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
825 (ubsan_expand_si_overflow_mul_check): Likewise.
826 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
827 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
828 (get_ancestor_addr_info): Likewise.
829 (ipa_modify_call_arguments): Likewise.
830 * loop-doloop.c (doloop_modify): Likewise.
831 (doloop_optimize): Likewise.
832 * loop-iv.c (iv_number_of_iterations): Likewise.
833 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
834 (unroll_loop_constant_iterations): Likewise.
835 (decide_unroll_runtime_iterations): Likewise.
836 (unroll_loop_runtime_iterations): Likewise.
837 (decide_peel_simple): Likewise.
838 (decide_unroll_stupid): Likewise.
839 * lto-streamer-in.c (streamer_read_wi): Add.
840 (input_cfg): Use wide-int interfaces.
841 (lto_input_tree_1): Likewise.
842 * lto-streamer-out.c (streamer_write_wi): Add.
843 (hash_tree): Use wide-int interfaces.
844 (output_cfg): Likewise.
845 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
846 (GTFILES): Add wide-int.h and signop.h.
847 (TAGS): Look for .cc files too.
848 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
849 * optabs.c (expand_subword_shift): Likewise.
850 (expand_doubleword_shift): Likewise.
851 (expand_absneg_bit): Likewise.
852 (expand_copysign_absneg): Likewise.
853 (expand_copysign_bit): Likewise.
854 * postreload.c (reload_cse_simplify_set): Likewise.
855 * predict.c (predict_iv_comparison): Likewise.
856 * pretty-print.h: Include wide-int-print.h.
858 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
859 * print-tree.c: Include wide-int-print.h.
860 (print_node_brief): Use wide-int interfaces.
861 (print_node): Likewise.
862 * read-rtl.c (validate_const_wide_int): New.
863 (read_rtx_code): Add CONST_WIDE_INT case.
864 * real.c: Include wide-int.h.
865 (real_to_integer2): Delete.
866 (real_to_integer): New function, returning a wide_int.
867 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
868 (ten_to_ptwo): Update call to real_from_integer.
869 (real_digit): Likewise.
870 * real.h: Include signop.h, wide-int.h and insn-modes.h.
871 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
872 (REAL_VALUE_TO_INT): Delete.
873 (real_to_integer): Declare a wide-int form.
874 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
875 * recog.c (const_int_operand): Improve comment.
876 (const_scalar_int_operand): New.
877 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
878 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
879 (split_double): Likewise.
880 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
881 (rtx_size): Likewise.
882 (rtx_alloc_stat_v): New.
883 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
884 (cwi_output_hex): New.
885 (iterative_hash_rtx): Handle CONST_WIDE_INT.
886 (cwi_check_failed_bounds): New.
887 * rtl.def (CONST_WIDE_INT): New.
888 * rtl.h: Include <utility> and wide-int.h.
889 (struct hwivec_def): New.
890 (CWI_GET_NUM_ELEM): New.
891 (CWI_PUT_NUM_ELEM): New.
892 (struct rtx_def): Add num_elem and hwiv.
893 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
894 (CASE_CONST_UNIQUE): Likewise.
895 (CASE_CONST_ANY): Likewise.
896 (CONST_SCALAR_INT_P): Likewise.
897 (CONST_WIDE_INT_P): New.
900 (cwi_check_failed_bounds): New.
903 (CONST_WIDE_INT_VEC) New.
904 (CONST_WIDE_INT_NUNITS) New.
905 (CONST_WIDE_INT_ELT) New.
906 (rtx_mode_t): New type.
907 (wi::int_traits <rtx_mode_t>): New.
909 (wi::min_value): New.
910 (wi::max_value): New.
912 (const_wide_int_alloc): New.
913 (immed_wide_int_const): New.
914 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
915 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
916 * signop.h: New file.
917 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
918 (simplify_const_unary_operation): Use wide-int interfaces.
919 (simplify_binary_operation_1): Likewise.
920 (simplify_const_binary_operation): Likewise.
921 (simplify_const_relational_operation): Likewise.
922 (simplify_immed_subreg): Likewise.
923 * stmt.c (expand_case): Likewise.
924 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
925 signop rather than a bool.
926 * stor-layout.c (layout_type): Use wide-int interfaces.
927 (initialize_sizetypes): Update calls to
928 set_min_and_max_values_for_integral_type.
929 (set_min_and_max_values_for_integral_type): Take a signop rather
930 than a bool. Use wide-int interfaces.
931 (fixup_signed_type): Update accordingly. Remove
932 HOST_BITS_PER_DOUBLE_INT limit.
933 (fixup_unsigned_type): Likewise.
934 * system.h (STATIC_CONSTANT_P): New.
935 (STATIC_ASSERT): New.
936 * target.def (can_use_doloop_p): Take widest_ints rather than
938 * target.h: Include wide-int.h rather than double-int.h.
939 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
941 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
942 rather than INT_CST_LT_UNSIGNED.
943 (can_use_doloop_if_innermost): Take widest_ints rather than
945 * tree-affine.c: Include wide-int-print.h.
946 (double_int_ext_for_comb): Delete.
947 (wide_int_ext_for_comb): New.
948 (aff_combination_zero): Use wide-int interfaces.
949 (aff_combination_const): Take a widest_int instead of a double_int.
950 (aff_combination_elt): Use wide-int interfaces.
951 (aff_combination_scale): Take a widest_int instead of a double_int.
952 (aff_combination_add_elt): Likewise.
953 (aff_combination_add_cst): Likewise.
954 (aff_combination_add): Use wide-int interfaces.
955 (aff_combination_convert): Likewise.
956 (tree_to_aff_combination): Likewise.
957 (add_elt_to_tree): Take a widest_int instead of a double_int.
958 (aff_combination_to_tree): Use wide-int interfaces.
959 (aff_combination_remove_elt): Likewise.
960 (aff_combination_add_product): Take a widest_int instead of
962 (aff_combination_mult): Use wide-int interfaces.
963 (aff_combination_expand): Likewise.
964 (double_int_constant_multiple_p): Delete.
965 (wide_int_constant_multiple_p): New.
966 (aff_combination_constant_multiple_p): Take a widest_int pointer
967 instead of a double_int pointer.
968 (print_aff): Use wide-int interfaces.
969 (get_inner_reference_aff): Take a widest_int pointer
970 instead of a double_int pointer.
971 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
972 * tree-affine.h: Include wide-int.h.
973 (struct aff_comb_elt): Change type of coef to widest_int.
974 (struct affine_tree_combination): Change type of offset to widest_int.
975 (double_int_ext_for_comb): Delete.
976 (wide_int_ext_for_comb): New.
977 (aff_combination_const): Use widest_int instead of double_int.
978 (aff_combination_scale): Likewise.
979 (aff_combination_add_elt): Likewise.
980 (aff_combination_constant_multiple_p): Likewise.
981 (get_inner_reference_aff): Likewise.
982 (aff_comb_cannot_overlap_p): Likewise.
983 (aff_combination_zero_p): Use wide-int interfaces.
984 * tree.c: Include tree.h.
985 (init_ttree): Use make_int_cst.
986 (tree_code_size): Removed code for INTEGER_CST case.
987 (tree_size): Add INTEGER_CST case.
988 (make_node_stat): Update comment.
989 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
990 (build_int_cst_type): Use wide-int interfaces.
991 (double_int_to_tree): Likewise.
992 (double_int_fits_to_tree_p): Delete.
993 (force_fit_type_double): Delete.
994 (force_fit_type): New.
995 (int_cst_hash_hash): Use wide-int interfaces.
996 (int_cst_hash_eq): Likewise.
997 (build_int_cst_wide): Delete.
998 (wide_int_to_tree): New.
999 (cache_integer_cst): Use wide-int interfaces.
1000 (build_low_bits_mask): Likewise.
1001 (cst_and_fits_in_hwi): Likewise.
1002 (real_value_from_int_cst): Likewise.
1003 (make_int_cst_stat): New.
1004 (integer_zerop): Use wide_int interfaces.
1005 (integer_onep): Likewise.
1006 (integer_all_onesp): Likewise.
1007 (integer_pow2p): Likewise.
1008 (integer_nonzerop): Likewise.
1009 (tree_log2): Likewise.
1010 (tree_floor_log2): Likewise.
1011 (tree_ctz): Likewise.
1012 (int_size_in_bytes): Likewise.
1013 (mem_ref_offset): Return an offset_int rather than a double_int.
1014 (build_type_attribute_qual_variant): Use wide_int interfaces.
1015 (type_hash_eq): Likewise
1016 (tree_int_cst_equal): Likewise.
1017 (tree_int_cst_lt): Delete.
1018 (tree_int_cst_compare): Likewise.
1019 (tree_fits_shwi_p): Use wide_int interfaces.
1020 (tree_fits_uhwi_p): Likewise.
1021 (tree_int_cst_sign_bit): Likewise.
1022 (tree_int_cst_sgn): Likewise.
1023 (tree_int_cst_min_precision): Take a signop rather than a bool.
1024 (simple_cst_equal): Use wide_int interfaces.
1025 (compare_tree_int): Likewise.
1026 (iterative_hash_expr): Likewise.
1027 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1029 (get_type_static_bounds): Use wide_int interfaces.
1030 (tree_int_cst_elt_check_failed): New.
1031 (build_common_tree_nodes): Reordered to set prec before filling in
1033 (int_cst_value): Check cst_and_fits_in_hwi.
1034 (widest_int_cst_value): Use wide_int interfaces.
1035 (upper_bound_in_type): Likewise.
1036 (lower_bound_in_type): Likewise.
1037 (num_ending_zeros): Likewise.
1038 (drop_tree_overflow): Likewise.
1039 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1040 (gen_conditions_for_pow_cst_base): Likewise.
1041 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1042 (group_case_labels_stmt): Use wide-int interfaces.
1043 (verify_gimple_assign_binary): Likewise.
1044 (print_loop): Likewise.
1045 * tree-chrec.c (tree_fold_binomial): Likewise.
1046 * tree-core.h (struct tree_base): Add int_length.
1047 (struct tree_int_cst): Change rep of value.
1048 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1049 (dr_may_alias_p): Likewise.
1050 (max_stmt_executions_tree): Likewise.
1051 * tree.def (INTEGER_CST): Update comment.
1052 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1053 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1054 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1055 (dequeue_and_dump): Use wide-int interfaces.
1056 * tree.h: Include wide-int.h.
1057 (NULL_TREE): Moved to earlier loc in file.
1058 (TREE_INT_CST_ELT_CHECK): New.
1059 (tree_int_cst_elt_check_failed): New.
1061 (TREE_INT_CST): Delete.
1062 (TREE_INT_CST_LOW): Use wide-int interfaces.
1063 (TREE_INT_CST_HIGH): Delete.
1064 (TREE_INT_CST_NUNITS): New.
1065 (TREE_INT_CST_EXT_NUNITS): Likewise.
1066 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1067 (TREE_INT_CST_ELT): Likewise.
1068 (INT_CST_LT): Delete.
1069 (tree_int_cst_elt_check): New (two forms).
1070 (type_code_size): Update comment.
1071 (make_int_cst_stat, make_int_cst): New.
1072 (tree_to_double_int): Delete.
1073 (double_int_fits_to_tree_p): Delete.
1074 (force_fit_type_double): Delete.
1075 (build_int_cstu): Replace with out-of-line function.
1076 (build_int_cst_wide): Delete.
1077 (tree_int_cst_lt): Define inline.
1078 (tree_int_cst_le): New.
1079 (tree_int_cst_compare): Define inline.
1080 (tree_int_cst_min_precision): Take a signop rather than a bool.
1081 (wi::int_traits <const_tree>): New.
1082 (wi::int_traits <tree>): New.
1083 (wi::extended_tree): New.
1084 (wi::int_traits <wi::extended_tree>): New.
1085 (wi::to_widest): New.
1086 (wi::to_offset): New.
1087 (wi::fits_to_tree_p): New.
1088 (wi::min_value): New.
1089 (wi::max_value): New.
1090 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1091 (copy_tree_body_r): Likewise.
1092 * tree-object-size.c (compute_object_offset): Likewise.
1093 (addr_object_size): Likewise.
1094 * tree-predcom.c: Include wide-int-print.h.
1095 (struct dref_d): Change type of offset to widest_int.
1096 (dump_dref): Call wide-int printer.
1097 (aff_combination_dr_offset): Use wide-int interfaces.
1098 (determine_offset): Take a widest_int pointer rather than a
1100 (split_data_refs_to_components): Use wide-int interfaces.
1101 (suitable_component_p): Likewise.
1102 (order_drefs): Likewise.
1103 (add_ref_to_chain): Likewise.
1104 (valid_initializer_p): Likewise.
1105 (determine_roots_comp): Likewise.
1106 * tree-pretty-print.c: Include wide-int-print.h.
1107 (dump_generic_node): Use wide-int interfaces.
1108 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1109 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1110 (move_fixed_address_to_symbol): Likewise.
1111 (move_hint_to_base): Likewise.
1112 (move_pointer_to_base): Likewise.
1113 (move_variant_to_index): Likewise.
1114 (most_expensive_mult_to_index): Likewise.
1115 (addr_to_parts): Likewise.
1116 (copy_ref_info): Likewise.
1117 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1118 (indirect_refs_may_alias_p): Likewise.
1119 (stmt_kills_ref_p_1): Likewise.
1120 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1121 * tree-ssa-ccp.c: Update comment at top of file. Include
1123 (struct prop_value_d): Change type of mask to widest_int.
1124 (extend_mask): New function.
1125 (dump_lattice_value): Use wide-int interfaces.
1126 (get_default_value): Likewise.
1127 (set_constant_value): Likewise.
1128 (set_value_varying): Likewise.
1129 (valid_lattice_transition): Likewise.
1130 (set_lattice_value): Likewise.
1131 (value_to_double_int): Delete.
1132 (value_to_wide_int): New.
1133 (get_value_from_alignment): Use wide-int interfaces.
1134 (get_value_for_expr): Likewise.
1135 (do_dbg_cnt): Likewise.
1136 (ccp_finalize): Likewise.
1137 (ccp_lattice_meet): Likewise.
1138 (bit_value_unop_1): Use widest_ints rather than double_ints.
1139 (bit_value_binop_1): Likewise.
1140 (bit_value_unop): Use wide-int interfaces.
1141 (bit_value_binop): Likewise.
1142 (bit_value_assume_aligned): Likewise.
1143 (evaluate_stmt): Likewise.
1144 (ccp_fold_stmt): Likewise.
1145 (visit_cond_stmt): Likewise.
1146 (ccp_visit_stmt): Likewise.
1147 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1148 (constant_pointer_difference): Likewise.
1149 (associate_pointerplus): Likewise.
1150 (combine_conversions): Likewise.
1151 * tree-ssa-loop.h: Include wide-int.h.
1152 (struct tree_niter_desc): Change type of max to widest_int.
1153 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1154 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1155 (remove_redundant_iv_tests): Likewise.
1156 (canonicalize_loop_induction_variables): Likewise.
1157 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1158 (constant_multiple_of): Take a widest_int pointer instead of
1159 a double_int pointer.
1160 (get_computation_aff): Use wide-int interfaces.
1161 (ptr_difference_cost): Likewise.
1162 (difference_cost): Likewise.
1163 (get_loop_invariant_expr_id): Likewise.
1164 (get_computation_cost_at): Likewise.
1165 (iv_elimination_compare_lt): Likewise.
1166 (may_eliminate_iv): Likewise.
1167 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1168 instead of double_int.
1169 (max_loop_iterations): Likewise.
1170 (max_stmt_executions): Likewise.
1171 (estimated_stmt_executions): Likewise.
1172 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1173 (split_to_var_and_offset): Use wide-int interfaces.
1174 (determine_value_range): Likewise.
1175 (bound_difference_of_offsetted_base): Likewise.
1176 (bounds_add): Take a widest_int instead of a double_int.
1177 (number_of_iterations_ne_max): Use wide-int interfaces.
1178 (number_of_iterations_ne): Likewise.
1179 (number_of_iterations_lt_to_ne): Likewise.
1180 (assert_loop_rolls_lt): Likewise.
1181 (number_of_iterations_lt): Likewise.
1182 (number_of_iterations_le): Likewise.
1183 (number_of_iterations_cond): Likewise.
1184 (number_of_iterations_exit): Likewise.
1185 (finite_loop_p): Likewise.
1186 (derive_constant_upper_bound_assign): Likewise.
1187 (derive_constant_upper_bound): Return a widest_int.
1188 (derive_constant_upper_bound_ops): Likewise.
1189 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1190 (record_estimate): Take a widest_int rather than a double_int.
1191 (record_nonwrapping_iv): Use wide-int interfaces.
1192 (double_int_cmp): Delete.
1193 (wide_int_cmp): New.
1194 (bound_index): Take a widest_int rather than a double_int.
1195 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1196 (maybe_lower_iteration_bound): Likewise.
1197 (estimate_numbers_of_iterations_loop): Likewise.
1198 (estimated_loop_iterations): Take a widest_int pointer than than
1199 a double_int pointer.
1200 (estimated_loop_iterations_int): Use wide-int interfaces.
1201 (max_loop_iterations): Take a widest_int pointer than than
1202 a double_int pointer.
1203 (max_loop_iterations_int): Use wide-int interfaces.
1204 (max_stmt_executions): Take a widest_int pointer than than
1205 a double_int pointer.
1206 (estimated_stmt_executions): Likewise.
1207 (n_of_executions_at_most): Use wide-int interfaces.
1208 (scev_probably_wraps_p): Likewise.
1209 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1211 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1213 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1214 double_ints. Adjust for trailing_wide_ints <3> representation.
1215 (set_nonzero_bits): Likewise.
1216 (get_range_info): Return wide_ints rather than double_ints.
1217 Adjust for trailing_wide_ints <3> representation.
1218 (get_nonzero_bits): Likewise.
1219 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1221 * tree-ssanames.h (struct range_info_def): Replace min, max and
1222 nonzero_bits with a trailing_wide_ints <3>.
1223 (set_range_info): Use wide_int_refs rather than double_ints.
1224 (set_nonzero_bits): Likewise.
1225 (get_range_info): Return wide_ints rather than double_ints.
1226 (get_nonzero_bits): Likewise.
1227 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1228 * tree-ssa-pre.c (phi_translate_1): Likewise.
1229 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1230 (acceptable_pow_call): Likewise.
1231 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1233 (vn_reference_fold_indirect): Likewise.
1234 (vn_reference_maybe_forwprop_address): Likewise.
1235 (valueize_refs_1): Likewise.
1236 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1237 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1238 tree_int_cst_lt and tree_int_cst_le.
1239 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1241 (streamer_alloc_tree): Likewise.
1242 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1243 (streamer_write_tree_header): Likewise.
1244 (streamer_write_integer_cst): Likewise.
1245 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1246 (build_constructors): Likewise.
1247 (array_value_type): Likewise.
1248 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1249 (vect_check_gather): Likewise.
1250 * tree-vect-generic.c (build_replicated_const): Likewise.
1251 (expand_vector_divmod): Likewise.
1252 * tree-vect-loop.c (vect_transform_loop): Likewise.
1253 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1254 (vect_do_peeling_for_alignment): Likewise.
1255 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1256 * tree-vrp.c: Include wide-int.h.
1257 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1258 (extract_range_from_assert): Use wide-int interfaces.
1259 (vrp_int_const_binop): Likewise.
1260 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1261 double_int pointers.
1262 (ranges_from_anti_range): Use wide-int interfaces.
1263 (quad_int_cmp): Delete.
1264 (quad_int_pair_sort): Likewise.
1265 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1266 (extract_range_from_unary_expr_1): Likewise.
1267 (adjust_range_with_scev): Likewise.
1268 (masked_increment): Take and return wide_ints rather than double_ints.
1269 (register_edge_assert_for_2): Use wide-int interfaces.
1270 (check_array_ref): Likewise.
1271 (search_for_addr_array): Likewise.
1272 (maybe_set_nonzero_bits): Likewise.
1273 (union_ranges): Pass an integer of the correct type instead of
1274 using integer_one_node.
1275 (intersect_ranges): Likewise.
1276 (simplify_truth_ops_using_ranges): Likewise.
1277 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1278 (range_fits_type_p): Likewise.
1279 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1281 (simplify_conversion_using_ranges): Use wide-int interfaces.
1282 (simplify_float_conversion_using_ranges): Likewise.
1283 (vrp_finalize): Likewise.
1284 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1285 (gimple_stringops_transform): Likewise.
1286 * varasm.c (decode_addr_const): Likewise.
1287 (const_hash_1): Likewise.
1288 (const_rtx_hash_1): Likewise
1289 (output_constant): Likewise.
1290 (array_size_for_constructor): Likewise.
1291 (output_constructor_regular_field): Likewise.
1292 (output_constructor_bitfield): Likewise.
1293 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1294 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1296 * gencheck.c: Define BITS_PER_UNIT.
1299 * wide-int-print.cc: New.
1300 * wide-int-print.h: New.
1302 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1304 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1306 2014-05-06 Richard Biener <rguenther@suse.de>
1308 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1309 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1310 (TODO_verify_all): Adjust.
1311 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1312 TODO_verify_stmts and TODO_verify_rtl_sharing.
1313 * bb-reorder.c: Likewise.
1314 * cfgexpand.c: Likewise.
1315 * cprop.c: Likewise.
1317 * function.c: Likewise.
1318 * fwprop.c: Likewise.
1320 * gimple-ssa-isolate-paths.c: Likewise.
1321 * gimple-ssa-strength-reduction.c: Likewise.
1322 * ipa-split.c: Likewise.
1323 * loop-init.c: Likewise.
1324 * loop-unroll.c: Likewise.
1325 * lower-subreg.c: Likewise.
1326 * modulo-sched.c: Likewise.
1327 * postreload-gcse.c: Likewise.
1328 * predict.c: Likewise.
1329 * recog.c: Likewise.
1330 * sched-rgn.c: Likewise.
1331 * store-motion.c: Likewise.
1332 * tracer.c: Likewise.
1333 * trans-mem.c: Likewise.
1334 * tree-call-cdce.c: Likewise.
1335 * tree-cfg.c: Likewise.
1336 * tree-cfgcleanup.c: Likewise.
1337 * tree-complex.c: Likewise.
1338 * tree-eh.c: Likewise.
1339 * tree-emutls.c: Likewise.
1340 * tree-if-conv.c: Likewise.
1341 * tree-into-ssa.c: Likewise.
1342 * tree-loop-distribution.c: Likewise.
1343 * tree-object-size.c: Likewise.
1344 * tree-parloops.c: Likewise.
1345 * tree-pass.h: Likewise.
1346 * tree-sra.c: Likewise.
1347 * tree-ssa-ccp.c: Likewise.
1348 * tree-ssa-copy.c: Likewise.
1349 * tree-ssa-copyrename.c: Likewise.
1350 * tree-ssa-dce.c: Likewise.
1351 * tree-ssa-dom.c: Likewise.
1352 * tree-ssa-dse.c: Likewise.
1353 * tree-ssa-forwprop.c: Likewise.
1354 * tree-ssa-ifcombine.c: Likewise.
1355 * tree-ssa-loop-ch.c: Likewise.
1356 * tree-ssa-loop-ivcanon.c: Likewise.
1357 * tree-ssa-loop.c: Likewise.
1358 * tree-ssa-math-opts.c: Likewise.
1359 * tree-ssa-phiopt.c: Likewise.
1360 * tree-ssa-phiprop.c: Likewise.
1361 * tree-ssa-pre.c: Likewise.
1362 * tree-ssa-reassoc.c: Likewise.
1363 * tree-ssa-sink.c: Likewise.
1364 * tree-ssa-strlen.c: Likewise.
1365 * tree-ssa-tail-merge.c: Likewise.
1366 * tree-ssa-uncprop.c: Likewise.
1367 * tree-switch-conversion.c: Likewise.
1368 * tree-tailcall.c: Likewise.
1369 * tree-vect-generic.c: Likewise.
1370 * tree-vectorizer.c: Likewise.
1371 * tree-vrp.c: Likewise.
1373 * var-tracking.c: Likewise.
1374 * bt-load.c: Likewise.
1375 * cfgcleanup.c: Likewise.
1376 * combine-stack-adj.c: Likewise.
1377 * combine.c: Likewise.
1378 * compare-elim.c: Likewise.
1379 * config/epiphany/resolve-sw-modes.c: Likewise.
1380 * config/i386/i386.c: Likewise.
1381 * config/mips/mips.c: Likewise.
1382 * config/s390/s390.c: Likewise.
1383 * config/sh/sh_treg_combine.cc: Likewise.
1384 * config/sparc/sparc.c: Likewise.
1387 * final.c: Likewise.
1388 * ifcvt.c: Likewise.
1389 * mode-switching.c: Likewise.
1390 * passes.c: Likewise.
1391 * postreload.c: Likewise.
1393 * reg-stack.c: Likewise.
1394 * regcprop.c: Likewise.
1395 * regrename.c: Likewise.
1398 2014-05-06 Richard Biener <rguenther@suse.de>
1401 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1402 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1404 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1407 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1409 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1410 Tom de Vries <tom@codesourcery.com>
1412 * target.def (call_fusage_contains_non_callee_clobbers): New
1414 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1416 (@node Miscellaneous Register Hooks): New node.
1417 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1418 * doc/tm.texi: Regenerate.
1420 2014-05-05 Marek Polacek <polacek@redhat.com>
1423 * opts.c (common_handle_option): Call error_at instead of warning_at.
1425 2014-05-05 Richard Biener <rguenther@suse.de>
1427 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1428 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1429 under the TODO_verify_il umbrella.
1431 2014-05-05 Richard Biener <rguenther@suse.de>
1433 * passes.c (execute_function_todo): Move TODO_verify_flow under
1434 the TODO_verify_ul umbrella.
1436 2014-05-05 Richard Biener <rguenther@suse.de>
1439 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1440 X & CST away from a CST that is the mask of a mode.
1442 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1444 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1445 int argument to enum machine_mode.
1446 (picochip_class_max_nregs): Ditto.
1447 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1448 (picochip_class_max_nregs): Ditto.
1450 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1452 * target.def: Add new target hook.
1453 * doc/tm.texi: Regenerate.
1454 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1455 * targhooks.c (default_keep_leaf_when_profiled): New function.
1457 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1458 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1460 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1462 PR tree-optimization/60363
1463 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1464 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1465 (update_destination_phis): New parameter.
1466 (create_edge_and_update_destination_phis): Ditto.
1467 (ssa_fix_duplicate_block_edges): Pass new arguments.
1468 (thread_single_edge): Ditto.
1470 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1472 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1473 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1474 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1475 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1476 Use RS6000_BTM_HARD_FLOAT.
1477 (BU_MISC_2): Likewise.
1478 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1479 RS6000_BTM_HARD_FLOAT.
1480 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1482 (rs6000_invalid_builtin): Add hard floating builtin support.
1483 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1484 hard float builtins.
1485 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1487 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1489 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1490 Add missing function* argument.
1492 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1494 * lra-constraints.c (valid_address_p): Move earlier in file.
1495 Add a constraint argument to the address_info version.
1496 (satisfies_memory_constraint_p): New function.
1497 (satisfies_address_constraint_p): Likewise.
1498 (process_alt_operands, curr_insn_transform): Use them.
1499 (process_address): Pass the constraint to valid_address_p when
1500 checking address operands.
1502 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1504 * config/mips/mips.c (mips_isa_rev): New variable.
1505 (mips_set_architecture): Set it.
1506 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1508 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1509 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1510 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1511 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1512 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1513 conditions in terms of mips_isa_rev.
1514 (mips_isa_rev): Declare.
1516 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1518 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1519 (prob_unlikely, prob_likely): Make variables const.
1521 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1523 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1525 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1527 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1529 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1531 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1532 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1533 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1535 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1536 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1538 Replace usage of ROUND_REG with sh_round_reg.
1539 Use CEIL instead of ROUND_ADVANCE.
1541 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1544 * config/sh/sh.c: Include stdlib headers before everything else.
1546 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1548 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1549 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1550 (gimplify_adjust_omp_clauses): Simd region is never
1551 directly nested in combined parallel. Instead, for linear
1552 with copyin/copyout, if in combined for simd loop, make decl
1553 firstprivate/lastprivate on OMP_FOR.
1554 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1555 expand_omp_for_static_chunk): When setting endvar, also set
1556 fd->loop.v to the same value.
1558 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1560 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1562 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1564 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1567 2014-05-02 Marek Polacek <polacek@redhat.com>
1569 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1571 2014-05-02 Kito Cheng <kito@0xlab.org>
1573 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1574 to a C expression marco.
1575 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1576 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1577 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1578 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1579 HONOR_REG_ALLOC_ORDER.
1580 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1582 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1584 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1586 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1588 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1590 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1592 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1593 (convert_scalar_cond_reduction): Likewise.
1594 (predicate_scalar_phi): Add recognition and transformation
1595 of simple conditioanl reduction to be vectorizable.
1597 2014-05-01 Marek Polacek <polacek@redhat.com>
1600 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1602 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1604 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1605 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1606 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1607 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1608 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1609 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1610 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1611 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1613 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1615 * config/arc/arc.opt (mlra): Move comment above option name
1616 to avoid mis-parsing as language options.
1618 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1620 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1621 * config/sol2.h: ... here.
1622 * config/sol2-10.h: Remove.
1624 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1625 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1626 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1627 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1628 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1629 * config/sol2.h: ... here.
1630 (SECTION_NAME_FORMAT): Don't redefine.
1631 (STARTFILE_ARCH32_SPEC): Rename to ...
1632 (STARTFILE_ARCH_SPEC): ... this.
1633 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1634 * config/sparc/sol2.h: ... here.
1635 (SECTION_NAME_FORMAT): Don't undef.
1636 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1637 (SUBTARGET_EXTRA_SPECS): Remove.
1638 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1640 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1641 (MD_STARTFILE_PREFIX): Remove.
1642 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1643 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1644 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1645 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1646 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1647 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1648 * config/i386/sol2.h: ... here.
1649 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1650 * config/i386/sol2-bi.h: Remove.
1651 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1652 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1654 * config/i386/t-sol2-64: Rename to ...
1655 * config/i386/t-sol2: ... this.
1656 * config/sparc/t-sol2-64: Rename to ...
1657 * config/sparc/t-sol2: ... this.
1659 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1660 sol2_tm_file_head, sol2_tm_file_tail.
1661 Include ${cpu_type}/sol2.h before sol2.h.
1663 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1664 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1665 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1666 Reflect i386/t-sol2-64 renaming.
1667 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1668 Reflect sparc/t-sol2-64 renaming.
1670 2014-04-30 Richard Biener <rguenther@suse.de>
1672 * passes.c (execute_function_todo): Move TODO_verify_stmts
1673 and TODO_verify_ssa under the TODO_verify_il umbrella.
1674 * tree-ssa.h (verify_ssa): Adjust prototype.
1675 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1676 we should verify SSA operands.
1677 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1678 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1679 whether we should verify whether not throwing stmts have EH info.
1680 * graphite-scop-detection.c (create_sese_edges): Adjust.
1681 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1682 * tree-eh.c (lower_try_finally_switch): Do not add the
1683 default case label twice.
1685 2014-04-30 Marek Polacek <polacek@redhat.com>
1687 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1688 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1689 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1690 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1692 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1694 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1695 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1696 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1697 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1698 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1699 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1700 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1701 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1703 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1705 * tree-cfg.c (dump_function_to_file): Dump the return type of
1706 functions, in a line to itself before the function body, mimicking
1707 the layout of a C function.
1709 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1711 PR tree-optimization/60971
1712 * tree-tailcall.c (process_assignment): Reject conversions which
1715 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1717 * calls.c (initialize_argument_information): Always treat
1718 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1719 (expand_call): Likewise.
1720 (emit_library_call_calue_1): Likewise.
1721 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1722 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1725 2014-04-29 Nick Clifton <nickc@redhat.com>
1727 * config/msp430/msp430.md (umulsidi): Fix typo.
1728 (mulhisi3): Enable even inside interrupt handlers.
1729 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1730 bigger return address pushed in large mode.
1732 2014-04-29 Nick Clifton <nickc@redhat.com>
1734 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1735 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1737 * config/m32r/m32r.c (init_reg_tables): Likewise.
1738 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1739 enum to hold the modes.
1741 2014-04-29 Richard Biener <rguenther@suse.de>
1743 * dominance.c (free_dominance_info): Add overload with
1745 (dom_info_state): Likewise.
1746 (dom_info_available_p): Likewise.
1747 * basic-block.h (free_dominance_info, dom_info_state,
1748 dom_info_available_p): Declare overloads.
1749 * passes.c (execute_function_todo): Verify that verifiers
1750 don't change dominator info state. Drop dominator info
1751 for IPA pass invocations.
1752 * cgraph.c (release_function_body): Restore asserts that
1753 dominator information is released.
1755 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1757 * doc/invoke.texi: Fix typo.
1758 * tree-vrp.c: Fix typos.
1759 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1761 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1763 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1765 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1767 * config/aarch64/aarch64-builtins.c
1768 (aarch64_types_storestruct_lane_qualifiers): New.
1769 (TYPES_STORESTRUCT_LANE): Likewise.
1770 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1771 (st3_lane): Likewise.
1772 (st4_lane): Likewise.
1773 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1774 (vec_store_lanesci_lane<mode>): Likewise.
1775 (vec_store_lanesxi_lane<mode>): Likewise.
1776 (aarch64_st2_lane<VQ:mode>): Likewise.
1777 (aarch64_st3_lane<VQ:mode>): Likewise.
1778 (aarch64_st4_lane<VQ:mode>): Likewise.
1779 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1780 * config/aarch64/arm_neon.h
1781 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1782 use new macro arguments.
1783 (__ST3_LANE_FUNC): Likewise.
1784 (__ST4_LANE_FUNC): Likewise.
1785 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1786 (V_THREE_ELEM): Likewise.
1787 (V_FOUR_ELEM): Likewise.
1789 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1791 * doc/gimple.texi: Replace the description of the now-defunct
1792 union gimple_statement_d with a diagram showing the
1793 gimple_statement_base class hierarchy and its relationships to
1794 the GSS_ and GIMPLE_ enums.
1796 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1798 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1799 * config/aarch64/aarch64.c
1800 (aarch64_cannot_change_mode_class): Weaken conditions.
1801 (aarch64_modes_tieable_p): New.
1802 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1804 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1806 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1807 (loadsync_<mode>): Change mode.
1808 (load_quadpti, store_quadpti): New.
1809 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1810 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1812 2014-04-28 Martin Jambor <mjambor@suse.cz>
1814 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1815 same alias type as the original statement.
1816 (subreplacement_assignment_data): New type.
1817 (handle_unscalarized_data_in_subtree): New type of parameter,
1818 generate new memory accesses with same alias type as the original
1820 (load_assign_lhs_subreplacements): Likewise.
1821 (sra_modify_constructor_assign): Generate new memory accesses with
1822 same alias type as the original statement.
1824 2014-04-28 Richard Biener <rguenther@suse.de>
1826 * tree-pass.h (TODO_verify_il): Define.
1827 (TODO_verify_all): Complete properly.
1828 * passes.c (execute_function_todo): Move existing loop-closed
1829 SSA verification under TODO_verify_il.
1830 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1831 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1832 Fix tree sharing issue.
1834 2014-04-28 Richard Biener <rguenther@suse.de>
1837 * builtins.def (DEF_C11_BUILTIN): Add.
1838 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1839 * coretypes.h (enum function_class): Add function_c11_misc.
1840 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1841 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1842 (call_may_clobber_ref_p_1): Likewise.
1843 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1844 (mark_all_reaching_defs_necessary_1): Likewise.
1845 (propagate_necessity): Likewise.
1846 (eliminate_unnecessary_stmts): Likewise.
1847 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1849 2014-04-28 Richard Biener <rguenther@suse.de>
1851 * tree-vrp.c (vrp_var_may_overflow): Remove.
1852 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1853 with overflow immediately bump to one before that value and
1854 let iteration figure out overflow status.
1856 2014-04-28 Richard Biener <rguenther@suse.de>
1858 * configure.ac: Do valgrind header checks unconditionally.
1859 Add --enable-valgrind-annotations.
1860 * system.h: Guard valgrind header inclusion with
1861 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1862 * alloc-pool.c (pool_alloc, pool_free): Use
1863 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1864 to guard possibly dead code.
1865 * config.in: Regenerated.
1866 * configure: Likewise.
1868 2014-04-28 Jeff Law <law@redhat.com>
1870 PR tree-optimization/60902
1871 * tree-ssa-threadedge.c
1872 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1873 over real defs when invalidating outputs from statements that do not
1874 produce useful outputs for threading.
1876 2014-04-28 Richard Biener <rguenther@suse.de>
1878 PR tree-optimization/60979
1879 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1880 SCOPs that end in a block with a successor with abnormal
1883 2014-04-28 Richard Biener <rguenther@suse.de>
1885 * tree-pass.h (execute_pass_list): Adjust prototype.
1886 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1887 (do_per_function): Change callback signature, push all actual
1888 work to the callbals.
1889 (do_per_function_toporder): Likewise.
1890 (execute_function_dump): Adjust.
1891 (execute_function_todo): Likewise.
1892 (clear_last_verified): Likewise.
1893 (verify_curr_properties): Likewise.
1894 (update_properties_after_pass): Likewise.
1895 (execute_pass_list_1): Split out from ...
1896 (execute_pass_list): ... here. Adjust.
1897 (execute_ipa_pass_list): Likewise.
1898 * cgraphunit.c (cgraph_add_new_function): Adjust.
1899 (analyze_function): Likewise.
1900 (expand_function): Likewise.
1901 * cgraph.c (release_function_body): Free dominance info
1902 here instead of asserting it was magically freed elsewhere.
1904 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1906 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1907 * configure: Regenerate.
1908 * config/sparc/sparc.opt (muser-mode): New option.
1909 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1911 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1912 * doc/invoke.texi (SPARC options): Document -muser-mode.
1914 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1916 * cselib.c (find_slot_memmode): Delete.
1917 (cselib_hasher): Change compare_type to a struct.
1918 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1920 (preserve_constants_and_equivs): Adjust for new compare_type.
1921 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1922 (wrap_constant): Delete.
1923 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1925 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1927 * doc/install.texi (Building with profile feedback): Remove
1930 2014-04-26 Tom de Vries <tom@codesourcery.com>
1932 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1935 2014-04-25 Cary Coutant <ccoutant@google.com>
1938 * dwarf2out.c (should_move_die_to_comdat): A type definition
1939 can contain a subprogram definition, but don't move it to a
1941 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1942 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1944 (clone_tree_hash): Rename to...
1945 (clone_tree_partial): ...this; change callers. Copy
1946 DW_TAG_subprogram DIEs as declarations.
1947 (copy_decls_walk): Don't copy children of a declaration into a
1950 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1953 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1956 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1958 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1959 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1961 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1964 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1966 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1968 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1970 PR tree-optimization/60930
1971 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1972 creating a multiply candidate by folding two constant
1973 multiplicands when the result overflows.
1975 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1977 PR tree-optimization/60960
1978 * tree-vect-generic.c (expand_vector_operation): Only call
1979 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1981 2014-04-25 Tom de Vries <tom@codesourcery.com>
1983 * expr.c (clobber_reg_mode): New function.
1984 * expr.h (clobber_reg): New function.
1986 2014-04-25 Tom de Vries <tom@codesourcery.com>
1988 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1991 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1992 Tom de Vries <tom@codesourcery.com>
1994 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1996 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1997 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1998 new argument to find_all_hard_reg_sets call.
2000 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2002 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2003 Use HOST_WIDE_INT_C for mask literal.
2004 (aarch_rev16_shleft_mask_imm_p): Likewise.
2006 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2009 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2011 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2013 PR preprocessor/56540
2014 * config/i386/i386-c.c (ix86_target_macros): Define
2015 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2017 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2019 * configure.ac (tga_func): Remove.
2020 (LIB_TLS_SPEC): Remove.
2021 * configure: Regenerate.
2022 * config.in: Regenerate.
2023 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2025 2014-04-25 Richard Biener <rguenther@suse.de>
2028 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2029 call stmt use/clobber sets during stmt walk instead of
2030 walking the possibly incomplete set of caller edges.
2032 2014-04-25 Richard Biener <rguenther@suse.de>
2035 * passes.c (apply_ipa_transforms): Inline into only caller ...
2036 (execute_one_pass): ... here. Properly bring in function
2037 bodies for nodes we want to apply IPA transforms to.
2039 2014-04-24 Cong Hou <congh@google.com>
2041 PR tree-optimization/60896
2042 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2043 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2044 (vect_mark_pattern_stmts): Set the def type of all statements in
2045 PATTERN_DEF_SEQ as vect_internal_def.
2047 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2049 * doc/extend.texi (PowerPC Built-in Functions): Document new
2050 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2051 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2053 * config/rs6000/predicates.md (const_0_to_3_operand): New
2054 predicate to match 0..3 integer constants.
2056 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2057 to support adding miscellaneous builtin functions.
2058 (BU_DFP_MISC_2): Likewise.
2059 (BU_P7_MISC_1): Likewise.
2060 (BU_P7_MISC_2): Likewise.
2061 (BU_P8V_MISC_3): Likewise.
2062 (BU_MISC_1): Likewise.
2063 (BU_MISC_2): Likewise.
2064 (DIVWE): Add extended divide builtin functions.
2067 (DIVWEUO): Likewise.
2071 (DIVDEUO): Likewise.
2072 (DXEX): Add decimal floating-point builtin functions.
2075 (DDEDPDQ): Likewise.
2077 (DENBCDQ): Likewise.
2084 (CDTBCD): Add new BCD builtin functions.
2088 (BCDADD_LT): Likewise.
2089 (BCDADD_EQ): Likewise.
2090 (BCDADD_GT): Likewise.
2091 (BCDADD_OV): Likewise.
2093 (BCDSUB_LT): Likewise.
2094 (BCDSUB_EQ): Likewise.
2095 (BCDSUB_GT): Likewise.
2096 (BCDSUB_OV): Likewise.
2097 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2098 (UNPACK_TD): Likewise.
2099 (PACK_TF): Likewise.
2100 (UNPACK_TF): Likewise.
2101 (UNPACK_TF_0): Likewise.
2102 (UNPACK_TF_1): Likewise.
2103 (PACK_V1TI): Likewise.
2104 (UNPACK_V1TI): Likewise.
2106 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2107 support for decimal floating point builtin functions.
2108 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2109 functions that take constant arguments.
2110 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2111 (rs6000_init_builtins): Setup long double, _Decimal64, and
2112 _Decimal128 types for new builtin functions.
2113 (builtin_function_type): Set the unsigned flags appropriately for
2114 the new builtin functions.
2115 (rs6000_opt_masks): Add support for decimal floating point builtin
2118 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2119 floating point builtin functions.
2120 (RS6000_BTM_COMMON): Likewise.
2121 (RS6000_BTI_long_double): Likewise.
2122 (RS6000_BTI_dfloat64): Likewise.
2123 (RS6000_BTI_dfloat128): Likewise.
2124 (long_double_type_internal_node): Likewise.
2125 (dfloat64_type_internal_node): Likewise.
2126 (dfloat128_type_internal_node): Likewise.
2128 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2129 2.07 bcd arithmetic instructions.
2130 (UNSPEC_BCDSUB): Likewise.
2131 (UNSPEC_BCD_OVERFLOW): Likewise.
2132 (UNSPEC_BCD_ADD_SUB): Likewise.
2133 (bcd_add_sub): Likewise.
2134 (BCD_TEST): Likewise.
2135 (bcd<bcd_add_sub>): Likewise.
2136 (bcd<bcd_add_sub>_test): Likewise.
2137 (bcd<bcd_add_sub>_test2): Likewise.
2138 (bcd<bcd_add_sub>_<code>): Likewise.
2139 (peephole2 for combined bcd ops): Likewise.
2141 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2142 decimal floating point builtin functions.
2143 (UNSPEC_DENBCD): Likewise.
2144 (UNSPEC_DXEX): Likewise.
2145 (UNSPEC_DIEX): Likewise.
2146 (UNSPEC_DSCLI): Likewise.
2147 (UNSPEC_DSCRI): Likewise.
2148 (D64_D128): Likewise.
2149 (dfp_suffix): Likewise.
2150 (dfp_ddedpd_<mode>): Likewise.
2151 (dfp_denbcd_<mode>): Likewise.
2152 (dfp_dxex_<mode>): Likewise.
2153 (dfp_diex_<mode>): Likewise.
2154 (dfp_dscli_<mode>): Likewise.
2155 (dfp_dscri_<mode>): Likewise.
2157 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2159 (UNSPEC_CDTBCD): Likewise.
2160 (UNSPEC_CBCDTD): Likewise.
2161 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2162 (UNSPEC_DIVEO): Likewise.
2163 (UNSPEC_DIVEU): Likewise.
2164 (UNSPEC_DIVEUO): Likewise.
2165 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2166 pack/unpack 128-bit types.
2167 (UNSPEC_PACK_128BIT): Likewise.
2168 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2169 (udiv<mode>3): Use idiv_ldiv mode attribute.
2170 (div<mode>3): Likewise.
2171 (addg6s): Add new BCD builtin functions.
2174 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2175 (div_extend): Likewise.
2176 (div<div_extend>_<mode>"): Likewise.
2177 (FP128_64): Add support for new builtin functions to pack/unpack
2179 (unpack<mode>): Likewise.
2180 (unpacktf_0): Likewise.
2181 (unpacktf_1): Likewise.
2182 (unpack<mode>_dm): Likewise.
2183 (unpack<mode>_nodm): Likewise.
2184 (pack<mode>): Likewise.
2185 (unpackv1ti): Likewise.
2186 (packv1ti): Likewise.
2188 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2190 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2193 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2195 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2196 * gimplify.c (omp_is_private): Change last argument's type to int.
2197 Only diagnose lastprivate if the simd argument is 1, only diagnose
2198 linear if the simd argument is 2.
2199 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2200 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2201 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2202 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2203 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2204 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2205 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2206 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2207 * tree-nested.c (convert_nonlocal_omp_clauses,
2208 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2210 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2213 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2216 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2218 * flag-types.h (enum ivar_visibility): Add.
2220 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2222 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2223 function * argument.
2225 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2227 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2229 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2230 Tom de Vries <tom@codesourcery.com>
2232 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2233 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2235 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2236 * emit-rtl.c (try_split): Same.
2238 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2239 Tom de Vries <tom@codesourcery.com>
2241 * common.opt (fuse-caller-save): New option.
2243 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2245 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2246 elements for big-endian.
2248 2014-04-24 Richard Biener <rguenther@suse.de>
2250 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2251 during TER and instead use the sepops interface for expanding
2252 non-GIMPLE_SINGLE_RHS.
2254 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2256 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2257 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2259 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2261 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2262 assembler 64-bit option.
2263 * configure: Regenerate.
2265 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2267 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2268 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2269 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2270 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2271 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2273 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2275 * config/aarch64/aarch64-builtins.c
2276 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2277 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2278 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2279 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2281 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2282 (Vrevsuff): New mode attribute.
2284 2014-04-24 Terry Guo <terry.guo@arm.com>
2286 * config/arm/arm.h (machine_function): Define variable
2287 after_arm_reorg here.
2288 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2289 (arm_split_constant): Update the way to access variable
2292 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2294 2014-04-23 Tom de Vries <tom@codesourcery.com>
2296 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2298 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2300 * is-a.h: Update comments to reflect the following changes to the
2301 "pointerness" of the API, making the template parameter match the
2302 return type, allowing use of is-a.h with typedefs of pointers.
2303 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2304 that the return type matches the parameter to the is_a_helper.
2306 (dyn_cast): Likewise.
2308 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2309 pointer from the is-a.h API.
2311 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2312 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2314 (is_a_helper <varpool_node>::test): Likewise, convert to...
2315 (is_a_helper <varpool_node *>::test): ...this.
2317 (varpool_first_variable): Update for removal of implicit pointer
2318 from the is-a.h API.
2319 (varpool_next_variable): Likewise.
2320 (varpool_first_static_initializer): Likewise.
2321 (varpool_next_static_initializer): Likewise.
2322 (varpool_first_defined_variable): Likewise.
2323 (varpool_next_defined_variable): Likewise.
2324 (cgraph_first_defined_function): Likewise.
2325 (cgraph_next_defined_function): Likewise.
2326 (cgraph_first_function): Likewise.
2327 (cgraph_next_function): Likewise.
2328 (cgraph_first_function_with_gimple_body): Likewise.
2329 (cgraph_next_function_with_gimple_body): Likewise.
2330 (cgraph_alias_target): Likewise.
2331 (varpool_alias_target): Likewise.
2332 (cgraph_function_or_thunk_node): Likewise.
2333 (varpool_variable_node): Likewise.
2334 (symtab_real_symbol_p): Likewise.
2335 * cgraphunit.c (referred_to_p): Likewise.
2336 (analyze_functions): Likewise.
2337 (handle_alias_pairs): Likewise.
2338 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2339 * gimple-ssa.h (gimple_vuse_op): Likewise.
2340 (gimple_vdef_op): Likewise.
2341 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2342 * gimple.c (gimple_build_asm_1): Likewise.
2343 (gimple_build_try): Likewise.
2344 (gimple_build_resx): Likewise.
2345 (gimple_build_eh_dispatch): Likewise.
2346 (gimple_build_omp_for): Likewise.
2347 (gimple_omp_for_set_clauses): Likewise.
2349 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2350 (is_a_helper <gimple_statement_asm *>::test): ...this.
2351 (is_a_helper <gimple_statement_bind>::test): Convert to...
2352 (is_a_helper <gimple_statement_bind *>::test): ...this.
2353 (is_a_helper <gimple_statement_call>::test): Convert to...
2354 (is_a_helper <gimple_statement_call *>::test): ...this.
2355 (is_a_helper <gimple_statement_catch>::test): Convert to...
2356 (is_a_helper <gimple_statement_catch *>::test): ...this.
2357 (is_a_helper <gimple_statement_resx>::test): Convert to...
2358 (is_a_helper <gimple_statement_resx *>::test): ...this.
2359 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2360 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2361 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2362 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2363 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2364 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2365 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2366 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2367 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2368 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2369 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2370 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2371 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2372 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2373 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2374 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2375 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2376 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2377 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2378 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2379 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2380 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2381 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2382 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2383 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2384 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2385 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2386 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2387 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2388 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2389 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2390 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2391 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2392 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2393 (is_a_helper <gimple_statement_phi>::test): Convert to...
2394 (is_a_helper <gimple_statement_phi *>::test): ...this.
2395 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2396 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2397 (is_a_helper <gimple_statement_try>::test): Convert to...
2398 (is_a_helper <gimple_statement_try *>::test): ...this.
2399 (is_a_helper <gimple_statement_wce>::test): Convert to...
2400 (is_a_helper <gimple_statement_wce *>::test): ...this.
2401 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2402 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2403 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2404 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2405 (is_a_helper <const gimple_statement_call>::test): Convert to...
2406 (is_a_helper <const gimple_statement_call *>::test): ...this.
2407 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2408 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2409 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2410 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2411 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2412 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2413 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2414 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2415 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2417 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2419 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2421 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2423 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2424 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2425 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2427 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2428 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2430 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2431 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2432 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2433 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2434 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2435 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2437 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2438 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2439 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2440 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2442 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2443 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2444 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2445 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2446 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2447 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2448 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2449 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2450 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2451 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2452 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2453 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2454 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2455 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2456 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2457 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2459 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2461 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2462 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2464 (gimple_use_ops): Update for removal of implicit pointer from the
2466 (gimple_set_use_ops): Likewise.
2467 (gimple_vuse): Likewise.
2468 (gimple_vdef): Likewise.
2469 (gimple_vuse_ptr): Likewise.
2470 (gimple_vdef_ptr): Likewise.
2471 (gimple_set_vuse): Likewise.
2472 (gimple_set_vdef): Likewise.
2473 (gimple_omp_return_set_lhs): Likewise.
2474 (gimple_omp_return_lhs): Likewise.
2475 (gimple_omp_return_lhs_ptr): Likewise.
2476 (gimple_call_fntype): Likewise.
2477 (gimple_call_set_fntype): Likewise.
2478 (gimple_call_set_internal_fn): Likewise.
2479 (gimple_call_use_set): Likewise.
2480 (gimple_call_clobber_set): Likewise.
2481 (gimple_bind_vars): Likewise.
2482 (gimple_bind_set_vars): Likewise.
2483 (gimple_bind_body_ptr): Likewise.
2484 (gimple_bind_set_body): Likewise.
2485 (gimple_bind_add_stmt): Likewise.
2486 (gimple_bind_block): Likewise.
2487 (gimple_bind_set_block): Likewise.
2488 (gimple_asm_ninputs): Likewise.
2489 (gimple_asm_noutputs): Likewise.
2490 (gimple_asm_nclobbers): Likewise.
2491 (gimple_asm_nlabels): Likewise.
2492 (gimple_asm_input_op): Likewise.
2493 (gimple_asm_input_op_ptr): Likewise.
2494 (gimple_asm_output_op): Likewise.
2495 (gimple_asm_output_op_ptr): Likewise.
2496 (gimple_asm_set_output_op): Likewise.
2497 (gimple_asm_clobber_op): Likewise.
2498 (gimple_asm_set_clobber_op): Likewise.
2499 (gimple_asm_label_op): Likewise.
2500 (gimple_asm_set_label_op): Likewise.
2501 (gimple_asm_string): Likewise.
2502 (gimple_catch_types): Likewise.
2503 (gimple_catch_types_ptr): Likewise.
2504 (gimple_catch_handler_ptr): Likewise.
2505 (gimple_catch_set_types): Likewise.
2506 (gimple_catch_set_handler): Likewise.
2507 (gimple_eh_filter_types): Likewise.
2508 (gimple_eh_filter_types_ptr): Likewise.
2509 (gimple_eh_filter_failure_ptr): Likewise.
2510 (gimple_eh_filter_set_types): Likewise.
2511 (gimple_eh_filter_set_failure): Likewise.
2512 (gimple_eh_must_not_throw_fndecl): Likewise.
2513 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2514 (gimple_eh_else_n_body_ptr): Likewise.
2515 (gimple_eh_else_e_body_ptr): Likewise.
2516 (gimple_eh_else_set_n_body): Likewise.
2517 (gimple_eh_else_set_e_body): Likewise.
2518 (gimple_try_eval_ptr): Likewise.
2519 (gimple_try_cleanup_ptr): Likewise.
2520 (gimple_try_set_eval): Likewise.
2521 (gimple_try_set_cleanup): Likewise.
2522 (gimple_wce_cleanup_ptr): Likewise.
2523 (gimple_wce_set_cleanup): Likewise.
2524 (gimple_phi_capacity): Likewise.
2525 (gimple_phi_num_args): Likewise.
2526 (gimple_phi_result): Likewise.
2527 (gimple_phi_result_ptr): Likewise.
2528 (gimple_phi_set_result): Likewise.
2529 (gimple_phi_arg): Likewise.
2530 (gimple_phi_set_arg): Likewise.
2531 (gimple_resx_region): Likewise.
2532 (gimple_resx_set_region): Likewise.
2533 (gimple_eh_dispatch_region): Likewise.
2534 (gimple_eh_dispatch_set_region): Likewise.
2535 (gimple_omp_critical_name): Likewise.
2536 (gimple_omp_critical_name_ptr): Likewise.
2537 (gimple_omp_critical_set_name): Likewise.
2538 (gimple_omp_for_clauses): Likewise.
2539 (gimple_omp_for_clauses_ptr): Likewise.
2540 (gimple_omp_for_set_clauses): Likewise.
2541 (gimple_omp_for_collapse): Likewise.
2542 (gimple_omp_for_index): Likewise.
2543 (gimple_omp_for_index_ptr): Likewise.
2544 (gimple_omp_for_set_index): Likewise.
2545 (gimple_omp_for_initial): Likewise.
2546 (gimple_omp_for_initial_ptr): Likewise.
2547 (gimple_omp_for_set_initial): Likewise.
2548 (gimple_omp_for_final): Likewise.
2549 (gimple_omp_for_final_ptr): Likewise.
2550 (gimple_omp_for_set_final): Likewise.
2551 (gimple_omp_for_incr): Likewise.
2552 (gimple_omp_for_incr_ptr): Likewise.
2553 (gimple_omp_for_set_incr): Likewise.
2554 (gimple_omp_for_pre_body_ptr): Likewise.
2555 (gimple_omp_for_set_pre_body): Likewise.
2556 (gimple_omp_parallel_clauses): Likewise.
2557 (gimple_omp_parallel_clauses_ptr): Likewise.
2558 (gimple_omp_parallel_set_clauses): Likewise.
2559 (gimple_omp_parallel_child_fn): Likewise.
2560 (gimple_omp_parallel_child_fn_ptr): Likewise.
2561 (gimple_omp_parallel_set_child_fn): Likewise.
2562 (gimple_omp_parallel_data_arg): Likewise.
2563 (gimple_omp_parallel_data_arg_ptr): Likewise.
2564 (gimple_omp_parallel_set_data_arg): Likewise.
2565 (gimple_omp_task_clauses): Likewise.
2566 (gimple_omp_task_clauses_ptr): Likewise.
2567 (gimple_omp_task_set_clauses): Likewise.
2568 (gimple_omp_task_child_fn): Likewise.
2569 (gimple_omp_task_child_fn_ptr): Likewise.
2570 (gimple_omp_task_set_child_fn): Likewise.
2571 (gimple_omp_task_data_arg): Likewise.
2572 (gimple_omp_task_data_arg_ptr): Likewise.
2573 (gimple_omp_task_set_data_arg): Likewise.
2574 (gimple_omp_taskreg_clauses): Likewise.
2575 (gimple_omp_taskreg_clauses_ptr): Likewise.
2576 (gimple_omp_taskreg_set_clauses): Likewise.
2577 (gimple_omp_taskreg_child_fn): Likewise.
2578 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2579 (gimple_omp_taskreg_set_child_fn): Likewise.
2580 (gimple_omp_taskreg_data_arg): Likewise.
2581 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2582 (gimple_omp_taskreg_set_data_arg): Likewise.
2583 (gimple_omp_task_copy_fn): Likewise.
2584 (gimple_omp_task_copy_fn_ptr): Likewise.
2585 (gimple_omp_task_set_copy_fn): Likewise.
2586 (gimple_omp_task_arg_size): Likewise.
2587 (gimple_omp_task_arg_size_ptr): Likewise.
2588 (gimple_omp_task_set_arg_size): Likewise.
2589 (gimple_omp_task_arg_align): Likewise.
2590 (gimple_omp_task_arg_align_ptr): Likewise.
2591 (gimple_omp_task_set_arg_align): Likewise.
2592 (gimple_omp_single_clauses): Likewise.
2593 (gimple_omp_single_clauses_ptr): Likewise.
2594 (gimple_omp_single_set_clauses): Likewise.
2595 (gimple_omp_target_clauses): Likewise.
2596 (gimple_omp_target_clauses_ptr): Likewise.
2597 (gimple_omp_target_set_clauses): Likewise.
2598 (gimple_omp_target_child_fn): Likewise.
2599 (gimple_omp_target_child_fn_ptr): Likewise.
2600 (gimple_omp_target_set_child_fn): Likewise.
2601 (gimple_omp_target_data_arg): Likewise.
2602 (gimple_omp_target_data_arg_ptr): Likewise.
2603 (gimple_omp_target_set_data_arg): Likewise.
2604 (gimple_omp_teams_clauses): Likewise.
2605 (gimple_omp_teams_clauses_ptr): Likewise.
2606 (gimple_omp_teams_set_clauses): Likewise.
2607 (gimple_omp_sections_clauses): Likewise.
2608 (gimple_omp_sections_clauses_ptr): Likewise.
2609 (gimple_omp_sections_set_clauses): Likewise.
2610 (gimple_omp_sections_control): Likewise.
2611 (gimple_omp_sections_control_ptr): Likewise.
2612 (gimple_omp_sections_set_control): Likewise.
2613 (gimple_omp_for_set_cond): Likewise.
2614 (gimple_omp_for_cond): Likewise.
2615 (gimple_omp_atomic_store_set_val): Likewise.
2616 (gimple_omp_atomic_store_val): Likewise.
2617 (gimple_omp_atomic_store_val_ptr): Likewise.
2618 (gimple_omp_atomic_load_set_lhs): Likewise.
2619 (gimple_omp_atomic_load_lhs): Likewise.
2620 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2621 (gimple_omp_atomic_load_set_rhs): Likewise.
2622 (gimple_omp_atomic_load_rhs): Likewise.
2623 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2624 (gimple_omp_continue_control_def): Likewise.
2625 (gimple_omp_continue_control_def_ptr): Likewise.
2626 (gimple_omp_continue_set_control_def): Likewise.
2627 (gimple_omp_continue_control_use): Likewise.
2628 (gimple_omp_continue_control_use_ptr): Likewise.
2629 (gimple_omp_continue_set_control_use): Likewise.
2630 (gimple_transaction_body_ptr): Likewise.
2631 (gimple_transaction_label): Likewise.
2632 (gimple_transaction_label_ptr): Likewise.
2633 (gimple_transaction_set_body): Likewise.
2634 (gimple_transaction_set_label): Likewise.
2636 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2637 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2638 * ipa-ref.c (ipa_record_reference): Likewise.
2639 * ipa-reference.c (analyze_function): Likewise.
2640 (ipa_reference_write_optimization_summary): Likewise.
2641 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2642 (address_taken_from_non_vtable_p): Likewise.
2643 (comdat_can_be_unshared_p_1): Likewise.
2644 * lto-cgraph.c (lto_output_ref): Likewise.
2645 (add_references): Likewise.
2646 (compute_ltrans_boundary): Likewise.
2647 (output_symtab): Likewise.
2648 (input_ref): Likewise.
2649 (input_cgraph_1): Likewise.
2650 (output_cgraph_opt_summary): Likewise.
2651 * lto-streamer-out.c (lto_output): Likewise.
2652 (output_symbol_p): Likewise.
2653 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2654 (lsei_start_function_in_partition): Likewise.
2655 (lsei_next_variable_in_partition): Likewise.
2656 (lsei_start_variable_in_partition): Likewise.
2657 * symtab.c (insert_to_assembler_name_hash): Likewise.
2658 (unlink_from_assembler_name_hash): Likewise.
2659 (symtab_unregister_node): Likewise.
2660 (symtab_remove_node): Likewise.
2661 (dump_symtab_node): Likewise.
2662 (verify_symtab_base): Likewise.
2663 (verify_symtab_node): Likewise.
2664 (symtab_make_decl_local): Likewise.
2665 (symtab_alias_ultimate_target): Likewise.
2666 (symtab_resolve_alias): Likewise.
2667 (symtab_get_symbol_partitioning_class): Likewise.
2668 * tree-phinodes.c (allocate_phi_node): Likewise.
2669 (reserve_phi_args_for_new_edge): Likewise.
2670 (remove_phi_args): Likewise.
2671 * varpool.c (varpool_node_for_asm): Likewise.
2672 (varpool_remove_unreferenced_decls): Likewise.
2674 2014-04-23 Jeff Law <law@redhat.com>
2676 PR tree-optimization/60902
2677 * tree-ssa-threadedge.c
2678 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2679 invalidate outputs from statements that do not produce useful
2680 outputs for threading.
2682 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2684 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2685 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2686 machine descriptions for Stack Smashing Protector.
2688 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2690 * aarch64.md (<optab>_rol<mode>3): New pattern.
2691 (<optab>_rolsi3_uxtw): Likewise.
2692 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2694 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2696 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2697 (arm_cortex_a12_tune): Likewise.
2699 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2701 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2703 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2705 * config/arm/arm.md (arm_rev16si2): New pattern.
2706 (arm_rev16si2_alt): Likewise.
2707 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2709 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2711 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2712 (rev16<mode>2_alt): Likewise.
2713 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2714 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2715 (aarch_rev16_shleft_mask_imm_p): Likewise.
2716 (aarch_rev16_p_1): Likewise.
2717 (aarch_rev16_p): Likewise.
2718 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2719 (aarch_rev16_shright_mask_imm_p): Likewise.
2720 (aarch_rev16_shleft_mask_imm_p): Likewise.
2722 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2724 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2725 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2727 (cortex_a53_extra_costs): Likewise.
2728 (cortex_a57_extra_costs): Likewise.
2729 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2730 (cortexa7_extra_costs): Likewise.
2731 (cortexa8_extra_costs): Likewise.
2732 (cortexa12_extra_costs): Likewise.
2733 (cortexa15_extra_costs): Likewise.
2734 (v7m_extra_costs): Likewise.
2735 (arm_new_rtx_costs): Handle BSWAP.
2737 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2739 * config/arm/arm.c (cortexa8_extra_costs): New table.
2740 (arm_cortex_a8_tune): New tuning struct.
2741 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2743 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2745 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2747 2014-04-23 Richard Biener <rguenther@suse.de>
2749 * Makefile.in (OBJS): Remove loop-unswitch.o.
2750 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2751 * passes.def (pass_rtl_unswitch): Likewise.
2752 * loop-init.c (gate_rtl_unswitch): Likewise.
2753 (rtl_unswitch): Likewise.
2754 (pass_data_rtl_unswitch): Likewise.
2755 (pass_rtl_unswitch): Likewise.
2756 (make_pass_rtl_unswitch): Likewise.
2757 * rtl.h (reversed_condition): Likewise.
2758 (compare_and_jump_seq): Likewise.
2759 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2761 * loop-unroll.c (compare_and_jump_seq): Likewise.
2763 2014-04-23 Richard Biener <rguenther@suse.de>
2765 PR tree-optimization/60903
2766 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2767 commented code block.
2768 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2769 loop flags to newly created BBs and edges.
2771 2014-04-23 Nick Clifton <nickc@redhat.com>
2773 * config/msp430/msp430.c (msp430_handle_option): Move function
2775 (msp430_option_override): Simplify mcu and mcpu option handling.
2776 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2777 support for -mhwmult command line option.
2778 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2779 -mhwmult command line option.
2780 (msp430_hwmult_enabled): Delete.
2781 (msp43o_output_labelref): Add support for -mhwmult command line option.
2782 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2783 (umulsidi3): Likewise.
2784 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2785 (mcpu, mlarge, msmall): Likewise.
2786 (mhwmult): New option.
2787 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2789 (msp430_is_f5_mcu): Remove prototype.
2790 (msp430_use_f5_series_hwmult): Add prototype.
2791 * config/msp430/msp430-opts.h: New file.
2792 * common/config/msp430: New directory.
2793 * common/config/msp430/msp430-common.c: New file.
2794 * config.gcc (msp430): Remove target_has_targetm_common.
2795 * doc/invoke.texi: Document -mhwmult command line option.
2797 2014-04-23 Nick Clifton <nickc@redhat.com>
2799 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2800 default-manifest.o if it can be found in the search path.
2801 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2803 2014-04-23 Terry Guo <terry.guo@arm.com>
2805 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2807 2014-04-23 Richard Biener <rguenther@suse.de>
2810 * tree-inline.c (declare_return_variable): Use mark_addressable.
2812 2014-04-23 Richard Biener <rguenther@suse.de>
2815 * loop-init.c (loop_optimizer_init): Make sure to apply
2816 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2818 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2821 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2823 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2824 if flag_sanitize_undefined_trap_on_error.
2825 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2826 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2827 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2828 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2829 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2830 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2831 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2832 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2833 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2834 * ubsan.c (ubsan_instrument_unreachable): Return
2835 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2836 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2837 if flag_sanitize_undefined_trap_on_error and
2838 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2839 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2840 instrument_bool_enum_load): Emit __builtin_trap () if
2841 flag_sanitize_undefined_trap_on_error and
2842 __builtin_handle_*_abort () if !flag_sanitize_recover.
2843 * doc/invoke.texi (-fsanitize-recover,
2844 -fsanitize-undefined-trap-on-error): Document.
2846 2014-04-22 Christian Bruel <christian.bruel@st.com>
2848 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2849 Force immediates to SImode.
2851 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2853 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2855 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2856 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2857 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2858 (nios2_fpu_insn): Add entry for round.
2859 (N2FPU_NO_ERRNO_P): Define.
2860 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2862 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2864 2014-04-22 Richard Henderson <rth@redhat.com>
2866 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2867 (add<GPI>3_compare0): Remove leading * from name.
2868 (add<GPI>3_carryin): Likewise.
2869 (sub<GPI>3_compare0): Likewise.
2870 (sub<GPI>3_carryin): Likewise.
2871 (<su_optab>mulditi3): New expander.
2872 (multi3): New expander.
2873 (madd<GPI>): Remove leading * from name.
2875 2014-04-22 Martin Jambor <mjambor@suse.cz>
2877 * cgraphclones.c (cgraph_function_versioning): Copy
2878 ipa_transforms_to_apply instead of asserting it is empty.
2880 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2883 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2884 on count_exp to get mode.
2886 2014-04-22 Andrew Pinski <apinski@cavium.com>
2888 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2889 Handle TLS for ILP32.
2890 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2891 (tlsie_small_<mode>): this and handle PTR.
2892 (tlsie_small_sidi): New pattern.
2893 (tlsle_small): Change to an expand to handle ILP32.
2894 (tlsle_small_<mode>): New pattern.
2895 (tlsdesc_small): Rename to ...
2896 (tlsdesc_small_<mode>): this and handle PTR.
2898 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2900 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2902 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2904 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2905 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2906 (aarch64_types_signed_poly_qualifiers): Likewise.
2907 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2908 (aarch64_types_poly_signed_qualifiers): Likewise.
2909 (TYPES_REINTERP_SS): Type macro added.
2910 (TYPES_REINTERP_SU): Likewise.
2911 (TYPES_REINTERP_SP): Likewise.
2912 (TYPES_REINTERP_US): Likewise.
2913 (TYPES_REINTERP_PS): Likewise.
2914 (aarch64_fold_builtin): New expression folding added.
2915 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2916 Declarations removed.
2917 (REINTERP_SS): Declarations added.
2918 (REINTERP_US): Likewise.
2919 (REINTERP_PS): Likewise.
2920 (REINTERP_SU): Likewise.
2921 (REINTERP_SP): Likewise.
2922 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2923 (vreinterpretq_p8_f64): Likewise.
2924 (vreinterpret_p16_f64): Likewise.
2925 (vreinterpretq_p16_f64): Likewise.
2926 (vreinterpret_f32_f64): Likewise.
2927 (vreinterpretq_f32_f64): Likewise.
2928 (vreinterpret_f64_f32): Likewise.
2929 (vreinterpret_f64_p8): Likewise.
2930 (vreinterpret_f64_p16): Likewise.
2931 (vreinterpret_f64_s8): Likewise.
2932 (vreinterpret_f64_s16): Likewise.
2933 (vreinterpret_f64_s32): Likewise.
2934 (vreinterpret_f64_s64): Likewise.
2935 (vreinterpret_f64_u8): Likewise.
2936 (vreinterpret_f64_u16): Likewise.
2937 (vreinterpret_f64_u32): Likewise.
2938 (vreinterpret_f64_u64): Likewise.
2939 (vreinterpretq_f64_f32): Likewise.
2940 (vreinterpretq_f64_p8): Likewise.
2941 (vreinterpretq_f64_p16): Likewise.
2942 (vreinterpretq_f64_s8): Likewise.
2943 (vreinterpretq_f64_s16): Likewise.
2944 (vreinterpretq_f64_s32): Likewise.
2945 (vreinterpretq_f64_s64): Likewise.
2946 (vreinterpretq_f64_u8): Likewise.
2947 (vreinterpretq_f64_u16): Likewise.
2948 (vreinterpretq_f64_u32): Likewise.
2949 (vreinterpretq_f64_u64): Likewise.
2950 (vreinterpret_s64_f64): Likewise.
2951 (vreinterpretq_s64_f64): Likewise.
2952 (vreinterpret_u64_f64): Likewise.
2953 (vreinterpretq_u64_f64): Likewise.
2954 (vreinterpret_s8_f64): Likewise.
2955 (vreinterpretq_s8_f64): Likewise.
2956 (vreinterpret_s16_f64): Likewise.
2957 (vreinterpretq_s16_f64): Likewise.
2958 (vreinterpret_s32_f64): Likewise.
2959 (vreinterpretq_s32_f64): Likewise.
2960 (vreinterpret_u8_f64): Likewise.
2961 (vreinterpretq_u8_f64): Likewise.
2962 (vreinterpret_u16_f64): Likewise.
2963 (vreinterpretq_u16_f64): Likewise.
2964 (vreinterpret_u32_f64): Likewise.
2965 (vreinterpretq_u32_f64): Likewise.
2967 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2969 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2970 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2971 (vreinterpret_p8_s8): Likewise.
2972 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2973 (vreinterpret_p8_s16): Likewise.
2974 (vreinterpret_p8_s32): Likewise.
2975 (vreinterpret_p8_s64): Likewise.
2976 (vreinterpret_p8_f32): Likewise.
2977 (vreinterpret_p8_u8): Likewise.
2978 (vreinterpret_p8_u16): Likewise.
2979 (vreinterpret_p8_u32): Likewise.
2980 (vreinterpret_p8_u64): Likewise.
2981 (vreinterpret_p8_p16): Likewise.
2982 (vreinterpretq_p8_s8): Likewise.
2983 (vreinterpretq_p8_s16): Likewise.
2984 (vreinterpretq_p8_s32): Likewise.
2985 (vreinterpretq_p8_s64): Likewise.
2986 (vreinterpretq_p8_f32): Likewise.
2987 (vreinterpretq_p8_u8): Likewise.
2988 (vreinterpretq_p8_u16): Likewise.
2989 (vreinterpretq_p8_u32): Likewise.
2990 (vreinterpretq_p8_u64): Likewise.
2991 (vreinterpretq_p8_p16): Likewise.
2992 (vreinterpret_p16_s8): Likewise.
2993 (vreinterpret_p16_s16): Likewise.
2994 (vreinterpret_p16_s32): Likewise.
2995 (vreinterpret_p16_s64): Likewise.
2996 (vreinterpret_p16_f32): Likewise.
2997 (vreinterpret_p16_u8): Likewise.
2998 (vreinterpret_p16_u16): Likewise.
2999 (vreinterpret_p16_u32): Likewise.
3000 (vreinterpret_p16_u64): Likewise.
3001 (vreinterpret_p16_p8): Likewise.
3002 (vreinterpretq_p16_s8): Likewise.
3003 (vreinterpretq_p16_s16): Likewise.
3004 (vreinterpretq_p16_s32): Likewise.
3005 (vreinterpretq_p16_s64): Likewise.
3006 (vreinterpretq_p16_f32): Likewise.
3007 (vreinterpretq_p16_u8): Likewise.
3008 (vreinterpretq_p16_u16): Likewise.
3009 (vreinterpretq_p16_u32): Likewise.
3010 (vreinterpretq_p16_u64): Likewise.
3011 (vreinterpretq_p16_p8): Likewise.
3012 (vreinterpret_f32_s8): Likewise.
3013 (vreinterpret_f32_s16): Likewise.
3014 (vreinterpret_f32_s32): Likewise.
3015 (vreinterpret_f32_s64): Likewise.
3016 (vreinterpret_f32_u8): Likewise.
3017 (vreinterpret_f32_u16): Likewise.
3018 (vreinterpret_f32_u32): Likewise.
3019 (vreinterpret_f32_u64): Likewise.
3020 (vreinterpret_f32_p8): Likewise.
3021 (vreinterpret_f32_p16): Likewise.
3022 (vreinterpretq_f32_s8): Likewise.
3023 (vreinterpretq_f32_s16): Likewise.
3024 (vreinterpretq_f32_s32): Likewise.
3025 (vreinterpretq_f32_s64): Likewise.
3026 (vreinterpretq_f32_u8): Likewise.
3027 (vreinterpretq_f32_u16): Likewise.
3028 (vreinterpretq_f32_u32): Likewise.
3029 (vreinterpretq_f32_u64): Likewise.
3030 (vreinterpretq_f32_p8): Likewise.
3031 (vreinterpretq_f32_p16): Likewise.
3032 (vreinterpret_s64_s8): Likewise.
3033 (vreinterpret_s64_s16): Likewise.
3034 (vreinterpret_s64_s32): Likewise.
3035 (vreinterpret_s64_f32): Likewise.
3036 (vreinterpret_s64_u8): Likewise.
3037 (vreinterpret_s64_u16): Likewise.
3038 (vreinterpret_s64_u32): Likewise.
3039 (vreinterpret_s64_u64): Likewise.
3040 (vreinterpret_s64_p8): Likewise.
3041 (vreinterpret_s64_p16): Likewise.
3042 (vreinterpretq_s64_s8): Likewise.
3043 (vreinterpretq_s64_s16): Likewise.
3044 (vreinterpretq_s64_s32): Likewise.
3045 (vreinterpretq_s64_f32): Likewise.
3046 (vreinterpretq_s64_u8): Likewise.
3047 (vreinterpretq_s64_u16): Likewise.
3048 (vreinterpretq_s64_u32): Likewise.
3049 (vreinterpretq_s64_u64): Likewise.
3050 (vreinterpretq_s64_p8): Likewise.
3051 (vreinterpretq_s64_p16): Likewise.
3052 (vreinterpret_u64_s8): Likewise.
3053 (vreinterpret_u64_s16): Likewise.
3054 (vreinterpret_u64_s32): Likewise.
3055 (vreinterpret_u64_s64): Likewise.
3056 (vreinterpret_u64_f32): Likewise.
3057 (vreinterpret_u64_u8): Likewise.
3058 (vreinterpret_u64_u16): Likewise.
3059 (vreinterpret_u64_u32): Likewise.
3060 (vreinterpret_u64_p8): Likewise.
3061 (vreinterpret_u64_p16): Likewise.
3062 (vreinterpretq_u64_s8): Likewise.
3063 (vreinterpretq_u64_s16): Likewise.
3064 (vreinterpretq_u64_s32): Likewise.
3065 (vreinterpretq_u64_s64): Likewise.
3066 (vreinterpretq_u64_f32): Likewise.
3067 (vreinterpretq_u64_u8): Likewise.
3068 (vreinterpretq_u64_u16): Likewise.
3069 (vreinterpretq_u64_u32): Likewise.
3070 (vreinterpretq_u64_p8): Likewise.
3071 (vreinterpretq_u64_p16): Likewise.
3072 (vreinterpret_s8_s16): Likewise.
3073 (vreinterpret_s8_s32): Likewise.
3074 (vreinterpret_s8_s64): Likewise.
3075 (vreinterpret_s8_f32): Likewise.
3076 (vreinterpret_s8_u8): Likewise.
3077 (vreinterpret_s8_u16): Likewise.
3078 (vreinterpret_s8_u32): Likewise.
3079 (vreinterpret_s8_u64): Likewise.
3080 (vreinterpret_s8_p8): Likewise.
3081 (vreinterpret_s8_p16): Likewise.
3082 (vreinterpretq_s8_s16): Likewise.
3083 (vreinterpretq_s8_s32): Likewise.
3084 (vreinterpretq_s8_s64): Likewise.
3085 (vreinterpretq_s8_f32): Likewise.
3086 (vreinterpretq_s8_u8): Likewise.
3087 (vreinterpretq_s8_u16): Likewise.
3088 (vreinterpretq_s8_u32): Likewise.
3089 (vreinterpretq_s8_u64): Likewise.
3090 (vreinterpretq_s8_p8): Likewise.
3091 (vreinterpretq_s8_p16): Likewise.
3092 (vreinterpret_s16_s8): Likewise.
3093 (vreinterpret_s16_s32): Likewise.
3094 (vreinterpret_s16_s64): Likewise.
3095 (vreinterpret_s16_f32): Likewise.
3096 (vreinterpret_s16_u8): Likewise.
3097 (vreinterpret_s16_u16): Likewise.
3098 (vreinterpret_s16_u32): Likewise.
3099 (vreinterpret_s16_u64): Likewise.
3100 (vreinterpret_s16_p8): Likewise.
3101 (vreinterpret_s16_p16): Likewise.
3102 (vreinterpretq_s16_s8): Likewise.
3103 (vreinterpretq_s16_s32): Likewise.
3104 (vreinterpretq_s16_s64): Likewise.
3105 (vreinterpretq_s16_f32): Likewise.
3106 (vreinterpretq_s16_u8): Likewise.
3107 (vreinterpretq_s16_u16): Likewise.
3108 (vreinterpretq_s16_u32): Likewise.
3109 (vreinterpretq_s16_u64): Likewise.
3110 (vreinterpretq_s16_p8): Likewise.
3111 (vreinterpretq_s16_p16): Likewise.
3112 (vreinterpret_s32_s8): Likewise.
3113 (vreinterpret_s32_s16): Likewise.
3114 (vreinterpret_s32_s64): Likewise.
3115 (vreinterpret_s32_f32): Likewise.
3116 (vreinterpret_s32_u8): Likewise.
3117 (vreinterpret_s32_u16): Likewise.
3118 (vreinterpret_s32_u32): Likewise.
3119 (vreinterpret_s32_u64): Likewise.
3120 (vreinterpret_s32_p8): Likewise.
3121 (vreinterpret_s32_p16): Likewise.
3122 (vreinterpretq_s32_s8): Likewise.
3123 (vreinterpretq_s32_s16): Likewise.
3124 (vreinterpretq_s32_s64): Likewise.
3125 (vreinterpretq_s32_f32): Likewise.
3126 (vreinterpretq_s32_u8): Likewise.
3127 (vreinterpretq_s32_u16): Likewise.
3128 (vreinterpretq_s32_u32): Likewise.
3129 (vreinterpretq_s32_u64): Likewise.
3130 (vreinterpretq_s32_p8): Likewise.
3131 (vreinterpretq_s32_p16): Likewise.
3132 (vreinterpret_u8_s8): Likewise.
3133 (vreinterpret_u8_s16): Likewise.
3134 (vreinterpret_u8_s32): Likewise.
3135 (vreinterpret_u8_s64): Likewise.
3136 (vreinterpret_u8_f32): Likewise.
3137 (vreinterpret_u8_u16): Likewise.
3138 (vreinterpret_u8_u32): Likewise.
3139 (vreinterpret_u8_u64): Likewise.
3140 (vreinterpret_u8_p8): Likewise.
3141 (vreinterpret_u8_p16): Likewise.
3142 (vreinterpretq_u8_s8): Likewise.
3143 (vreinterpretq_u8_s16): Likewise.
3144 (vreinterpretq_u8_s32): Likewise.
3145 (vreinterpretq_u8_s64): Likewise.
3146 (vreinterpretq_u8_f32): Likewise.
3147 (vreinterpretq_u8_u16): Likewise.
3148 (vreinterpretq_u8_u32): Likewise.
3149 (vreinterpretq_u8_u64): Likewise.
3150 (vreinterpretq_u8_p8): Likewise.
3151 (vreinterpretq_u8_p16): Likewise.
3152 (vreinterpret_u16_s8): Likewise.
3153 (vreinterpret_u16_s16): Likewise.
3154 (vreinterpret_u16_s32): Likewise.
3155 (vreinterpret_u16_s64): Likewise.
3156 (vreinterpret_u16_f32): Likewise.
3157 (vreinterpret_u16_u8): Likewise.
3158 (vreinterpret_u16_u32): Likewise.
3159 (vreinterpret_u16_u64): Likewise.
3160 (vreinterpret_u16_p8): Likewise.
3161 (vreinterpret_u16_p16): Likewise.
3162 (vreinterpretq_u16_s8): Likewise.
3163 (vreinterpretq_u16_s16): Likewise.
3164 (vreinterpretq_u16_s32): Likewise.
3165 (vreinterpretq_u16_s64): Likewise.
3166 (vreinterpretq_u16_f32): Likewise.
3167 (vreinterpretq_u16_u8): Likewise.
3168 (vreinterpretq_u16_u32): Likewise.
3169 (vreinterpretq_u16_u64): Likewise.
3170 (vreinterpretq_u16_p8): Likewise.
3171 (vreinterpretq_u16_p16): Likewise.
3172 (vreinterpret_u32_s8): Likewise.
3173 (vreinterpret_u32_s16): Likewise.
3174 (vreinterpret_u32_s32): Likewise.
3175 (vreinterpret_u32_s64): Likewise.
3176 (vreinterpret_u32_f32): Likewise.
3177 (vreinterpret_u32_u8): Likewise.
3178 (vreinterpret_u32_u16): Likewise.
3179 (vreinterpret_u32_u64): Likewise.
3180 (vreinterpret_u32_p8): Likewise.
3181 (vreinterpret_u32_p16): Likewise.
3182 (vreinterpretq_u32_s8): Likewise.
3183 (vreinterpretq_u32_s16): Likewise.
3184 (vreinterpretq_u32_s32): Likewise.
3185 (vreinterpretq_u32_s64): Likewise.
3186 (vreinterpretq_u32_f32): Likewise.
3187 (vreinterpretq_u32_u8): Likewise.
3188 (vreinterpretq_u32_u16): Likewise.
3189 (vreinterpretq_u32_u64): Likewise.
3190 (vreinterpretq_u32_p8): Likewise.
3191 (vreinterpretq_u32_p16): Likewise.
3193 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3195 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3197 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3199 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3200 (vqnegd_s64): Likewise.
3201 (vqabs_s64): Likewise.
3202 (vqabsd_s64): Likewise.
3204 2014-04-22 Richard Henderson <rth@redhat.com>
3206 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3207 computation to the top of the loop.
3209 2014-04-22 Renlin <renlin.li@arm.com>
3210 Jiong Wang <jiong.wang@arm.com>
3212 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3213 * config/aarch64/aarch64.c (aarch64_layout_frame)
3214 (aarch64_initial_elimination_offset): Likewise.
3216 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3218 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3221 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3223 * machmode.h (bitwise_mode_for_mode): Declare.
3224 * stor-layout.h (bitwise_type_for_mode): Likewise.
3225 * stor-layout.c (bitwise_mode_for_mode): New function.
3226 (bitwise_type_for_mode): Likewise.
3227 * builtins.c (fold_builtin_memory_op): Use it instead of
3228 int_mode_for_mode and build_nonstandard_integer_type.
3230 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3232 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3233 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3234 (*-*-solaris2*): Simplify.
3235 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3236 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3237 *-*-solaris2.9* handling.
3239 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3241 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3242 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3244 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3245 * configure: Regenerate.
3247 * config/i386/sol2-9.h: Remove.
3249 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3250 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3251 Remove Solaris 9 references.
3253 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3255 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3256 (floatuns<GPI:mode><GPF:mode>2): Remove.
3257 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3258 and floatuns conversions.
3259 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3260 and floatuns conversions.
3261 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3262 (w1,w2): New mode attributes for inequal width conversions.
3264 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3266 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3267 the output asm format.
3269 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3271 * config/aarch64/aarch64-simd.md
3272 (aarch64_cm<optab>di): Always split.
3273 (*aarch64_cm<optab>di): New.
3274 (aarch64_cmtstdi): Always split.
3275 (*aarch64_cmtstdi): New.
3277 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3279 PR tree-optimization/60823
3280 * omp-low.c (ipa_simd_modify_function_body): Go through
3281 all SSA_NAMEs and for those refering to vector arguments
3282 which are going to be replaced adjust SSA_NAME_VAR and,
3283 if it is a default definition, change it into a non-default
3284 definition assigned at the beginning of function from new_decl.
3285 (ipa_simd_modify_stmt_ops): Rewritten.
3286 * tree-dfa.c (set_ssa_default_def): When removing default def,
3287 check for NULL loc instead of NULL *loc.
3289 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3291 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3292 restrictions on core registers for DImode values in Thumb2.
3294 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3296 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3297 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3299 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3301 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3302 (*iordi_notzesidi_di): Likewise.
3303 (*iordi_notsesidi_di): Likewise.
3305 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3307 * config/arm/arm-protos.h (tune_params): New struct members.
3308 * config/arm/arm.c: Initialise tune_params per processor.
3309 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3310 for speed, based on new tune_params.
3312 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3314 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3315 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3316 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3317 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3318 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3319 (vrnda_f64): Likewise.
3320 (vrndi_f64): Likewise.
3321 (vrndm_f64): Likewise.
3322 (vrndn_f64): Likewise.
3323 (vrndp_f64): Likewise.
3324 (vrndx_f64): Likewise.
3326 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3328 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3329 GET_MODE_SIZE argument is enum machine_mode.
3331 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3334 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3335 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3337 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3340 * asan.c (asan_emit_stack_protection): Force the base to align to
3341 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3342 appropriate bits if STRICT_ALIGNMENT.
3343 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3345 (expand_used_vars): Leave a space in the stack frame for alignment
3346 if STRICT_ALIGNMENT.
3348 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3350 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3352 (gimple_store_p): Likewise.
3353 (gimple_assign_load_p): Likewise.
3354 (gimple_assign_cast_p): Likewise.
3355 (gimple_clobber_p): Likewise.
3357 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3358 rather than a gimple.
3359 (gimple_assign_cast_p): Likewise.
3361 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3364 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3365 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3367 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3368 more debug information for E500 if -mdebug=reg.
3370 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3373 * config/i386/i386.c (ix86_expand_builtin)
3374 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3375 register for target RTX.
3376 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3378 2014-04-18 Cong Hou <congh@google.com>
3380 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3381 the widen-mult pattern by handling two operands with different sizes,
3382 and operands whose size is smaller than half of the result type.
3384 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3386 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3387 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3388 (do_estimate_edge_time): Compute it.
3389 * ipa-inline.c (want_inline_small_function_p): Bypass
3390 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3392 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3394 * ipa-inline.c (spec_rem): New static variable.
3395 (dump_overall_stats): New function.
3396 (dump_inline_stats): New function.
3398 2014-04-18 Richard Henderson <rth@redhat.com>
3400 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3401 to GET_MODE_SIZE, not a reg_class_t.
3403 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3405 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3406 (vsx_xxmrglw_<mode>): Likewise.
3408 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3411 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3412 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3413 (rs6000_init_hard_regno_mode_ok): Likewise.
3415 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3417 * ipa-inline.c (inline_small_functions): Account only non-cold
3419 * doc/invoke.texi (inline-unit-growth): Update documentation.
3421 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3423 * config/rs6000/rs6000.md (addti3, subti3): New.
3425 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3428 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3429 comment. Check optimize_insn_for_size_p instead of
3430 optimize_insn_for_speed_p.
3432 2014-04-17 Martin Jambor <mjambor@suse.cz>
3434 * gimple-iterator.c (gsi_start_edge): New function.
3435 * gimple-iterator.h (gsi_start_edge): Declare.
3436 * tree-sra.c (single_non_eh_succ): New function.
3437 (disqualify_ops_if_throwing_stmt): Renamed to
3438 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3439 having one non-EH successor BB.
3440 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3441 generate loads into replacements.
3442 (sra_modify_assign): Likewise and and also use the simple path for
3444 (sra_modify_function_body): Commit statements on edges.
3446 2014-04-17 Richard Biener <rguenther@suse.de>
3449 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3450 comparison results and add clarifying comment.
3452 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3454 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3455 (blank_mode): Initialize it.
3456 (emit_mode_size_inline, emit_mode_nunits_inline,
3457 emit_mode_inner_inline): New functions.
3458 (emit_insn_modes_h): Call them and surround their output with
3459 #if GCC_VERSION >= 4001 ... #endif.
3460 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3461 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3462 mode_* arrays if the argument is __builtin_constant_p.
3463 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3464 is enum machine_mode.
3466 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3468 * passes.c (opt_pass::execute): Adjust.
3469 (pass_manager::execute_pass_mode_switching): Likewise.
3470 (early_local_passes::execute): Likewise.
3471 (execute_one_pass): Pass cfun to the pass's execute method.
3472 * tree-pass.h (opt_pass::execute): Add function * argument.
3473 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3474 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3475 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3476 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3477 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3478 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3479 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3480 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3481 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3482 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3483 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3484 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3485 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3486 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3487 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3488 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3489 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3490 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3491 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3492 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3493 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3494 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3495 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3496 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3497 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3498 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3499 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3500 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3501 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3504 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3506 * passes.c (opt_pass::gate): Take function * argument.
3507 (gate_all_early_local_passes): Merge into
3508 (early_local_passes::gate): this.
3509 (gate_all_early_optimizations): Merge into
3510 (all_early_optimizations::gate): this.
3511 (gate_all_optimizations): Mege into
3512 (all_optimizations::gate): this.
3513 (gate_all_optimizations_g): Merge into
3514 (all_optimizations_g::gate): this.
3515 (gate_rest_of_compilation): Mege into
3516 (rest_of_compilation::gate): this.
3517 (gate_postreload): Merge into
3518 (postreload::gate): this.
3519 (dump_one_pass): Pass cfun to the pass's gate method.
3520 (execute_ipa_summary_passes): Likewise.
3521 (execute_one_pass): Likewise.
3522 (ipa_write_summaries_2): Likewise.
3523 (ipa_write_optimization_summaries_1): Likewise.
3524 (ipa_read_summaries_1): Likewise.
3525 (ipa_read_optimization_summaries_1): Likewise.
3526 (execute_ipa_stmt_fixups): Likewise.
3527 * tree-pass.h (opt_pass::gate): Add function * argument.
3528 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3529 combine-stack-adj.c, combine.c, compare-elim.c,
3530 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3531 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3532 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3533 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3534 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3535 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3536 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3537 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3538 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3539 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3540 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3541 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3542 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3543 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3544 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3545 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3546 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3547 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3548 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3549 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3550 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3551 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3552 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3553 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3554 var-tracking.c, vtable-verify.c, web.c: Adjust.
3556 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3558 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3559 * configure: Regenerate.
3561 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3563 * passes.c (dump_one_pass): don't check pass->has_gate.
3564 (execute_ipa_summary_passes): Likewise.
3565 (execute_one_pass): Likewise.
3566 (ipa_write_summaries_2): Likewise.
3567 (ipa_write_optimization_summaries_1): Likewise.
3568 (ipa_read_optimization_summaries_1): Likewise.
3569 (execute_ipa_stmt_fixups): Likewise.
3570 * tree-pass.h (pass_data::has_gate): Remove.
3571 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3572 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3573 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3574 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3575 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3576 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3577 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3578 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3579 gimple-low.c, gimple-ssa-isolate-paths.c,
3580 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3581 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3582 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3583 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3584 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3585 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3586 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3587 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3588 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3589 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3590 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3591 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3592 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3593 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3594 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3595 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3596 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3597 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3598 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3599 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3600 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3601 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3604 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3606 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3608 * passes.c (pass_manager::register_dump_files_1): Merge into
3609 (pass_manager::register_dump_files): this, and remove its handling of
3610 properties since the pass always has the properties anyway.
3611 (pass_manager::pass_manager): Adjust.
3613 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3615 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3616 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3617 dealing with properties.
3618 (pass_manager::register_dump_files): Adjust.
3620 2014-03-20 Mark Wielaard <mjw@redhat.com>
3622 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3623 then represent the bound as normal constant value.
3625 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3628 Forward port from 4.8 branch
3629 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3631 * config/i386/bmiintrin.h (_blsi_u32): New.
3635 (_blsmsk_u32): Ditto.
3636 (_blsmsk_u64): Ditto.
3637 (_tzcnt_u32): Ditto.
3638 (_tzcnt_u64): Ditto.
3640 2014-04-17 Kito Cheng <kito@0xlab.org>
3642 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3644 2014-04-17 Richard Biener <rguenther@suse.de>
3647 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3648 boolean results for comparisons.
3650 2014-04-17 Richard Biener <rguenther@suse.de>
3652 PR tree-optimization/60836
3653 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3654 initial PHI args to be gimple values.
3656 2014-04-17 Richard Biener <rguenther@suse.de>
3658 PR tree-optimization/60841
3659 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3660 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3661 of stmts to SLP build.
3662 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3663 (vect_analyze_slp): Likewise.
3664 (vect_analyze_slp_instance): Likewise.
3665 (vect_build_slp_tree): Limit overall SLP tree growth.
3666 * tree-vectorizer.h (vect_analyze_data_refs,
3667 vect_analyze_slp): Adjust prototypes.
3669 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3671 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3674 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3676 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3677 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3678 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3679 for TARGET_SLOW_PSHUFB
3681 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3683 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3684 * config/i386/i386.c (intel_cost): Ditto.
3686 2014-04-17 Joey Ye <joey.ye@arm.com>
3688 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3690 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3692 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3695 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3697 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3698 (type_all_derivations_known_p): New predicate.
3699 (type_all_ctors_visible_p): New predicate.
3700 (type_possibly_instantiated_p): New predicate.
3701 (get_odr_type): Compute all_derivations_known.
3702 (dump_odr_type): Dump the flag.
3703 (maybe_record_type): Cleanup.
3704 (record_target_from_binfo): Add bases_to_consider array;
3705 record bases for types w/o instances and skip CXX destructor.
3706 (possible_polymorphic_call_targets_1): Add bases_to_consider
3707 and consider_construction parameters; check if type may have instance.
3708 (get_polymorphic_call_info): Set maybe_in_construction to true
3709 when we know nothing.
3710 (record_targets_from_bases): Skip CXX destructors; they are
3711 never called for types in construction.
3712 (possible_polymorphic_call_targets): Do not record target when
3713 type may not have instance.
3715 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3718 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3719 external aliases alive, too.
3721 2014-04-16 Andrew Pinski <apinski@cavium.com>
3723 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3726 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3728 * final.c (compute_alignments): Do not apply loop alignment to a block
3729 falling through to the exit.
3731 2014-04-16 Catherine Moore <clm@codesourcery.com>
3733 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3734 Adjust constraints for microMIPS store patterns.
3736 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3738 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3740 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3742 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3743 (append_use): Run at -O0.
3744 (append_vdef): Likewise.
3745 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3746 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3748 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3750 PR tree-optimization/60844
3751 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3752 (propagate_op_to_single_use, remove_visited_stmt_chain,
3753 linearize_expr, repropagate_negates, reassociate_bb): Use it
3754 instead of gsi_remove.
3756 2014-04-16 Martin Jambor <mjambor@suse.cz>
3758 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3759 ipa_transforms_to_apply.
3760 (cgraph_function_versioning): Assert that old_node has empty
3761 ipa_transforms_to_apply.
3762 * trans-mem.c (ipa_tm_create_version): Likewise.
3763 * tree-inline.c (tree_function_versioning): Do not duplicate
3764 ipa_transforms_to_apply.
3766 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3769 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3771 Pass necessary as flags on 64-bit Solaris/x86.
3772 Use lowercase relocs for x86_64-*-*.
3773 * configure: Regenerate.
3775 2014-04-15 Jan Hubicka <jh@suse.cz>
3777 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3778 (maybe_record_node, likely_target_p): Use it.
3780 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3783 Revert following patch
3785 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3788 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3789 software floating point or no floating point registers, do not
3790 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3791 in GPRs that occurs after we tested for GPRs that would never be
3794 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3795 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3796 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3797 specifically allow DDmode, since that does not use the SPE SIMD
3800 2014-03-21 Mark Wielaard <mjw@redhat.com>
3802 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3803 as unsigned or int depending on type and value used.
3805 2014-04-15 Richard Biener <rguenther@suse.de>
3807 PR rtl-optimization/56965
3808 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3809 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3811 * alias.c (true_dependence_1): Do not call
3812 nonoverlapping_component_refs_p.
3813 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3814 nonoverlapping_component_refs_p.
3815 (indirect_refs_may_alias_p): Likewise.
3817 2014-04-15 Teresa Johnson <tejohnson@google.com>
3819 * cfg.c (dump_bb_info): Fix flags check.
3820 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3822 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3824 PR rtl-optimization/60663
3825 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3828 2014-04-15 Richard Biener <rguenther@suse.de>
3830 * lto-streamer.h (LTO_major_version): Bump to 4.
3832 2014-04-15 Richard Biener <rguenther@suse.de>
3834 * common.opt (lto_partition_model): New enum.
3835 (flto-partition=): Merge separate options with a single with argument,
3836 add -flto-partition=one support.
3837 * flag-types.h (enum lto_partition_model): Declare.
3838 * opts.c (finish_options): Remove duplicate -flto-partition=
3840 * lto-wrapper.c (run_gcc): Adjust.
3842 2014-04-15 Richard Biener <rguenther@suse.de>
3844 * alias.c (ncr_compar): New function.
3845 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3847 2014-04-15 Richard Biener <rguenther@suse.de>
3849 * alias.c (record_component_aliases): Do not walk BINFOs.
3851 2014-04-15 Richard Biener <rguenther@suse.de>
3853 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3854 Add struct function argument and adjust.
3855 (find_func_aliases_for_call): Likewise.
3856 (find_func_aliases): Likewise.
3857 (find_func_clobbers): Likewise.
3858 (intra_create_variable_infos): Likewise.
3859 (compute_points_to_sets): Likewise.
3860 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3862 2014-04-15 Richard Biener <rguenther@suse.de>
3864 * tree.c (iterative_hash_expr): Use enum tree_code_class
3865 to store TREE_CODE_CLASS.
3866 (tree_block): Likewise.
3867 (tree_set_block): Likewise.
3868 * tree.h (fold_build_pointer_plus_loc): Use
3869 convert_to_ptrofftype_loc.
3871 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3874 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3877 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3879 * cfgloop.h (struct loop): Move force_vectorize down.
3880 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3881 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3882 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3883 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3884 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3885 * tree-core.h (enum annot_expr_kind): Add new kind values.
3886 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3887 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3889 * tree.def (ANNOTATE_EXPR): Tweak comment.
3891 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3893 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3896 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3898 * tree.h (TYPE_IDENTIFIER): Declare.
3899 * tree.c (subrange_type_for_debug_p): Use it.
3900 * godump.c (go_format_type): Likewise.
3901 * dwarf2out.c (is_cxx_auto, modified_type_die,
3902 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3903 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3905 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3908 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3910 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3912 * config/i386/i386.c (examine_argument): Return bool. Return true if
3913 parameter should be passed in memory.
3914 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3915 (construct_container): Update calls to examine_argument.
3916 (function_arg_advance_64): Ditto.
3917 (return_in_memory_32): Merge with ix86_return_in_memory.
3918 (return_in_memory_64): Ditto.
3919 (return_in_memory_ms_64): Ditto.
3921 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3923 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3924 * coverage.c (coverage_compute_profile_id): Handle externally visible
3927 2014-04-14 Martin Jambor <mjambor@suse.cz>
3929 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3930 DECL_DISREGARD_INLINE_LIMITS functions.
3932 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3935 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3937 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3940 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3941 optimize_insn_for_speed_p instead of
3942 optimize_function_for_speed_p.
3944 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3946 * doc/invoke.texi (free): Document AArch64.
3948 2014-04-14 Richard Biener <rguenther@suse.de>
3950 PR tree-optimization/60042
3951 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3952 (insert_into_preds_of_block): Do not prevent PHI insertion
3953 for REFERENCE exprs here ...
3954 (eliminate_dom_walker::before_dom_children): ... but prevent
3955 their use here under similar conditions when applied to the
3956 IL after PRE optimizations.
3958 2014-04-14 Richard Biener <rguenther@suse.de>
3960 * passes.def: Move early points-to after early SRA.
3962 2014-04-14 Richard Biener <rguenther@suse.de>
3964 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3965 check for which sign-changes we allow when forwarding
3966 a converted value into a switch.
3968 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3970 * stor-layout.c (place_field): Finalize non-constant offset for the
3973 2014-04-14 Richard Biener <rguenther@suse.de>
3975 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3977 (expand_switch_using_bit_tests_p): Likewise.
3978 (process_switch): Compute and pass on speed_p based on the
3980 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3981 optimize_bb_for_speed_p.
3983 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3985 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3986 * function.h (struct function): Rename has_force_vect_loops into
3987 has_force_vectorize_loops.
3988 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3989 (input_struct_function_base): Likewise.
3990 * lto-streamer-out.c (output_cfg): Likewise.
3991 (output_struct_function_base): Likewise.
3992 * omp-low.c (expand_omp_simd): Likewise.
3993 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3994 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3995 (version_loop_for_if_conversion): Likewise.
3996 (tree_if_conversion): Likewise.
3997 (main_tree_if_conversion): Likewise.
3998 (gate_tree_if_conversion): Likewise.
3999 * tree-inline.c (copy_loops): Likewise.
4000 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4001 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4002 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4003 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4004 * tree-vectorizer.c (vectorize_loops): Likewise.
4005 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4007 2014-04-14 Richard Biener <rguenther@suse.de>
4010 * lto-streamer-out.c (wrap_refs): New function.
4011 (lto_output): Wrap symbol references in global initializes in
4012 type-preserving MEM_REFs.
4014 2014-04-14 Christian Bruel <christian.bruel@st.com>
4016 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4018 2014-04-14 Christian Bruel <christian.bruel@st.com>
4020 * config/sh/sh.md (setmemqi): New expand pattern.
4021 * config/sh/sh.h (CLEAR_RATIO): Define.
4022 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4023 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4025 2014-04-14 Richard Biener <rguenther@suse.de>
4028 * fold-const.c (negate_expr_p): Don't negate directional rounding
4030 (fold_negate_expr): Likewise.
4032 2014-04-14 Richard Biener <rguenther@suse.de>
4034 PR tree-optimization/59817
4035 PR tree-optimization/60453
4036 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4037 recursion to catch all CHRECs in the scalar evolution and restrict
4038 the predicate for the remains appropriately.
4040 2014-04-12 Catherine Moore <clm@codesourcery.com>
4042 * config/mips/constraints.md: Add new register constraint "kb".
4043 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4044 (*movhi_internal): Likewise.
4045 (*movqi_internal): Likewise.
4046 * config/mips/mips.h (M16_STORE_REGS): New register class.
4047 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4048 (REG_CLASS_CONTENTS): Likewise.
4049 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4051 2014-04-11 Tobias Burnus <burnus@net-b.de>
4054 * doc/invoke.texi (-Wformat-signedness): Document it.
4055 (Wformat=2): Mention that this enables -Wformat-signedness.
4057 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4059 * common/config/epiphany/epiphany-common.c
4060 (epiphany_option_optimization_table): Enable section anchors by
4061 default at -O1 or higher.
4062 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4063 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4064 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4065 carries no extra cost.
4066 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4067 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4068 * config/epiphany/predicates.md (memclob_operand): New predicate.
4069 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4070 Use memclob_operand predicate and X constraint for operand 3.
4072 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4074 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4075 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4078 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4080 PR rtl-optimization/60651
4081 * mode-switching.c (optimize_mode_switching): Make sure to emit
4082 sets of a lower numbered entity before sets of a higher numbered
4083 entity to a mode of the same or lower priority.
4084 When creating a seginfo for a basic block that starts with a code
4085 label, move the insertion point past the code label.
4086 (new_seginfo): Document and enforce requirement that
4087 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4088 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4089 * doc/tm.texi: Regenerate.
4091 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4094 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4096 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4098 * BASE-VER: Set to 4.10.0.
4100 2014-04-11 Tobias Burnus <burnus@net-b.de>
4103 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4104 * doc/gcc.texi (Service): Update description in the @menu
4105 * doc/invoke.texi (Option Summary): Remove misplaced and
4108 2014-04-11 Steve Ellcey <sellcey@mips.com>
4109 Jakub Jelinek <jakub@redhat.com>
4112 * expr.c (convert_move): Use emit_store_flag_force instead of
4113 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4116 2014-04-11 Richard Biener <rguenther@suse.de>
4119 * varasm.c (assemble_alias): Avoid endless error reporting
4120 recursion by setting TREE_ASM_WRITTEN.
4122 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4124 * config/s390/s390.md: Add a splitter for NOT rtx.
4126 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4128 PR rtl-optimization/60663
4129 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4131 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4132 Jakub Jelinek <jakub@redhat.com>
4135 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4136 flag from decl_node to node.
4138 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4141 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4142 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4143 ameliorating the cases where it can be.
4145 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4148 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4150 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4151 (loadsync_<mode>): Change mode.
4152 (load_quadpti, store_quadpti): New.
4153 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4154 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4155 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4157 2014-04-09 Cong Hou <congh@google.com>
4160 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4163 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4165 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4166 instead of vnor to exploit possible fusion opportunity in the
4168 (altivec_expand_vec_perm_const_le): Likewise.
4170 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4172 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4173 (loadsync_<mode>): Change mode.
4174 (load_quadpti, store_quadpti): New.
4175 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4176 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4178 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4181 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4182 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4183 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4185 2014-04-08 Richard Biener <rguenther@suse.de>
4188 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4189 a 64bit widest int print double-int similar to on HWI64 hosts.
4191 2014-04-08 Richard Biener <rguenther@suse.de>
4193 PR tree-optimization/60785
4194 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4195 default defs properly.
4197 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4199 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4200 (Weffc++): Likewise.
4202 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4204 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4205 set completep to false rather than true.
4207 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4210 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4211 ARM_TARGET2_DWARF_FORMAT.
4213 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4216 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4217 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4220 2014-04-07 Richard Biener <rguenther@suse.de>
4222 PR tree-optimization/60766
4223 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4224 (may_eliminate_iv): Convert cand_value_at result to desired type.
4226 2014-04-07 Jason Merrill <jason@redhat.com>
4229 * common.opt (-fno-gnu-unique): Add.
4230 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4232 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4234 * haifa-sched.c: Fix outdated function reference and minor
4235 grammar errors in introductory comment.
4237 2014-04-07 Richard Biener <rguenther@suse.de>
4240 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4242 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4244 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4247 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4248 size accounting for thunks.
4249 (pa_asm_output_mi_thunk): Use final_start_function() and
4250 final_end_function() to output function start and end directives.
4252 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4254 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4255 device specific ISA/ feature information. Remove short_sp and
4256 errata_skip ds. Add avr_device_specific_features enum to have device
4258 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4259 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4260 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4261 updated device specific info.
4262 * config/avr/avr-mcus.def: Merge device specific details to
4263 dev_attribute field.
4264 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4266 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4267 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4268 assembler if RMW isa supported by current device.
4269 * config/avr/genmultilib.awk: Update as device info structure changed.
4270 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4272 2014-04-04 Cong Hou <congh@google.com>
4274 PR tree-optimization/60656
4275 * tree-vect-stmts.c (supportable_widening_operation):
4276 Fix a bug that elements in a vector with vect_used_by_reduction
4277 property are incorrectly reordered when the operation on it is not
4278 consistant with the one in reduction operation.
4280 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4282 PR rtl-optimization/60155
4283 * gcse.c (record_set_data): New function.
4284 (single_set_gcse): New function.
4285 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4286 (hoist_code): Likewise.
4287 (get_pressure_class_and_nregs): Likewise.
4289 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4291 * explow.c (probe_stack_range): Emit a final optimization blockage.
4293 2014-04-04 Anthony Green <green@moxielogic.com>
4295 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4298 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4301 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4302 flags are set only during streaming.
4303 * ipa.c (process_references, walk_polymorphic_call_targets,
4304 symtab_remove_unreachable_nodes): Drop bodies of always inline
4305 after early inlining.
4306 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4308 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4309 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4312 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4315 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4318 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4320 (cortex_a53_fdivd): Likewise.
4322 2014-04-04 Martin Jambor <mjambor@suse.cz>
4325 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4327 * cgraph.c (clone_of_p): Also return true if thunks match.
4328 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4329 cgraph_function_or_thunk_node and an obsolete comment.
4330 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4332 (build_function_decl_skip_args): Likewise.
4333 (set_new_clone_decl_and_node_flags): New function.
4334 (duplicate_thunk_for_node): Likewise.
4335 (redirect_edge_duplicating_thunks): Likewise.
4336 (cgraph_clone_node): New parameter args_to_skip, pass it to
4337 redirect_edge_duplicating_thunks which is called instead of
4338 cgraph_redirect_edge_callee.
4339 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4340 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4342 2014-04-04 Jeff Law <law@redhat.com>
4345 * config/arm/predicates.md (const_int_I_operand): New predicate.
4346 (const_int_M_operand): Similarly.
4347 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4349 (insv_t2, extv_reg, extzv_t2): Likewise.
4350 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4351 (pop_multiple_with_writeback_and_return): Likewise.
4352 (vfp_pop_multiple_with_writeback): Likewise
4354 2014-04-04 Richard Biener <rguenther@suse.de>
4357 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4358 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4360 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4361 * gimplify.c (gimple_add_tmp_var_fn): New function.
4362 * gimple-expr.h (create_tmp_reg_fn): Declare.
4363 * gimple-expr.c (create_tmp_reg_fn): New function.
4364 * gimple-low.c (record_vars_into): Don't change cfun.
4365 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4366 code generation without cfun.
4368 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4371 * Makefile.in (install-driver): Fix shell scripting.
4373 2014-04-03 Cong Hou <congh@google.com>
4375 PR tree-optimization/60505
4376 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4377 threshold of number of iterations below which no vectorization
4379 * tree-vect-loop.c (new_loop_vec_info):
4380 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4381 * tree-vect-loop.c (vect_analyze_loop_operations):
4382 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4383 * tree-vect-loop.c (vect_transform_loop):
4384 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4385 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4386 of iterations of the loop and see if we should build the epilogue.
4388 2014-04-03 Richard Biener <rguenther@suse.de>
4390 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4391 (streamer_tree_cache_create): Adjust.
4392 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4393 to allow optional nodes array.
4394 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4395 (streamer_tree_cache_append): Likewise.
4396 (streamer_tree_cache_create): Create nodes array optionally
4397 as specified by parameter.
4398 * lto-streamer-out.c (create_output_block): Avoid maintaining
4399 the node array in the writer cache.
4400 (DFS_write_tree): Remove assertion.
4401 (produce_asm_for_decls): Free the out decl state hash table early.
4402 * lto-streamer-in.c (lto_data_in_create): Adjust for
4403 streamer_tree_cache_create prototype change.
4405 2014-04-03 Richard Biener <rguenther@suse.de>
4407 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4408 set TREE_CHAIN to NULL_TREE.
4410 2014-04-03 Richard Biener <rguenther@suse.de>
4412 PR tree-optimization/60740
4413 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4414 over all GIMPLE_COND operands.
4416 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4418 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4419 (Weffc++): Remove Scott's numbering, merge lists and reference
4422 2014-04-03 Nick Clifton <nickc@redhat.com>
4424 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4427 2014-04-03 Martin Jambor <mjambor@suse.cz>
4429 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4430 mention gcc_unreachable before failing.
4431 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4434 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4437 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4438 inconsistent code and instead mark the context inconsistent.
4439 (possible_polymorphic_call_targets): For inconsistent contexts
4440 return empty complete list.
4442 2014-04-02 Anthony Green <green@moxielogic.com>
4444 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4445 (extendqisi2, extendhisi2): Define.
4446 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4447 (WCHAR_TYPE): Change to unsigned int.
4449 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4451 PR tree-optimization/60733
4452 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4453 insertion point for PHI candidates to be the end of the feeding
4454 block for the PHI argument.
4456 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4458 PR rtl-optimization/60650
4459 * lra-constraints.c (process_alt_operands): Decrease reject for
4460 earlyclobber matching.
4462 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4464 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4466 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4468 * config/spu/spu.c (pad_bb): Do not crash when the last
4469 insn is CODE_FOR_blockage.
4471 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4473 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4474 lies outside the target mode.
4476 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4479 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4480 software floating point or no floating point registers, do not
4481 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4482 in GPRs that occurs after we tested for GPRs that would never be
4485 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4486 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4487 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4488 specifically allow DDmode, since that does not use the SPE SIMD
4491 2014-04-02 Richard Biener <rguenther@suse.de>
4494 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4495 MODE_INTs. Properly use negv_optab.
4496 (expand_abs): Likewise.
4498 2014-04-02 Richard Biener <rguenther@suse.de>
4501 * Makefile.in (install-driver): Guard extra installs with special
4504 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4506 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4509 2014-04-01 Richard Henderson <rth@redhat.com>
4512 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4513 alternative enabled before register allocation.
4515 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4517 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4518 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4520 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4521 (nios2_got_address): Update nios2_large_got_address call site.
4522 (nios2_delegitimize_address): New function.
4523 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4524 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4525 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4527 2014-04-01 Martin Husemann <martin@duskware.de>
4529 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4532 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4534 PR rtl-optimization/60604
4535 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4536 check from register_operand.
4537 (register_operand): Redefine in terms of general_operand.
4538 (nonmemory_operand): Use register_operand for the non-constant cases.
4540 2014-04-01 Richard Biener <rguenther@suse.de>
4542 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4544 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4546 * doc/invoke.texi (mapp-regs): Clarify.
4548 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4550 * config/i386/avx512fintrin.h (__v32hi): Define type.
4551 (__v64qi): Likewise.
4552 (_mm512_set1_epi8): Define.
4553 (_mm512_set1_epi16): Define.
4554 (_mm512_set4_epi32): Define.
4555 (_mm512_set4_epi64): Define.
4556 (_mm512_set4_pd): Define.
4557 (_mm512_set4_ps): Define.
4558 (_mm512_setr4_epi64): Define.
4559 (_mm512_setr4_epi32): Define.
4560 (_mm512_setr4_pd): Define.
4561 (_mm512_setr4_ps): Define.
4562 (_mm512_setzero_epi32): Define.
4564 2014-03-31 Martin Jambor <mjambor@suse.cz>
4567 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4568 callsite_arguments_match_p. Updated all callers. Also check types of
4569 corresponding formal parameters and actual arguments.
4570 (not_all_callers_have_enough_arguments_p) Renamed to
4571 some_callers_have_mismatched_arguments_p.
4573 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4575 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4577 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4580 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4583 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4585 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4586 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4588 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4589 Use FMAMODE_NOVF512 mode iterator.
4590 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4591 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4592 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4593 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4595 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4596 Use VF_128_256 mode iterator.
4597 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4600 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4602 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4603 static chain if needed.
4605 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4608 * lra-constraints.c (index_part_to_reg): New.
4609 (process_address): Use it.
4611 2014-03-27 Jeff Law <law@redhat.com>
4612 Jakub Jelinek <jakub@redhat.com>
4615 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4616 gen_rtx_{PLUS,MULT} to build up the address expression.
4618 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4619 creating non-canonical RTL.
4621 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4624 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4625 functions; reorganize to make cheap checks first.
4626 (inline_small_functions): Do not estimate growth when dumping;
4628 * ipa-inline.h (inline_summary): Add min_size.
4629 (growth_likely_positive): New function.
4630 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4631 (set_cond_stmt_execution_predicate): Cleanup.
4632 (estimate_edge_size_and_time): Compute min_size.
4633 (estimate_calls_size_and_time): Likewise.
4634 (estimate_node_size_and_time): Likewise.
4635 (inline_update_overall_summary): Update min_size.
4636 (do_estimate_edge_time): Likewise.
4637 (do_estimate_edge_size): Update.
4638 (do_estimate_edge_hints): Update.
4639 (growth_likely_positive): New function.
4641 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4644 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4645 also if addr has VOIDmode.
4647 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4649 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4650 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4652 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4653 instructions as well as AdvancedSIMD loads.
4655 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4657 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4658 Use crypto_aese type.
4659 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4660 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4661 crypto_aese, crypto_aesmc. Move to types.md.
4662 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4664 * config/arm/iterators.md (crypto_type): Likewise.
4666 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4668 * cgraph.c: Include expr.h and tree-dfa.h.
4669 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4672 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4675 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4676 regs from checking multi-reg pseudos.
4678 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4680 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4682 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4684 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4685 if it would clobber the stack pointer, even temporarily.
4687 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4689 * mode-switching.c: Make small adjustments to the top comment.
4691 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4693 * config/rs6000/constraints.md (wD constraint): New constraint to
4694 match the constant integer to get the top DImode/DFmode out of a
4695 vector in a VSX register.
4697 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4698 match the constant integer to get the top DImode/DFmode out of a
4699 vector in a VSX register.
4701 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4704 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4707 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4708 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4710 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4711 Optimize vec_extract of 64-bit values, where the value being
4712 extracted is in the top word, where we can use scalar
4713 instructions. Add direct move and store support. Combine the big
4714 endian/little endian vector select load support into a single insn.
4715 (vsx_extract_<mode>_internal1): Likewise.
4716 (vsx_extract_<mode>_internal2): Likewise.
4717 (vsx_extract_<mode>_load): Likewise.
4718 (vsx_extract_<mode>_store): Likewise.
4719 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4720 combined into vsx_extract_<mode>_load.
4721 (vsx_extract_<mode>_one_le): Likewise.
4723 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4724 define the top 64-bit vector element.
4726 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4729 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4730 Document vec_vbpermq builtin.
4733 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4734 enable use of xxsldwi and xxpermdi builtin functions.
4735 (vec_xxpermdi): Likewise.
4737 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4738 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4740 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4742 PR rtl-optimization/60650
4743 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4745 (find_spills_for): New.
4746 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4747 Spill all pseudos on the second iteration.
4749 2014-03-27 Marek Polacek <polacek@redhat.com>
4752 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4755 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4757 * config/s390/s390.c (s390_can_use_return_insn): Check for
4758 call-saved FPRs on 31 bit.
4760 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4763 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4764 if they need regimplification, just drop them instead of
4765 calling gimple_regimplify_operands on them.
4767 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4770 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4771 (aarch64_frame_pointer_required): Adjust logic.
4772 (aarch64_can_eliminate): Adjust logic.
4773 (aarch64_override_options_after_change): Adjust logic.
4775 2014-03-27 Dehao Chen <dehao@google.com>
4777 * ipa-inline.c (early_inliner): Update node's inline info.
4779 2014-03-26 Dehao Chen <dehao@google.com>
4781 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4782 compiler inserted conditional jumps for NAN float check.
4784 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4786 * ubsan.h (ubsan_create_data): Change second argument's type
4787 to const location_t *.
4788 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4790 (ubsan_create_data): Change second argument to const location_t *PLOC.
4791 Create Loc field whenever PLOC is non-NULL.
4792 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4793 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4797 * real.c (real_to_integer2): Change type of low to UHWI.
4799 2014-03-26 Tobias Burnus <burnus@net-b.de>
4801 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4802 (CILK_SELF_SPECS): New define.
4803 (driver_self_specs): Use it.
4805 2014-03-26 Richard Biener <rguenther@suse.de>
4807 * tree-pretty-print.c (percent_K_format): Implement special
4808 case for LTO and its stripped down BLOCK tree.
4810 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4813 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4815 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4816 one range is range_int_cst_p, but not both, at least optimize
4817 addition/subtraction of 0 and multiplication by 0 or 1.
4818 * gimple-fold.c (gimple_fold_call): Fold
4819 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4820 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4821 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4823 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4825 PR rtl-optimization/60452
4826 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4827 <case REG>: Return 1 for invalid offsets from the frame pointer.
4829 2014-03-26 Marek Polacek <polacek@redhat.com>
4832 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4835 2014-03-26 Marek Polacek <polacek@redhat.com>
4838 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4841 2014-03-26 Marek Polacek <polacek@redhat.com>
4844 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4845 multiplication in unsigned type.
4847 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4849 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4851 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4853 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4855 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4858 * cif-code.def (UNREACHABLE) New code.
4859 * ipa-inline.c (inline_small_functions): Skip edges to
4860 __builtlin_unreachable.
4861 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4862 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4863 predicate to __bulitin_unreachable.
4864 (set_cond_stmt_execution_predicate): Fix issue when
4865 invert_tree_comparison returns ERROR_MARK.
4866 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4867 propagate to inline clones.
4868 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4870 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4871 * cgraphclones.c (cgraph_clone_node): If call destination is already
4872 ureachable, do not redirect it back.
4873 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4876 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4878 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4879 Do not modify inline clones.
4881 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4883 * config/i386/i386.md (general_sext_operand): New mode attr.
4884 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4885 don't generate (sign_extend (const_int)).
4886 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4887 operands[2]. Use We constraint instead of <i> and
4888 <general_sext_operand> predicate instead of <general_operand>.
4889 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4890 * config/i386/constraints.md (We): New constraint.
4891 * config/i386/predicates.md (x86_64_sext_operand,
4892 sext_operand): New predicates.
4894 2014-03-25 Martin Jambor <mjambor@suse.cz>
4897 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4898 inconsistent devirtualizations to __builtin_unreachable.
4900 2014-03-25 Marek Polacek <polacek@redhat.com>
4903 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4905 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4907 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4908 order of elements for big-endian.
4910 2014-03-25 Richard Biener <rguenther@suse.de>
4913 * gimplify-me.c (gimple_regimplify_operands): Update the
4916 2014-03-25 Martin Jambor <mjambor@suse.cz>
4919 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4920 (lto_output_varpool_node): Likewise.
4921 (input_overwrite_node): Likewise.
4922 (input_varpool_node): Likewise.
4924 2014-03-25 Richard Biener <rguenther@suse.de>
4926 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4927 (run_gcc): Likewise.
4929 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4931 * combine.c (simplify_compare_const): Add MODE argument.
4932 Handle mode_width 0 as very large mode_width.
4933 (try_combine, simplify_comparison): Adjust callers.
4935 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4936 type to avoid signed integer overflow.
4937 * explow.c (plus_constant): Likewise.
4939 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4941 * doc/generic.texi: Correct typos.
4943 2014-03-24 Tobias Burnus <burnus@net-b.de>
4945 * doc/invoke.texi (-flto): Expand section about
4946 using static libraries with LTO.
4948 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4950 PR rtl-optimization/60501
4951 * optabs.def (addptr3_optab): New optab.
4952 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4953 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4954 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4956 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4958 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4960 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4962 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4965 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4966 (_mm256_undefined_ps): Define.
4967 (_mm256_undefined_pd): Define.
4968 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4969 (_mm_undefined_pd): Define.
4970 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4971 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4972 (_mm512_undefined_ps): Define.
4973 (_mm512_undefined_pd): Define.
4974 Use _mm*_undefined_*.
4975 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4977 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4979 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4980 (lshr_simd): DI mode added.
4981 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4982 (aarch64_ushr_simddi): Likewise.
4983 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4984 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4985 (vshrd_n_u64): Likewise.
4987 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4989 * Makefile.in (s-macro_list): Depend on cc1.
4991 2014-03-23 Teresa Johnson <tejohnson@google.com>
4993 * ipa-utils.c (ipa_print_order): Use specified dump file.
4995 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4997 PR rtl-optimization/60601
4998 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5000 * gcc.c (eval_spec_function): Initialize save_growing_value.
5002 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5005 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5006 code == MINUS_EXPR, never swap op0 with op1.
5008 * toplev.c (init_local_tick): Avoid signed integer multiplication
5010 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5011 shift by first operand's bitsize.
5013 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5016 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5018 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5019 TARGET_ISA_64BIT_P(x).
5021 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5023 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5024 pattern for vector nor instead of subtract from splat(-1).
5025 (altivec_expand_vec_perm_const_le): Likewise.
5027 2014-03-21 Richard Henderson <rth@twiddle.net>
5030 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5031 related insns after epilogue_completed.
5033 2014-03-21 Martin Jambor <mjambor@suse.cz>
5036 * cgraph.h (symtab_node): New flag body_removed.
5037 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5038 when removing bodies.
5039 * symtab.c (dump_symtab_base): Dump body_removed flag.
5040 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5041 had their bodies removed.
5043 2014-03-21 Martin Jambor <mjambor@suse.cz>
5046 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5049 2014-03-21 Richard Biener <rguenther@suse.de>
5051 PR tree-optimization/60577
5052 * tree-core.h (struct tree_base): Document nothrow_flag use
5053 in VAR_DECL_NONALIASED.
5054 * tree.h (VAR_DECL_NONALIASED): New.
5055 (may_be_aliased): Adjust.
5056 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5058 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5060 * expr.c (expand_expr_real_1): Remove outdated comment.
5062 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5065 * ira.c (adjust_cleared_regs): Call copy_rtx on
5066 *reg_equiv[REGNO (loc)].src_p before passing it to
5067 simplify_replace_fn_rtx.
5070 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5071 into CONST, put pic register as first operand of PLUS. Use
5072 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5074 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5076 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5078 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5080 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5081 around for store forwarding issue in the FPU on the UT699.
5082 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5083 loads and operations if -mfix-ut699 is specified.
5084 (divtf3_hq): Tweak attribute.
5085 (sqrttf2_hq): Likewise.
5087 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5089 * calls.c (store_one_arg): Remove incorrect const qualification on the
5090 type of the temporary.
5091 * cfgexpand.c (expand_return): Likewise.
5092 * expr.c (expand_constructor): Likewise.
5093 (expand_expr_real_1): Likewise.
5095 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5097 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5100 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5103 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5105 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5107 * config/arm/aarch-common-protos.h
5108 (alu_cost_table): Fix spelling of "extend".
5109 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5111 2014-03-19 Richard Biener <rguenther@suse.de>
5114 * tree-core.h (tree_type_common): Re-order pointer members
5115 to reduce recursion depth during GC walks.
5117 2014-03-19 Marek Polacek <polacek@redhat.com>
5120 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5121 before accessing it.
5123 2014-03-19 Richard Biener <rguenther@suse.de>
5126 * lto-streamer-in.c (input_function): In WPA stage do not drop
5129 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5131 PR tree-optimization/60559
5132 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5133 with build_zero_cst assignment.
5135 2014-03-18 Kai Tietz <ktietz@redhat.com>
5137 PR rtl-optimization/56356
5138 * sdbout.c (sdbout_parms): Verify that parms'
5139 incoming argument is valid.
5140 (sdbout_reg_parms): Likewise.
5142 2014-03-18 Richard Henderson <rth@redhat.com>
5145 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5146 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5147 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5149 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5151 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5152 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5153 Italicize plugin event names in description. Explain that
5154 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5155 Remind that no GCC functions should be called after PLUGIN_FINISH.
5156 Explain what pragmas with expansion are.
5158 2014-03-18 Martin Liska <mliska@suse.cz>
5160 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5161 gimple call statement is update.
5162 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5163 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5165 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5168 * ubsan.c (ubsan_instrument_unreachable): Call
5169 initialize_sanitizer_builtins.
5170 (ubsan_pass): Likewise.
5173 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5174 varpool_finalize_decl instead of rest_of_decl_compilation.
5176 2014-03-18 Richard Biener <rguenther@suse.de>
5178 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5179 by using bitmap_and_compl instead of bitmap_and_compl_into.
5180 (df_rd_transfer_function): Likewise.
5182 2014-03-18 Richard Biener <rguenther@suse.de>
5184 * doc/lto.texi (fresolution): Fix typo.
5186 2014-03-18 Richard Biener <rguenther@suse.de>
5188 * doc/invoke.texi (flto): Update for changes in 4.9.
5190 2014-03-18 Richard Biener <rguenther@suse.de>
5192 * doc/loop.texi: Remove section on the removed lambda framework.
5193 Update loop docs with recent changes in preserving loop structure.
5195 2014-03-18 Richard Biener <rguenther@suse.de>
5197 * doc/lto.texi (-fresolution): Document.
5199 2014-03-18 Richard Biener <rguenther@suse.de>
5201 * doc/contrib.texi: Adjust my name.
5203 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5206 * internal-fn.c: Include diagnostic-core.h.
5207 (expand_BUILTIN_EXPECT): New function.
5208 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5209 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5210 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5211 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5213 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5214 Revert 3 argument __builtin_expect code.
5215 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5216 * gimple-fold.c (gimple_fold_call): Likewise.
5217 * tree.h (fold_builtin_expect): New prototype.
5218 * builtins.c (build_builtin_expect_predicate): Add predictor
5219 argument, if non-NULL, create 3 argument __builtin_expect.
5220 (fold_builtin_expect): No longer static. Add ARG2 argument,
5221 pass it through to build_builtin_expect_predicate.
5222 (fold_builtin_2): Adjust caller.
5223 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5224 * internal-fn.def (BUILTIN_EXPECT): New.
5226 2014-03-18 Tobias Burnus <burnus@net-b.de>
5229 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5230 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5231 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5233 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5236 * predict.c (combine_predictions_for_bb): Fix up formatting.
5237 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5238 fill what it points to if non-NULL.
5239 (tree_predict_by_opcode): Adjust caller, use the predictor.
5240 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5242 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5244 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5245 proper constant for the store mode.
5247 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5249 * symtab.c (change_decl_assembler_name): Fix transparent alias
5252 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5254 * config/aarch64/aarch64.c: Correct the comments about the
5255 aarch64 stack layout.
5257 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5259 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5260 check for GF_OMP_FOR_KIND_FOR.
5262 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5264 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5265 ymm and zmm register names.
5267 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5270 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5271 note creation for the 2010-08-31 changes.
5273 2014-03-17 Marek Polacek <polacek@redhat.com>
5276 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5277 as -fno-tree-loop-vectorize.
5278 (expand_omp_simd): Likewise.
5280 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5282 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5283 (eligible_for_call_delay): New prototype.
5284 * config/sparc/sparc.c (tls_call_delay): Rename into...
5285 (eligible_for_call_delay): ...this. Return false if the instruction
5286 cannot be put in the delay slot of a branch.
5287 (eligible_for_restore_insn): Simplify.
5288 (eligible_for_return_delay): Return false if the instruction cannot be
5289 put in the delay slot of a branch and simplify.
5290 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5291 put in the delay slot of a branch.
5292 * config/sparc/sparc.md (fix_ut699): New attribute.
5293 (tls_call_delay): Delete.
5294 (in_call_delay): Reimplement.
5295 (eligible_for_sibcall_delay): Rename into...
5296 (in_sibcall_delay): ...this.
5297 (eligible_for_return_delay): Rename into...
5298 (in_return_delay): ...this.
5299 (in_branch_delay): Reimplement.
5300 (in_uncond_branch_delay): Delete.
5301 (in_annul_branch_delay): Delete.
5303 2014-03-14 Richard Henderson <rth@redhat.com>
5306 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5307 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5308 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5309 (floathi splitters): Remove.
5310 (float<SWI48x>xf2): New pattern.
5311 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5312 code that tried to handle DImode for 32-bit, but which was excluded
5313 by the pattern's condition. Drop allocation of stack temporary.
5314 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5315 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5316 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5317 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5318 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5319 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5320 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5321 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5322 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5323 (*float<SWI48x><X87MODEF>2_i387): Remove.
5324 (all float _with_temp splitters): Remove.
5325 (*float<SWI48x><MODEF>2_i387): New pattern.
5326 (*float<SWI48><MODEF>2_sse): New pattern.
5327 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5328 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5330 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5331 Marek Polacek <polacek@redhat.com>
5334 * common.opt (dump_base_name_prefixed): New Variable.
5335 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5336 if x_dump_base_name_prefixed is already set, set it at the end.
5338 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5340 PR rtl-optimization/60508
5341 * lra-constraints.c (get_reload_reg): Add new parameter
5343 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5344 Pass the new parameter values.
5346 2014-03-14 Richard Biener <rguenther@suse.de>
5348 * common.opt: Revert unintented changes from r205065.
5351 2014-03-14 Richard Biener <rguenther@suse.de>
5354 * cfghooks.c (split_block): Properly adjust all loops the
5355 block was a latch of.
5357 2014-03-14 Martin Jambor <mjambor@suse.cz>
5360 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5363 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5366 * config/avr/avr.c (avr_set_current_function): Pass function name
5367 through default_strip_name_encoding before sanity checking instead
5368 of skipping the first char of the assembler name.
5370 2014-03-13 Richard Henderson <rth@redhat.com>
5373 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5374 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5375 * config/i386/i386-protos.h: Likewise.
5376 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5377 in the expander instead of a splitter.
5378 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5379 any possibility of requiring a memory.
5380 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5381 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5382 (fp branch splitters): Update for ix86_split_fp_branch.
5383 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5384 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5385 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5386 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5387 (*fop_<MODEF>_3_i387): Likewise.
5388 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5389 (splitters for the fop_* register patterns): Remove.
5390 (fscalexf4_i387): Rename from *fscalexf4_i387.
5391 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5393 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5395 PR tree-optimization/59779
5396 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5397 type for bitsize and maxsize instead of HOST_WIDE_INT.
5399 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5401 PR rtl-optimization/57320
5402 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5403 the CFG after thread_prologue_and_epilogue_insns.
5405 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5407 PR rtl-optimization/57189
5408 * lra-constraints.c (process_alt_operands): Disfavor spilling
5411 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5413 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5415 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5417 PR tree-optimization/59025
5419 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5420 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5422 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5425 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5426 calls of avr_out_plus_1.
5428 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5430 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5431 BB's single pred and update the father loop's latch info later.
5433 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5435 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5439 (VEC_base): Likewise.
5440 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5441 registers, we need to swap double words in little endian mode.
5443 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5444 to be a container mode for 128-bit integer operations added in ISA
5445 2.07. Unlike TImode and PTImode, the preferred register set is
5446 the Altivec/VMX registers for the 128-bit operations.
5448 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5450 (rs6000_split_128bit_ok_p): Likewise.
5452 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5453 macros for creating ISA 2.07 normal and overloaded builtin
5454 functions with 3 arguments.
5455 (BU_P8V_OVERLOAD_3): Likewise.
5456 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5457 for use as overloaded functions.
5458 (VPERM_1TI_UNS): Likewise.
5459 (VSEL_1TI): Likewise.
5460 (VSEL_1TI_UNS): Likewise.
5461 (ST_INTERNAL_1ti): Likewise.
5462 (LD_INTERNAL_1ti): Likewise.
5463 (XXSEL_1TI): Likewise.
5464 (XXSEL_1TI_UNS): Likewise.
5465 (VPERM_1TI): Likewise.
5466 (VPERM_1TI_UNS): Likewise.
5467 (XXPERMDI_1TI): Likewise.
5468 (SET_1TI): Likewise.
5469 (LXVD2X_V1TI): Likewise.
5470 (STXVD2X_V1TI): Likewise.
5471 (VEC_INIT_V1TI): Likewise.
5472 (VEC_SET_V1TI): Likewise.
5473 (VEC_EXT_V1TI): Likewise.
5474 (EQV_V1TI): Likewise.
5475 (NAND_V1TI): Likewise.
5476 (ORC_V1TI): Likewise.
5477 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5478 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5480 (VADDUQM): Likewise.
5481 (VSUBCUQ): Likewise.
5482 (VADDEUQM): Likewise.
5483 (VADDECUQ): Likewise.
5484 (VSUBEUQM): Likewise.
5485 (VSUBECUQ): Likewise.
5487 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5488 __int128_t and __uint128_t types.
5489 (__uint128_type): Likewise.
5490 (altivec_categorize_keyword): Add support for vector __int128_t,
5491 vector __uint128_t, vector __int128, and vector unsigned __int128
5492 as a container type for TImode operations that need to be done in
5493 VSX/Altivec registers.
5494 (rs6000_macro_to_expand): Likewise.
5495 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5496 to support 128-bit integer instructions vaddcuq, vadduqm,
5497 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5498 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5500 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5501 for V1TImode, and set up preferences to use VSX/Altivec registers.
5502 Setup VSX reload handlers.
5503 (rs6000_debug_reg_global): Likewise.
5504 (rs6000_init_hard_regno_mode_ok): Likewise.
5505 (rs6000_preferred_simd_mode): Likewise.
5506 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5507 (easy_altivec_constant): Likewise.
5508 (output_vec_const_move): Likewise.
5509 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5511 (rs6000_expand_vector_extract): Likewise.
5512 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5514 (rs6000_const_vec): Add support for V1TImode.
5515 (rs6000_emit_le_vsx_load): Swap double words when loading or
5516 storing TImode/V1TImode.
5517 (rs6000_emit_le_vsx_store): Likewise.
5518 (rs6000_emit_le_vsx_move): Likewise.
5519 (rs6000_emit_move): Add support for V1TImode.
5520 (altivec_expand_ld_builtin): Likewise.
5521 (altivec_expand_st_builtin): Likewise.
5522 (altivec_expand_vec_init_builtin): Likewise.
5523 (altivec_expand_builtin): Likewise.
5524 (rs6000_init_builtins): Add support for V1TImode type. Add
5525 support for ISA 2.07 128-bit integer builtins. Define type names
5526 for the VSX/Altivec vector types.
5527 (altivec_init_builtins): Add support for overloaded vector
5528 functions with V1TImode type.
5529 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5530 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5532 (rs6000_split_128bit_ok_p): Likewise.
5533 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5534 __int128_t and vector __uint128_t.
5536 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5537 and mode attributes.
5544 (VS_scalar): Likewise.
5545 (VS_double): Likewise.
5546 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5548 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5549 we support the ISA 2.07 128-bit integer arithmetic instructions.
5550 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5551 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5552 and TImode types for use with the builtin functions.
5553 (V1TI_type_node): Likewise.
5554 (unsigned_V1TI_type_node): Likewise.
5555 (intTI_type_internal_node): Likewise.
5556 (uintTI_type_internal_node): Likewise.
5558 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5559 128-bit builtin functions.
5560 (UNSPEC_VADDEUQM): Likewise.
5561 (UNSPEC_VADDECUQ): Likewise.
5562 (UNSPEC_VSUBCUQ): Likewise.
5563 (UNSPEC_VSUBEUQM): Likewise.
5564 (UNSPEC_VSUBECUQ): Likewise.
5565 (VM): Add V1TImode to vector mode iterators.
5567 (VI_unit): Likewise.
5568 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5569 (altivec_vaddcuq): Likewise.
5570 (altivec_vsubuqm): Likewise.
5571 (altivec_vsubcuq): Likewise.
5572 (altivec_vaddeuqm): Likewise.
5573 (altivec_vaddecuq): Likewise.
5574 (altivec_vsubeuqm): Likewise.
5575 (altivec_vsubecuq): Likewise.
5577 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5579 (BOOL_128): Likewise.
5580 (BOOL_REGS_OUTPUT): Likewise.
5581 (BOOL_REGS_OP1): Likewise.
5582 (BOOL_REGS_OP2): Likewise.
5583 (BOOL_REGS_UNARY): Likewise.
5584 (BOOL_REGS_AND_CR0): Likewise.
5586 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5587 128-bit integer builtin support.
5588 (vec_vadduqm): Likewise.
5589 (vec_vaddecuq): Likewise.
5590 (vec_vaddeuqm): Likewise.
5591 (vec_vsubecuq): Likewise.
5592 (vec_vsubeuqm): Likewise.
5593 (vec_vsubcuq): Likewise.
5594 (vec_vsubuqm): Likewise.
5596 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5597 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5598 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5599 128-bit integer add/subtract to ISA 2.07.
5601 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5603 * config/arc/arc.c (arc_predicate_delay_insns):
5604 Fix third argument passed to conditionalize_nonjump.
5606 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5608 * config/aarch64/aarch64-builtins.c
5609 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5610 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5611 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5612 instead of __builtin_lfloor.
5613 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5615 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5617 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5618 (tree_ssa_ifcombine_bb_1): New function.
5619 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5620 is an empty forwarder block to then_bb or vice versa and then_bb
5621 and else_bb are effectively swapped.
5623 2014-03-12 Christian Bruel <christian.bruel@st.com>
5626 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5627 REG_CFA_DEF_CFA note.
5628 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5629 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5631 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5633 PR tree-optimization/60454
5634 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5636 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5638 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5639 Do not define target_cpu_default2 to generic.
5640 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5641 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5642 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5644 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5645 Marc Glisse <marc.glisse@inria.fr>
5647 PR tree-optimization/60502
5648 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5649 instead of build_low_bits_mask.
5651 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5654 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5655 if there are multiple uses, but op doesn't live on E edge.
5656 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5657 clobber stmts before __builtin_unreachable.
5659 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5661 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5662 hard_frame_pointer_rtx.
5663 * cse.c (cse_insn): Remove volatile check.
5664 * cselib.c (cselib_process_insn): Likewise.
5665 * dse.c (scan_insn): Likewise.
5667 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5669 * config/arc/arc.c (conditionalize_nonjump): New function,
5671 (arc_ifcvt): ... this.
5672 (arc_predicate_delay_insns): Use it.
5674 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5676 * config/arc/predicates.md (extend_operand): During/after reload,
5677 allow const_int_operand.
5678 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5679 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5680 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5682 (umulsi3_highpart_i): Likewise.
5684 2014-03-11 Richard Biener <rguenther@suse.de>
5686 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5687 Add asserts to guard possible wrong-code bugs.
5689 2014-03-11 Richard Biener <rguenther@suse.de>
5691 PR tree-optimization/60429
5692 PR tree-optimization/60485
5693 * tree-ssa-structalias.c (set_union_with_increment): Properly
5694 take into account all fields that overlap the shifted vars.
5695 (do_sd_constraint): Likewise.
5696 (do_ds_constraint): Likewise.
5697 (get_constraint_for_ptr_offset): Likewise.
5699 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5701 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5702 (nios2_compute_frame_layout):
5703 Add calculation of cfun->machine->fp_save_offset.
5704 (nios2_expand_prologue): Correct setting of frame pointer register
5706 (nios2_expand_epilogue): Update recovery of stack pointer from
5707 frame pointer accordingly.
5708 (nios2_initial_elimination_offset): Update calculation of offset
5709 for eliminating to HARD_FRAME_POINTER_REGNUM.
5711 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5714 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5715 cgraph_get_create_node on VAR_DECLs.
5717 2014-03-10 Richard Biener <rguenther@suse.de>
5720 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5722 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5724 * config/vms/vms.opt (vms_float_format): New variable.
5726 2014-03-08 Tobias Burnus <burnus@net-b.de>
5728 * doc/invoke.texi (-fcilkplus): Update implementation status.
5730 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5731 Richard Biener <rguenther@suse.de>
5733 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5734 consistently accross all TUs.
5735 (run_gcc): Enable -fshort-double automatically at link at link-time
5736 and disallow override.
5738 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5741 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5742 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5743 if they can't be used.
5745 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5747 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5748 for Solaris 11/x86 ld.
5749 * configure: Regenerate.
5751 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5753 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5754 (LIB_TLS_SPEC): Save as ld_tls_libs.
5755 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5756 (HAVE_AS_IX86_TLSLDM): New test.
5757 * configure, config.in: Regenerate.
5758 * config/i386/i386.c (legitimize_tls_address): Fall back to
5759 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5760 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5761 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5762 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5764 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5766 * common.opt (fira-loop-pressure): Mark as optimization.
5768 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5770 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5771 an OpenMP mappable type.
5773 2014-03-06 Matthias Klose <doko@ubuntu.com>
5775 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5776 MULTILIB_OSDIRNAMES is not defined.
5778 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5779 Meador Inge <meadori@codesourcery.com>
5782 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5783 (arm_legitimize_address): Call legitimize_tls_address for any
5784 arm_tls_referenced_p expression, handle constant addend. Call it
5785 before testing for !TARGET_ARM.
5786 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5788 2014-03-06 Richard Biener <rguenther@suse.de>
5794 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5796 * tree-streamer.c (record_common_node): Assert we don't record
5797 nodes with type double.
5798 (preload_common_node): Skip type double, complex double and double
5799 pointer since it is now frontend dependent due to fshort-double option.
5801 2014-03-06 Richard Biener <rguenther@suse.de>
5803 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5804 or -fno-lto is specified and the linker has full plugin support.
5805 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5806 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5807 * lto-wrapper.c (merge_and_complain): Merge compile-time
5808 optimization levels.
5809 (run_gcc): And pass it through to the link options.
5811 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5815 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5817 * cselib.c (remove_useless_values): Skip to avoid quadratic
5818 behavior if the condition moved from...
5819 (cselib_process_insn): ... here holds.
5821 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5824 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5825 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5828 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5829 (TM_H): Add x86-tune.def.
5831 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5833 * config/aarch64/aarch64.c (generic_tunings):
5834 Use cortexa57_extra_costs.
5836 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5839 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5840 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5841 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5844 2014-03-04 Heiher <r@hev.cc>
5846 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5847 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5849 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5851 * config/i386/predicates.md (const2356_operand): Change to ...
5852 (const2367_operand): ... this.
5853 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5855 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5856 (*avx512pf_scatterpf<mode>sf): Ditto.
5857 (avx512pf_scatterpf<mode>df): Ditto.
5858 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5859 (*avx512pf_scatterpf<mode>df): Ditto.
5860 * config/i386/i386.c (ix86_expand_builtin): Update
5861 incorrect hint operand error message.
5863 2014-03-04 Richard Biener <rguenther@suse.de>
5865 * lto-section-in.c (lto_get_section_data): Fix const cast.
5867 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5869 * tree-streamer.c (record_common_node): Assert we don't record
5870 nodes with type double.
5871 (preload_common_node): Skip type double, complex double and double
5872 pointer since it is now frontend dependent due to fshort-double option.
5874 2014-03-04 Richard Biener <rguenther@suse.de>
5877 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5878 (lto_input_toplevel_asms): Likewise.
5879 * lto-section-in.c (lto_get_section_data): Instead do it here
5882 2014-03-04 Richard Biener <rguenther@suse.de>
5884 PR tree-optimization/60382
5885 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5886 dead PHIs a reduction.
5888 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5890 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5892 (_mm_prefetch): Move out of GCC target("sse") pragma.
5893 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5894 GCC target("prfchw") pragma.
5895 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5897 * config/i386/i386.c (ix86_option_override_internal): Enable
5898 -mprfchw with -mprefetchwt1.
5900 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5902 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5905 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5907 * opts.h (CL_PCH_IGNORE): Define.
5908 * targhooks.c (option_affects_pch_p):
5909 Return false for options that have CL_PCH_IGNORE set.
5910 * opt-functions.awk: Process PchIgnore.
5911 * doc/options.texi: Document PchIgnore.
5913 * config/arc/arc.opt (misize): Add PchIgnore property.
5915 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5917 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5918 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5919 constraint on constants to permit them being loaded into
5920 GENERAL_REGS or BASE_REGS.
5922 2014-03-03 Nick Clifton <nickc@redhat.com>
5924 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5925 anti-cacnonical alternatives.
5926 (negandhi3_real): New pattern.
5927 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5929 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5931 * config/avr/avr-mcus.def: Remove atxmega16x1.
5932 * config/avr/avr-tables.opt: Regenerate.
5933 * config/avr/t-multilib: Regenerate.
5934 * doc/avr-mmcu.texi: Regenerate.
5936 2014-03-03 Tobias Grosser <tobias@grosser.es>
5937 Mircea Namolaru <mircea.namolaru@inria.fr>
5939 PR tree-optimization/58028
5940 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5943 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5945 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5946 not handled by recognizers.
5948 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5951 * function.c (expand_function_end): Don't emit
5952 clobber_return_register sequence if clobber_after is a BARRIER.
5953 * cfgexpand.c (construct_exit_block): Append instructions before
5954 return_label to prev_bb.
5956 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5958 * config/rs6000/constraints.md: Document reserved use of "wc".
5960 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5963 * ipa.c (function_and_variable_visibility): When dissolving comdat
5964 group, also set all symbols to local.
5966 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5971 2013-12-14 Jan Hubicka <jh@suse.cz>
5973 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5975 2014-03-02 Jon Beniston <jon@beniston.com>
5981 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5982 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5983 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5984 (simple_return, *simple_return): New patterns
5985 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5986 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5988 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5990 * dwarf2out.c (gen_subprogram_die): Tidy.
5992 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5995 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5996 (*mov_t_msb_neg_negc): ... this new insn.
5998 2014-02-28 Jason Merrill <jason@redhat.com>
6001 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6004 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6007 * dwarf2out.c (decltype_auto_die): New static.
6008 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6009 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6010 (is_cxx_auto): Likewise.
6012 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6014 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6015 we are not using general regs only.
6017 2014-02-28 Richard Biener <rguenther@suse.de>
6020 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6021 previous fix and only allow to remove trivial pre-headers
6022 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6023 (remove_forwarder_block): Properly update the latch of a loop.
6025 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6028 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6029 (cselib_preserved_hash_table): New.
6030 (preserve_constants_and_equivs): Move preserved vals to it.
6031 (cselib_find_slot): Look it up first.
6032 (cselib_init): Initialize it.
6033 (cselib_finish): Release it.
6034 (dump_cselib_table): Dump it.
6036 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6039 * cselib.c (remove_useless_values): Skip to avoid quadratic
6040 behavior if the condition moved from...
6041 (cselib_process_insn): ... here holds.
6043 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6046 * var-tracking.c (vt_initialize): Apply the same condition to
6047 preserve the CFA base value.
6049 2014-02-28 Joey Ye <joey.ye@arm.com>
6052 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6053 if reload in progress or completed.
6055 2014-02-28 Tobias Burnus <burnus@net-b.de>
6058 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6061 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6063 * doc/tm.texi.in (Condition Code Status): Update documention for
6064 relative locations of cc0-setter and cc0-user.
6066 2014-02-27 Jeff Law <law@redhat.com>
6068 PR rtl-optimization/52714
6069 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6070 into two independent simple sets, if I3 is a jump, ensure the
6071 pattern we place into I3 is a (set (pc) ...).
6073 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6074 Jeff Law <law@redhat.com>
6076 PR rtl-optimization/49847
6077 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6078 are in different blocks.
6079 * doc/tm.texi (Condition Code Status): Update documention for
6080 relative locations of cc0-setter and cc0-user.
6082 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6085 * lra.c (lra_emit_add): Check SUBREG too.
6087 2014-02-27 Andreas Schwab <schwab@suse.de>
6089 * config/m68k/m68k.c (m68k_option_override): Disable
6090 -flive-range-shrinkage for classic m68k.
6091 (m68k_override_options_after_change): Likewise.
6093 2014-02-27 Marek Polacek <polacek@redhat.com>
6096 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6097 -Wmaybe-uninitialized.
6099 2014-02-27 Alan Modra <amodra@gmail.com>
6102 * reload1.c (emit_input_reload_insns): When reload_override_in,
6103 set old to rl->in_reg when rl->in_reg is a subreg.
6105 2014-02-26 Richard Biener <rguenther@suse.de>
6108 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6110 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6112 * common/config/i386/predicates.md (const1256_operand): Remove.
6113 (const2356_operand): New.
6114 (const_1_to_2_operand): Remove.
6115 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6116 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6117 (*avx512pf_gatherpf<mode>sf): Ditto.
6118 (avx512pf_gatherpf<mode>df): Ditto.
6119 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6120 (*avx512pf_gatherpf<mode>df): Ditto.
6121 (avx512pf_scatterpf<mode>sf): Ditto.
6122 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6123 (*avx512pf_scatterpf<mode>sf): Ditto.
6124 (avx512pf_scatterpf<mode>df): Ditto.
6125 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6126 (*avx512pf_scatterpf<mode>df): Ditto.
6127 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6129 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6131 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6132 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6133 (_mm512_mask_testn_epi64_mask): Move to ...
6134 * config/i386/avx512cdintrin.h: Here.
6135 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6136 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6137 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6138 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6139 TARGET_AVX512F from TARGET_AVX512CD.
6141 2014-02-26 Richard Biener <rguenther@suse.de>
6144 * ipa.c (walk_polymorphic_call_targets): Properly guard
6145 call to inline_update_overall_summary.
6147 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6150 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6151 and latches only if requested. Fix latch if it is removed.
6152 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6153 LOOPS_HAVE_PREHEADERS.
6155 2014-02-25 Andrew Pinski <apinski@cavium.com>
6157 * builtins.c (expand_builtin_thread_pointer): Create a new target
6158 when the target is NULL.
6160 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6162 PR rtl-optimization/60317
6163 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6164 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6165 * lra-assigns.c: Include params.h.
6166 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6167 other reload pseudos considerations.
6169 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6171 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6172 to use canonical form for nor<mode>3.
6174 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6177 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6180 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6182 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6183 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6184 (ix86_handle_option): Handle OPT_mprefetchwt1.
6185 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6186 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6188 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6189 OPTION_MASK_ISA_PREFETCHWT1.
6190 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6191 (PTA_PREFETCHWT1): New.
6192 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6193 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6194 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6195 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6196 (*prefetch_avx512pf_<mode>_: Change into ...
6197 (*prefetch_prefetchwt1_<mode>: This.
6198 * config/i386/i386.opt (mprefetchwt1): New.
6199 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6200 (_mm_prefetch): Handle intent to write.
6201 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6203 2014-02-25 Richard Biener <rguenther@suse.de>
6206 * emit-rtl.c (mem_attrs_htab): Remove.
6207 (mem_attrs_htab_hash): Likewise.
6208 (mem_attrs_htab_eq): Likewise.
6209 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6210 (init_emit_once): Do not allocate mem_attrs_htab.
6212 2014-02-25 Richard Biener <rguenther@suse.de>
6215 * lto-opts.c (lto_write_options): Output non-explicit conservative
6216 -fwrapv, -fno-trapv and -fno-strict-overflow.
6217 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6218 (run_gcc): And pass them through.
6220 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6222 * sel-sched.c (calculate_new_fences): New parameter ptime.
6223 Calculate it as a maximum over all fence cycles.
6224 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6225 Print the final schedule timing when sched_verbose.
6227 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6229 PR rtl-optimization/60292
6230 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6231 bit fot the fence instruction.
6233 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6235 * calls.h: Fix typo in comment.
6237 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6239 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6240 adjusting offsetable addresses.
6242 2014-02-24 Guozhi Wei <carrot@google.com>
6244 * sparseset.h (sparseset_pop): Fix the wrong index.
6246 2014-02-24 Walter Lee <walt@tilera.com>
6248 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6249 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6251 * common/config/tilegx/tilegx-common.c
6252 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6253 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6255 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6256 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6257 (tilegx_gimplify_va_arg_expr): Handle big endian.
6258 (tilegx_expand_unaligned_load): Ditto.
6259 (tilegx_expand_unaligned_store): Ditto.
6260 (TARGET_RETURN_IN_MSB): New.
6261 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6262 (TARGET_ENDIAN_DEFAULT): New.
6263 (TARGET_BIG_ENDIAN): Handle big endian.
6264 (BYTES_BIG_ENDIAN): Ditto.
6265 (WORDS_BIG_ENDIAN): Ditto.
6266 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6269 * config/tilegx/tilegx.md (extv): Handle big endian.
6271 (insn_st<n>): Ditto.
6272 (insn_st<n>_add<bitsuffix>): Ditto.
6273 (insn_stnt<n>): Ditto.
6274 (insn_stnt<n>_add<bitsuffix>):Ditto.
6275 (vec_interleave_highv8qi): Handle big endian.
6276 (vec_interleave_highv8qi_be): New.
6277 (vec_interleave_highv8qi_le): New.
6278 (insn_v1int_h): Handle big endian.
6279 (vec_interleave_lowv8qi): Handle big endian.
6280 (vec_interleave_lowv8qi_be): New.
6281 (vec_interleave_lowv8qi_le): New.
6282 (insn_v1int_l): Handle big endian.
6283 (vec_interleave_highv4hi): Handle big endian.
6284 (vec_interleave_highv4hi_be): New.
6285 (vec_interleave_highv4hi_le): New.
6286 (insn_v2int_h): Handle big endian.
6287 (vec_interleave_lowv4hi): Handle big endian.
6288 (vec_interleave_lowv4hi_be): New.
6289 (vec_interleave_lowv4hi_le): New.
6290 (insn_v2int_l): Handle big endian.
6291 (vec_interleave_highv2si): Handle big endian.
6292 (vec_interleave_highv2si_be): New.
6293 (vec_interleave_highv2si_le): New.
6294 (insn_v4int_h): Handle big endian.
6295 (vec_interleave_lowv2si): Handle big endian.
6296 (vec_interleave_lowv2si_be): New.
6297 (vec_interleave_lowv2si_le): New.
6298 (insn_v4int_l): Handle big endian.
6299 * config/tilegx/tilegx.opt (mbig-endian): New option.
6300 (mlittle-endian): New option.
6301 * doc/install.texi: Document tilegxbe-linux.
6302 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6304 2014-02-24 Martin Jambor <mjambor@suse.cz>
6307 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6308 there are no parameter descriptors.
6310 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6312 PR rtl-optimization/60268
6313 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6314 initialization to ...
6315 (sched_rgn_init): ... here.
6316 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6318 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6320 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6323 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6325 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6328 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6330 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6331 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6333 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6335 * config/microblaze/predicates.md: Add cmp_op predicate.
6336 * config/microblaze/microblaze.md: Add branch_compare instruction
6337 which uses cmp_op predicate and emits cmp insn before branch.
6338 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6339 to microblaze_expand_conditional_branch and consolidate logic.
6340 (microblaze_expand_conditional_branch): emit branch_compare
6341 insn instead of handling cmp op separate from branch insn.
6343 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6345 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6348 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6350 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6351 define_insn with define_expand and new define_insn
6352 *altivec_lve<VI_char>x_internal.
6353 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6354 and new define_insn *altivec_stve<VI_char>x_internal.
6355 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6357 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6359 (altivec_expand_stvex_be): New function.
6361 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6363 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6364 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6365 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6366 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6368 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6371 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6372 instead of emit_move_insn.
6374 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6376 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6378 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6379 gen_altivec_vsumsws.
6381 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6383 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6384 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6385 (altivec_lvxl_<mode>): New define_expand incorporating
6386 -maltivec=be semantics where needed.
6387 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6388 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6389 semantics where needed.
6390 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6391 (altivec_stvx_<mode>): New define_expand incorporating
6392 -maltivec=be semantics where needed.
6393 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6394 VM2 iterator instead of V4SI.
6395 (altivec_stvxl_<mode>): New define_expand incorporating
6396 -maltivec=be semantics where needed.
6397 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6398 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6399 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6400 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6401 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6402 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6403 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6404 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6405 ALTIVEC_BUILTIN_STVXL.
6406 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6407 (altivec_expand_stvx_be): Likewise.
6408 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6409 (altivec_expand_lvx_be): Likewise.
6410 (altivec_expand_stvx_be): Likewise.
6411 (altivec_expand_builtin): Add cases for
6412 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6413 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6414 (altivec_init_builtins): Add definitions for
6415 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6416 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6418 2014-02-21 Catherine Moore <clm@codesourcery.com>
6420 * doc/invoke.texi (mvirt, mno-virt): Document.
6421 * config/mips/mips.opt (mvirt): New option.
6422 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6424 2014-02-21 Richard Biener <rguenther@suse.de>
6426 PR tree-optimization/60276
6427 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6428 (STMT_VINFO_MIN_NEG_DIST): New macro.
6429 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6430 STMT_VINFO_MIN_NEG_DIST.
6431 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6432 made for negative dependence distances still hold.
6434 2014-02-21 Richard Biener <rguenther@suse.de>
6437 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6438 DECL_INITIAL for globals not in the current function context.
6440 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6442 PR tree-optimization/56490
6443 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6444 * tree-ssa-uninit.c: Include params.h.
6445 (compute_control_dep_chain): Add num_calls argument, return false
6446 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6447 num_calls to recursive call.
6448 (find_predicates): Change dep_chain into normal array,
6449 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6450 variable and adjust compute_control_dep_chain caller.
6451 (find_def_preds): Likewise.
6453 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6455 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6456 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6458 2014-02-21 Nick Clifton <nickc@redhat.com>
6460 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6461 (pushhi1): Likewise.
6462 (popqi1): Add mode to pre_dec.
6465 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6467 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6468 mode for mask of V8SFmode permutation.
6470 2014-02-20 Richard Henderson <rth@redhat.com>
6473 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6474 a new pseudo for OLDVAL.
6476 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6479 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6480 gen_reg_rtx if d->testing_p.
6481 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6482 if d->testing_p and we will certainly return true.
6483 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6486 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6488 * emit-rtl.c (gen_reg_rtx): Assert that
6489 crtl->emit.regno_pointer_align_length is non-zero.
6491 2014-02-20 Richard Henderson <rth@redhat.com>
6494 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6495 on failure the store back into EXPECT.
6497 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6498 Sandra Loosemore <sandra@codesourcery.com>
6500 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6501 * config/nios2/nios2.c (nios2_function_profiler): Add
6502 -fPIC (flag_pic == 2) support.
6503 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6504 (nios2_large_offset_p): New function.
6505 (nios2_unspec_reloc_p): Move up position, update to use
6506 nios2_large_offset_p.
6507 (nios2_unspec_address): Remove function.
6508 (nios2_unspec_offset): New function.
6509 (nios2_large_got_address): New function.
6510 (nios2_got_address): Add large offset support.
6511 (nios2_legitimize_tls_address): Update usage of removed and new
6513 (nios2_symbol_binds_local_p): New function.
6514 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6515 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6516 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6517 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6518 processing for (const (unspec ...)).
6519 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6521 2014-02-20 Richard Biener <rguenther@suse.de>
6523 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6524 doing the substitution.
6525 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6527 2014-02-20 Martin Jambor <mjambor@suse.cz>
6530 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6531 info when checking whether lattices are bottom.
6533 2014-02-20 Richard Biener <rguenther@suse.de>
6536 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6539 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6542 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6543 parameter specifying the scaling.
6544 (inline_call): Update.
6545 (want_inline_recursively): Guard division by zero.
6546 (recursive_inlining): Update.
6547 * ipa-inline.h (clone_inlined_nodes): Update.
6549 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6552 * config/i386/i386.c (classify_argument): Pass structures of size
6553 64 bytes or less in register.
6555 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6556 Kirill Yukhin <kirill.yukhin@intel.com>
6558 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6559 (_mm_rcp28_round_ss): Ditto.
6560 (_mm_rsqrt28_round_sd): Ditto.
6561 (_mm_rsqrt28_round_ss): Ditto.
6562 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6563 (_mm_rcp14_round_ss): Ditto.
6564 (_mm_rsqrt14_round_sd): Ditto.
6565 (_mm_rsqrt14_round_ss): Ditto.
6566 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6567 the first input operand, get rid of match_dup.
6568 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6570 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6572 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6573 operand as the first input operand, set type attribute.
6574 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6576 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6577 operand as the first input operand, set type attribute.
6579 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6581 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6584 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6587 * config/i386/i386.c (construct_container): Remove TFmode check
6588 for X86_64_INTEGER_CLASS.
6590 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6593 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6594 only when -Wpsabi is enabled.
6596 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6599 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6600 passing arrays in registers are the same as for structs, so remove the
6601 special case for them.
6603 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6605 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6606 destination type, extract only the valid bits if the source type is not
6607 integral and has a different mode.
6609 2014-02-19 Richard Biener <rguenther@suse.de>
6612 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6615 2014-02-19 Richard Biener <rguenther@suse.de>
6618 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6619 (ipa_modify_call_arguments): Emit an argument load explicitely and
6620 preserve virtual SSA form there and for the replacement call.
6621 Do not update SSA form nor free dominance info.
6623 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6625 * ipa.c (function_and_variable_visibility): Also clear WEAK
6626 flag when disolving COMDAT_GROUP.
6628 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6630 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6631 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6633 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6634 do more sanity checks.
6635 (detect_type_change): Return true when giving up early.
6636 (compute_complex_assign_jump_func): Fix type parameter of
6637 ipa_set_ancestor_jf.
6638 (compute_complex_ancestor_jump_func): Likewise.
6639 (update_jump_functions_after_inlining): Fix updating of
6641 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6643 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6645 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6646 inline clones when edge disappears.
6648 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6651 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6652 Split 64-bit moves into 2 patterns. Do not allow the use of
6653 direct move for TDmode in little endian, since the decimal value
6654 has little endian bytes within a word, but the 64-bit pieces are
6655 ordered in a big endian fashion, and normal subreg's of TDmode are
6657 (mov<mode>_64bit_dm): Likewise.
6658 (movtd_64bit_nodm): Likewise.
6660 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6662 PR tree-optimization/60174
6663 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6664 statement of an SSA_NAME that occurs in an abnormal PHI node.
6666 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6669 * final.c (SEEN_BB): Remove.
6670 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6671 (final_scan_insn): Don't force_source_line on second
6672 NOTE_INSN_BASIC_BLOCK.
6674 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6677 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6678 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6679 (type_natural_mode): Warn ABI change when %zmm register is not
6680 available for AVX512F vector value passing.
6682 2014-02-18 Kai Tietz <ktietz@redhat.com>
6685 * config/i386/i386.c (ix86_expand_prologue): Use value in
6686 rax register as displacement when restoring %r10 or %rax.
6687 Fix wrong offset when restoring both registers.
6689 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6691 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6692 assertion with conditional return.
6694 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6695 Uros Bizjak <ubizjak@gmail.com>
6698 * config/i386/driver-i386.c (host_detect_local_cpu): If
6699 YMM state is not saved by the OS, also clear has_f16c. Move
6700 CPUID 0x80000001 handling before YMM state saving checking.
6702 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6704 PR rtl-optimization/58960
6705 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6706 factored out from ...
6707 (sched_init): ... here.
6708 (free_global_sched_pressure_data): New, factored out from ...
6709 (sched_finish): ... here.
6710 * sched-int.h (free_global_sched_pressure_data): Declare.
6711 * sched-rgn.c (nr_regions_initial): New static global.
6712 (haifa_find_rgns): Initialize it.
6713 (schedule_region): Disable sched-pressure for the newly
6716 2014-02-17 Richard Biener <rguenther@suse.de>
6718 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6719 release SSA defs of pattern stmts.
6721 2014-02-17 Richard Biener <rguenther@suse.de>
6723 * tree-inline.c (expand_call_inline): Release the virtual
6724 operand defined by the call we are about to inline.
6726 2014-02-17 Richard Biener <rguenther@suse.de>
6728 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6730 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6731 Ilya Tocar <ilya.tocar@intel.com>
6733 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6734 arguments order in builtin.
6735 (_mm512_permutexvar_epi64): Ditto.
6736 (_mm512_mask_permutexvar_epi64): Ditto
6737 (_mm512_maskz_permutexvar_epi32): Ditto
6738 (_mm512_permutexvar_epi32): Ditto
6739 (_mm512_mask_permutexvar_epi32): Ditto
6741 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6743 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6744 (p8_vmrgow): Likewise.
6746 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6748 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6751 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6754 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6755 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6756 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6757 using direct move instructions on ISA 2.07. Also adjust
6758 instruction length for 64-bit.
6759 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6760 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6762 2014-02-15 Alan Modra <amodra@gmail.com>
6766 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6767 find_replacement on parts of insn rtl that might be reloaded.
6769 2014-02-15 Richard Biener <rguenther@suse.de>
6771 PR tree-optimization/60183
6772 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6773 (tree_ssa_phiprop): Calculate and free post-dominators.
6775 2014-02-14 Jeff Law <law@redhat.com>
6777 PR rtl-optimization/60131
6778 * ree.c (get_extended_src_reg): New function.
6779 (combine_reaching_defs): Use it rather than assuming location of REG.
6780 (find_and_remove_re): Verify first operand of extension is
6781 a REG before adding the insns to the copy list.
6783 2014-02-14 Roland McGrath <mcgrathr@google.com>
6785 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6786 * configure: Regenerated.
6787 * config.in: Regenerated.
6788 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6789 instead of ASM_SHORT.
6791 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6792 Richard Earnshaw <rearnsha@arm.com>
6794 PR rtl-optimization/59535
6795 * lra-constraints.c (process_alt_operands): Encourage alternative
6796 when unassigned pseudo class is superset of the alternative class.
6797 (inherit_reload_reg): Don't inherit when optimizing for code size.
6798 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6799 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6800 modes not less than 4 for Thumb1.
6802 2014-02-14 Kyle McMartin <kyle@redhat.com>
6805 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6807 2014-02-14 Richard Biener <rguenther@suse.de>
6809 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6810 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6811 Do not strip INDIRECT_REFs.
6813 2014-02-14 Richard Biener <rguenther@suse.de>
6816 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6817 DECL_FUNCTION_SPECIFIC_TARGET.
6818 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6819 * tree-streamer-out.c (pack_ts_target_option): Remove.
6820 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6821 (write_ts_function_decl_tree_pointers): Do not stream
6822 DECL_FUNCTION_SPECIFIC_TARGET.
6823 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6824 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6825 (lto_input_ts_function_decl_tree_pointers): Do not stream
6826 DECL_FUNCTION_SPECIFIC_TARGET.
6828 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6830 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6831 (get_initial_def_for_induction, vectorizable_induction): Ignore
6832 debug stmts when looking for exit_phi.
6833 (vectorizable_live_operation): Fix up condition.
6835 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6837 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6838 nreverse() because it changes the content of original tree list.
6840 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6842 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6843 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6845 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6847 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6848 GNU coding standards.
6850 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6853 * dwarf2out.c (gen_subprogram_die): Don't call
6854 add_calling_convention_attribute if subr_die is old_die.
6856 2014-02-13 Sharad Singhai <singhai@google.com>
6858 * doc/optinfo.texi: Fix order of nodes.
6860 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6862 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6863 operands[2], not operands[3].
6865 2014-02-13 Richard Biener <rguenther@suse.de>
6868 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6869 mention the possibility of an in-tree build.
6870 (CLooG): Update recommended version to 0.18.1, mention the
6871 possibility of an in-tree build and clarify that the ISL
6872 bundled with CLooG does not work.
6874 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6877 * expr.c (compress_float_constant): If x is a hard register,
6878 extend into a pseudo and then move to x.
6880 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6882 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6883 caused by bad second argument to warning_at() with -mhotpatch and
6884 nested functions (e.g. with gfortran).
6886 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6888 * opts.c (option_name): Remove "enabled by default" rider.
6890 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6892 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6894 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6895 Uros Bizjak <ubizjak@gmail.com>
6898 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6899 * configure: Regenerated.
6901 2014-02-12 Richard Biener <rguenther@suse.de>
6903 * vec.c (vec_prefix::calculate_allocation): Move as
6904 inline variant to vec.h.
6905 (vec_prefix::calculate_allocation_1): New out-of-line version.
6906 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6907 (vec_prefix::m_has_auto_buf): Rename to ...
6908 (vec_prefix::m_using_auto_storage): ... this.
6909 (vec_prefix::calculate_allocation): Inline the easy cases
6910 and dispatch to calculate_allocation_1 which doesn't need the
6912 (va_heap::reserve): Use gcc_checking_assert.
6913 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6914 m_using_auto_storage.
6915 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6917 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6918 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6919 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6921 2014-02-12 Richard Biener <rguenther@suse.de>
6923 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6924 when we found a dependence.
6926 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6928 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6930 (maybe_fold_stmt): ... into this new function.
6931 * omp-low.c (lower_omp): Update comment.
6933 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6936 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6939 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6941 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6942 identifiers in comments.
6943 (cortexa53_extra_costs): Likewise.
6944 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6945 (cortexa7_extra_costs): Likewise.
6946 (cortexa12_extra_costs): Likewise.
6947 (cortexa15_extra_costs): Likewise.
6948 (v7m_extra_costs): Likewise.
6950 2014-02-12 Richard Biener <rguenther@suse.de>
6953 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6954 of posix_memalign being successful.
6955 (lower_stmt): Restrict lowering of posix_memalign to when
6956 -ftree-bit-ccp is enabled.
6958 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6960 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6962 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6964 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6966 PR rtl-optimization/60116
6967 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6968 other_insn once the combination has been validated.
6970 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6973 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6975 * ipa-devirt.c: Include demangle.h
6976 (odr_violation_reported): New static variable.
6977 (add_type_duplicate): Update odr_violations.
6978 (maybe_record_node): Add completep parameter; update it.
6979 (record_target_from_binfo): Add COMPLETEP parameter;
6980 update it as needed.
6981 (possible_polymorphic_call_targets_1): Likewise.
6982 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6983 rename FINAL to COMPLETE.
6984 (record_targets_from_bases): Sanity check we found the binfo;
6985 fix COMPLETEP updating.
6986 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6987 parameter, fix computing of COMPLETEP.
6988 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6989 at LTO time do demangling.
6990 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6991 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6993 (gimple_get_virt_method_for_binfo): Likewise.
6994 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6995 gimple_get_virt_method_for_vtable): Update prototypes.
6997 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7000 * genautomata.c (add_presence_absence): Fix typo with
7001 {final_}presence_list.
7003 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7006 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7007 for VSX/Altivec vectors that land in GPR registers.
7009 2014-02-11 Richard Henderson <rth@redhat.com>
7010 Jakub Jelinek <jakub@redhat.com>
7013 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7014 around drhs if type conversion to lacc->type is not useless.
7016 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7018 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7020 (cortex-a57.cortex-a53): Likewise.
7021 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7023 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7025 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7028 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7030 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7031 add_options_for_arm_vfp3.
7033 2014-02-11 Jeff Law <law@redhat.com>
7036 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7037 object with an undesirable mode.
7039 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7042 * config/i386/sol2-9.h: New file.
7043 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7044 *-*-solaris2.9*): Use it.
7046 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7048 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7049 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7051 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7053 * config/microblaze/microblaze.c: Extend mcpu version format
7055 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7057 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7059 2014-02-10 Richard Henderson <rth@redhat.com>
7062 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7063 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7064 ms-abi vs -mno-accumulate-outgoing-args.
7065 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7066 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7069 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7072 * cfgexpand.c (expand_asm_operands): Attach source location to
7073 ASM_INPUT rtx objects.
7074 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7076 2014-02-10 Nick Clifton <nickc@redhat.com>
7078 * config/mn10300/mn10300.c (popcount): New function.
7079 (mn10300_expand_prologue): Include saved registers in stack usage
7082 2014-02-10 Jeff Law <law@redhat.com>
7085 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7086 when changing the SET_DEST of a prior insn to avoid an input reload.
7088 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7090 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7091 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7092 -mcall-openbsd, or -mcall-linux.
7093 (CC1_ENDIAN_BIG_SPEC): Remove.
7094 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7095 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7096 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7097 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7098 and %cc1_endian_default.
7099 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7101 2014-02-10 Richard Biener <rguenther@suse.de>
7103 PR tree-optimization/60115
7104 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7105 MEM_REF handling. Properly verify that the accesses are not
7106 out of the objects bound.
7108 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7110 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7113 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7115 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7116 proper constants and fix formatting.
7117 (possible_polymorphic_call_targets): Fix formatting.
7119 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7120 Ilya Tocar <ilya.tocar@intel.com>
7122 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7123 (_mm512_loadu_epi32): Renamed into...
7124 (_mm512_loadu_si512): This.
7125 (_mm512_storeu_epi32): Renamed into...
7126 (_mm512_storeu_si512): This.
7127 (_mm512_maskz_ceil_ps): Removed.
7128 (_mm512_maskz_ceil_pd): Ditto.
7129 (_mm512_maskz_floor_ps): Ditto.
7130 (_mm512_maskz_floor_pd): Ditto.
7131 (_mm512_floor_round_ps): Ditto.
7132 (_mm512_floor_round_pd): Ditto.
7133 (_mm512_ceil_round_ps): Ditto.
7134 (_mm512_ceil_round_pd): Ditto.
7135 (_mm512_mask_floor_round_ps): Ditto.
7136 (_mm512_mask_floor_round_pd): Ditto.
7137 (_mm512_mask_ceil_round_ps): Ditto.
7138 (_mm512_mask_ceil_round_pd): Ditto.
7139 (_mm512_maskz_floor_round_ps): Ditto.
7140 (_mm512_maskz_floor_round_pd): Ditto.
7141 (_mm512_maskz_ceil_round_ps): Ditto.
7142 (_mm512_maskz_ceil_round_pd): Ditto.
7143 (_mm512_expand_pd): Ditto.
7144 (_mm512_expand_ps): Ditto.
7145 * config/i386/i386.c (ix86_builtins): Remove
7146 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7147 (bdesc_args): Ditto.
7148 * config/i386/predicates.md (const1256_operand): New.
7149 (const_1_to_2_operand): Ditto.
7150 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7151 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7152 (*avx512pf_gatherpf<mode>sf): Ditto.
7153 (avx512pf_gatherpf<mode>df): Ditto.
7154 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7155 (*avx512pf_gatherpf<mode>df): Ditto.
7156 (avx512pf_scatterpf<mode>sf): Ditto.
7157 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7158 (*avx512pf_scatterpf<mode>sf): Ditto.
7159 (avx512pf_scatterpf<mode>df): Ditto.
7160 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7161 (*avx512pf_scatterpf<mode>df): Ditto.
7162 (avx512f_expand<mode>): Removed.
7163 (<shift_insn><mode>3<mask_name>): Change predicate type.
7165 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7167 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7168 not at the end of datarefs vector use ordered_remove to avoid
7169 reordering datarefs vector.
7172 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7173 mark local addressable non-static vars as GOVD_PRIVATE
7174 instead of GOVD_LOCAL.
7175 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7176 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7180 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7181 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7182 assume_aligned or alloc_align attributes.
7183 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7184 arguments. Handle also assume_aligned and alloc_align attributes.
7185 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7186 calls to functions with assume_aligned or alloc_align attributes.
7187 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7189 2014-02-08 Terry Guo <terry.guo@arm.com>
7191 * doc/invoke.texi: Document ARM -march=armv7e-m.
7193 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7195 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7196 flag on __cilkrts_rethrow builtin.
7199 * ipa-cp.c (determine_versionability): Fail at -O0
7200 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7201 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7204 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7207 * tree-inline.c (copy_forbidden): Fail for
7208 __attribute__((optimize (0))) functions.
7210 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7212 * varpool.c: Include pointer-set.h.
7213 (varpool_remove_unreferenced_decls): Variables in other partitions
7214 will not be output; be however careful to not lose information
7217 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7219 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7220 lookup in the vtable constructor.
7222 2014-02-07 Jeff Law <law@redhat.com>
7225 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7226 define_insn_and_split.
7228 * ipa-inline.c (inline_small_functions): Fix typos.
7230 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7232 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7233 (s390_can_use_return_insn): Declare.
7234 * config/s390/s390.h (EPILOGUE_USES): Define.
7235 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7237 (s390_chunkify_start): Handle return JUMP_LABELs.
7238 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7239 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7240 (s390_can_use_return_insn): New functions.
7241 (s390_fix_long_loop_prediction): Handle conditional returns.
7242 (TARGET_SET_UP_BY_PROLOGUE): Define.
7243 * config/s390/s390.md (ANY_RETURN): New code iterator.
7244 (*creturn, *csimple_return, return, simple_return): New patterns.
7246 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7248 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7249 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7250 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7251 REG_CFA_RESTORE list when deciding not to restore a register.
7253 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7255 * config/s390/s390.c: Include tree-pass.h and context.h.
7256 (s390_early_mach): New function, split out from...
7257 (s390_emit_prologue): ...here.
7258 (pass_data_s390_early_mach): New pass structure.
7259 (pass_s390_early_mach): New class.
7260 (s390_option_override): Create and register early_mach pass.
7261 Move to end of file.
7263 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7265 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7266 to match for the exit block.
7268 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7270 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7271 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7272 Reject misaligned operands.
7274 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7276 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7278 2014-02-07 Richard Biener <rguenther@suse.de>
7281 * gimple-low.c (lower_builtin_posix_memalign): New function.
7282 (lower_stmt): Call it to lower posix_memalign in a way
7283 to make alignment info accessible.
7285 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7288 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7289 __builtin_setjmp_receiver.
7291 2014-02-07 Richard Biener <rguenther@suse.de>
7294 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7295 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7296 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7297 Handle BUILT_IN_POSIX_MEMALIGN.
7298 (find_func_clobbers): Likewise.
7299 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7300 (call_may_clobber_ref_p_1): Likewise.
7302 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7305 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7308 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7311 * lto-cgraph.c (lto_output_node): Use
7312 symtab_get_symbol_partitioning_class.
7313 (lto_output_varpool_node): likewise.
7314 (symtab_get_symbol_partitioning_class): Move here from
7316 * cgraph.h (symbol_partitioning_class): Likewise.
7317 (symtab_get_symbol_partitioning_class): Declare.
7319 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7321 * ggc.h (ggc_internal_cleared_alloc): New macro.
7322 * vec.h (vec_safe_copy): Handle memory stats.
7323 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7324 * target-globals.c (save_target_globals): Likewise.
7326 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7329 * expr.c (emit_move_resolve_push): Export; be bit more selective
7330 on when to clear alias set.
7331 * expr.h (emit_move_resolve_push): Declare.
7332 * function.h (struct function): Add tail_call_marked.
7333 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7334 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7335 * config/i386/i386.md (TImode move expander): De not call
7337 (FP push expanders): Preserve memory attributes.
7338 * config/i386/sse.md (push<mode>1): Remove.
7339 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7340 (ix86_expand_push): Remove.
7341 * config/i386/mmx.md (push<mode>1): Remove.
7343 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7345 PR rtl-optimization/60030
7346 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7347 lopart with paradoxical subreg before shifting it up by hprec.
7349 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7351 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7352 Remove extra newline at end of file.
7353 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7354 (arm_issue_rate): Handle cortexa57.
7355 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7356 (cortex-a57.cortex-a53): Likewise.
7358 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7361 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7362 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7364 (arm_expand_prologue): Adjust all callers.
7365 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7366 info, registers also at the lowest numbered registers side. Use
7367 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7371 * var-tracking.c (adjust_mems): Before adding a SET to
7372 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7374 2014-02-06 Alan Modra <amodra@gmail.com>
7377 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7378 change SDmode to DDmode when lra_in_progress.
7380 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7383 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7384 free_data_ref on the dr first, and before goto again also set dr
7385 to the next dr. For simd_lane_access, free old datarefs[i] before
7386 overwriting it. For get_vectype_for_scalar_type failure, don't
7387 free_data_ref if simd_lane_access.
7389 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7392 * tree.h (opts_for_fn): New inline function.
7393 (opt_for_fn): Define.
7394 * config/i386/i386.c (ix86_function_regparm): Use
7395 opt_for_fn (decl, optimize) instead of optimize.
7397 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7399 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7400 for SYMBOL_REF in large memory model.
7402 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7404 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7406 (cortex-a57): Likewise.
7407 (cortex-a57.cortex-a53): Likewise.
7409 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7410 Kugan Vivekanandarajah <kuganv@linaro.org>
7412 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7414 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7416 2014-02-06 Richard Biener <rguenther@suse.de>
7418 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7419 set_loop_copy and initialize_original_copy_tables.
7421 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7423 * config/aarch64/aarch64-simd.md
7424 (aarch64_ashr_simddi): Change QI to SI.
7426 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7427 Jakub Jelinek <jakub@redhat.com>
7430 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7433 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7435 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7436 CODE_FOR_altivec_vpku[hw]um to
7437 CODE_FOR_altivec_vpku[hw]um_direct.
7438 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7439 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7440 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7441 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7443 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7445 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7446 generation for -maltivec=be.
7447 (altivec_vsumsws): Simplify redundant test.
7449 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7451 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7452 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7453 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7454 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7455 gen_altivec_vpkuwum.
7456 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7458 (altivec_vpks<VI_char>ss): Likewise.
7459 (altivec_vpks<VI_char>us): Likewise.
7460 (altivec_vpku<VI_char>us): Likewise.
7461 (altivec_vpku<VI_char>um): Likewise.
7462 (altivec_vpku<VI_char>um_direct): New (copy of
7463 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7465 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7466 target is little endian and -maltivec=be is not specified.
7467 (*altivec_vupkhs<VU_char>_direct): New (copy of
7468 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7469 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7470 target is little endian and -maltivec=be is not specified.
7471 (*altivec_vupkls<VU_char>_direct): New (copy of
7472 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7473 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7474 little endian and -maltivec=be is not specified.
7475 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7476 little endian and -maltivec=be is not specified.
7478 2014-02-05 Richard Henderson <rth@redhat.com>
7481 * combine-stack-adj.c: Revert r206943.
7482 * sched-int.h (struct deps_desc): Add last_args_size.
7483 * sched-deps.c (init_deps): Initialize it.
7484 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7485 contain REG_ARGS_SIZE notes.
7487 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7489 * lto-cgraph.c (asm_nodes_output): Make global.
7490 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7491 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7492 (driver_handle_option): Handle OPT_fwpa.
7494 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7497 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7498 a comment typo and formatting issue. If odr_hash hasn't been
7499 created, return vNULL and set *completep to false.
7502 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7503 bb with no successors.
7505 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7508 * doc/invoke.texi (-march): Clarify documentation for ARM.
7512 2014-02-05 Richard Biener <rguenther@suse.de>
7514 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7515 when not vectorizing because of too many alias checks.
7516 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7517 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7519 2014-02-05 Nick Clifton <nickc@redhat.com>
7521 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7522 accept extended registers in any mode when compiling for the MN10300.
7524 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7526 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7527 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7528 sanitization attributes.
7529 (can_inline_edge_p): Likewise.
7530 (sanitize_attrs_match_for_inline_p): New function.
7532 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7534 * ipa-prop.c (detect_type_change): Shor circuit testing of
7535 type changes on THIS pointer.
7537 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7540 * config/pa/pa.c (legitimize_tls_address): Return original address
7541 if not passed a SYMBOL_REF rtx.
7542 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7544 (pa_emit_move_sequence): Simplify TLS source operands.
7545 (pa_legitimate_constant_p): Reject all TLS constants.
7546 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7547 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7549 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7551 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7552 groups when we know they are controlled by LTO.
7553 * varasm.c (default_binds_local_p_1): If object is in other partition,
7554 it will be resolved locally.
7556 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7558 * config/host-linux.c (linux_gt_pch_use_address): Don't
7559 use SSIZE_MAX because it is not always defined.
7561 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7564 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7565 threshold for pseudo splitting.
7566 (update_ebb_live_info): Process call argument hard registers and
7567 hard registers from insn definition too.
7568 (max_small_class_regs_num): New constant.
7569 (inherit_in_ebb): Update live hard regs through EBBs. Update
7570 reloads_num only for small register classes. Don't split for
7573 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7576 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7579 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7581 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7584 2014-02-04 Marek Polacek <polacek@redhat.com>
7586 * gdbinit.in (pel): Define.
7588 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7590 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7593 2014-02-04 Richard Biener <rguenther@suse.de>
7596 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7597 in function context local.
7598 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7599 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7600 similar to LTO_imported_decl_ref.
7602 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7604 PR tree-optimization/60002
7605 * cgraphclones.c (build_function_decl_skip_args): Clear
7608 PR tree-optimization/60023
7609 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7610 false to gsi_replace.
7611 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7612 has been in some EH region and vec_stmt could throw, add
7613 vec_stmt into the same EH region.
7614 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7615 has no lhs, ignore it.
7616 * internal-fn.c (expand_MASK_LOAD): Likewise.
7619 * tree-inline.c (copy_forbidden): Fail for
7620 __attribute__((optimize (0))) functions.
7623 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7624 is set, copy one less argument.
7625 (expand_simd_clones): Don't subtract clone_info->inbranch
7626 from simd_clone_struct_alloc argument.
7628 PR rtl-optimization/57915
7629 * recog.c (simplify_while_replacing): If all unary/binary/relational
7630 operation arguments are constant, attempt to simplify those.
7633 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7634 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7636 2014-02-04 Richard Biener <rguenther@suse.de>
7638 PR tree-optimization/60012
7639 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7640 TBAA disambiguation to all DDRs.
7642 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7645 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7646 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7648 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7651 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7653 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7655 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7656 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7658 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7661 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7662 to figure out targets of polymorphic calls with known decl.
7663 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7664 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7665 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7666 (get_polymorphic_call_info): ... here.
7667 (get_polymorphic_call_info_from_invariant): New function.
7669 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7671 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7672 lookup via vtable pointer; check for type consistency
7673 and turn inconsitent facts into UNREACHABLE.
7674 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7675 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7676 type inconsistent querries; return UNREACHABLE instead.
7678 2014-02-03 Richard Henderson <rth@twiddle.net>
7681 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7682 already processed this node.
7683 (normalize_one_pred_1): Pass along mark_set.
7684 (normalize_one_pred): Create and destroy a pointer_set_t.
7685 (normalize_one_pred_chain): Likewise.
7687 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7689 PR gcov-profile/58602
7690 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7692 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7695 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7696 -fno-devirtualize; try to devirtualize by the knowledge of
7697 virtual table pointer given by aggregate propagation.
7698 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7699 (ipa_print_node_jump_functions): Dump also offset that
7700 is relevant for polymorphic calls.
7701 (determine_known_aggregate_parts): Add arg_type parameter; use it
7702 instead of determining the type from pointer type.
7703 (ipa_compute_jump_functions_for_edge): Update call of
7704 determine_known_aggregate_parts.
7705 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7706 (gimple_get_virt_method_for_binfo): ... here; simplify using
7707 vtable_pointer_value_to_vtable.
7708 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7709 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7710 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7711 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7713 (vtable_pointer_value_to_binfo): ... here.
7714 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7716 2014-02-03 Teresa Johnson <tejohnson@google.com>
7718 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7719 redef of outer loop index variable.
7721 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7725 * doc/extend.texi (Function Attributes): Typo.
7727 2014-02-03 Cong Hou <congh@google.com>
7729 PR tree-optimization/60000
7730 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7731 if the vectorized statement is a store. A store statement can only
7732 appear at the end of pattern statements.
7734 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7736 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7737 (ix86_option_override_internal): Default long double to 64-bit for
7738 32-bit Bionic and to 128-bit for 64-bit Bionic.
7740 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7741 TARGET_LONG_DOUBLE_128 is true.
7742 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7744 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7745 (mlong-double-64): Negate -mlong-double-128.
7746 (mlong-double-128): New option.
7748 * config/i386/i386-c.c (ix86_target_macros): Define
7749 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7751 * doc/invoke.texi: Document -mlong-double-128.
7753 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7755 PR rtl-optimization/60024
7756 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7758 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7760 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7762 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7764 PR rtl-optimization/57662
7765 * sel-sched.c (code_motion_path_driver): Do not mark already not
7766 existing blocks in the visiting bitmap.
7768 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7770 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7771 on the insn being emitted.
7773 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7774 Will Deacon <will.deacon@arm.com>
7776 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7778 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7780 * config/arm/arm-tables.opt: Regenerate.
7782 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7784 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7785 for vector types other than V16QImode.
7786 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7787 define_expand, and call altivec_expand_vec_perm_le when producing
7788 code with little endian element order.
7789 (*altivec_vperm_<mode>_internal): New insn having previous
7790 behavior of altivec_vperm_<mode>.
7791 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7792 altivec_expand_vec_perm_le when producing code with little endian
7794 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7795 behavior of altivec_vperm_<mode>_uns.
7797 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7799 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7800 (altivec_vsumsws): Add handling for -maltivec=be with a little
7802 (altivec_vsumsws_direct): New.
7803 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7804 gen_altivec_vsumsws.
7806 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7808 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7809 vtable_pointer_value_to_binfo): New functions.
7810 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7811 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7813 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7815 * config/nios2/nios2.md (load_got_register): Initialize GOT
7816 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7817 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7819 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7821 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7822 preserverd by passthrough, do not propagate the type.
7824 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7826 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7827 (mips_atomic_assign_expand_fenv): New function.
7828 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7830 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7832 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7833 (__builtin_mips_set_fcsr): Likewise.
7834 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7835 MIPS_USI_FTYPE_VOID.
7836 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7837 (mips16_expand_set_fcsr): Likewise.
7838 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7839 (mips16_set_fcsr_stub): Likewise.
7840 (mips16_get_fcsr_one_only_stub): New class.
7841 (mips16_set_fcsr_one_only_stub): Likewise.
7842 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7843 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7844 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7845 (hard_float): New availability predicate.
7846 (mips_builtins): Add get_fcsr and set_fcsr.
7847 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7848 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7849 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7850 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7851 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7854 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7856 * config/mips/mips.c (mips_one_only_stub): New class.
7857 (mips_need_mips16_rdhwr_p): Replace with...
7858 (mips16_rdhwr_stub): ...this new variable.
7859 (mips16_stub_call_address): New function.
7860 (mips16_rdhwr_one_only_stub): New class.
7861 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7862 (mips_output_mips16_rdhwr): Delete.
7863 (mips_finish_stub): New function.
7864 (mips_code_end): Use it to handle rdhwr stubs.
7866 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7869 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7870 when calculating size of integer atomic types.
7872 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7874 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7876 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7878 PR tree-optimization/60003
7879 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7880 * profile.c (branch_prob): Use gimple_call_builtin_p
7881 to check for BUILT_IN_SETJMP_RECEIVER.
7882 * tree-inline.c (copy_bb): Call notice_special_calls.
7884 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7887 * lra-constraints.c (process_alt_operands): Update reload_sum only
7890 2014-01-31 Richard Henderson <rth@redhat.com>
7893 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7894 until after else_eh is processed.
7896 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7898 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7899 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7900 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7901 in smmintrin.h, remove them.
7902 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7903 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7904 * config/i386/i386.md (ROUND_SAE): Fix value.
7905 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7906 (const48_operand): New.
7907 * config/i386/subst.md (round), (round_expand): Use
7908 const_4_or_8_to_11_operand.
7909 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7911 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7913 * config/i386/constraints.md (Yk): Swap meaning with k.
7914 * config/i386/i386.md (movhi_internal): Change Yk to k.
7915 (movqi_internal): Ditto.
7916 (*k<logic><mode>): Ditto.
7919 (kandn<mode>): Ditto.
7920 (*<code>hi_1): Ditto.
7921 (*<code>qi_1): Ditto.
7922 (kxnor<mode>): Ditto.
7923 (kortestzhi): Ditto.
7924 (kortestchi): Ditto.
7926 (*one_cmplhi2_1): Ditto.
7927 (*one_cmplqi2_1): Ditto.
7928 * config/i386/sse.md (): Change k to Yk.
7929 (avx512f_load<mode>_mask): Ditto.
7930 (avx512f_blendm<mode>): Ditto.
7931 (avx512f_store<mode>_mask): Ditto.
7932 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7933 (avx512f_storedqu<mode>_mask): Ditto.
7934 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7936 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7937 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7938 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7939 (avx512f_maskcmp<mode>3): Ditto.
7940 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7941 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7942 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7943 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7944 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7945 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7946 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7947 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7948 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7949 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7950 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7951 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7952 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7953 (vec_extract_lo_<mode>_maskm): Ditto.
7954 (vec_extract_hi_<mode>_maskm): Ditto.
7955 (avx512f_vternlog<mode>_mask): Ditto.
7956 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7957 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7958 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7959 (avx512f_<code>v8div16qi2_mask): Ditto.
7960 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7961 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7962 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7963 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7964 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7965 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7966 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7967 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7968 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7969 (avx512cd_maskb_vec_dupv8di): Ditto.
7970 (avx512cd_maskw_vec_dupv16si): Ditto.
7971 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7972 (avx512f_vpermi2var<mode>3_mask): Ditto.
7973 (avx512f_vpermi2var<mode>3_mask): Ditto.
7974 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7975 (*avx512f_gathersi<mode>): Ditto.
7976 (*avx512f_gathersi<mode>_2): Ditto.
7977 (*avx512f_gatherdi<mode>): Ditto.
7978 (*avx512f_gatherdi<mode>_2): Ditto.
7979 (*avx512f_scattersi<mode>): Ditto.
7980 (*avx512f_scatterdi<mode>): Ditto.
7981 (avx512f_compress<mode>_mask): Ditto.
7982 (avx512f_compressstore<mode>_mask): Ditto.
7983 (avx512f_expand<mode>_mask): Ditto.
7984 * config/i386/subst.md (mask): Change k to Yk.
7985 (mask_scalar_merge): Ditto.
7988 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7990 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7992 2014-01-31 Richard Biener <rguenther@suse.de>
7995 * builtins.c (fold_builtin_memory_op): Make sure to not
7996 use a floating-point mode or a boolean or enumeral type for
7999 2014-01-30 DJ Delorie <dj@redhat.com>
8001 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8002 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8003 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8004 whenever main() has an epilogue.
8006 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8008 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8009 unused variable "field".
8010 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8011 (vsx_mergeh_<mode>): Likewise.
8012 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8013 (altivec_vmrghh): Likewise.
8014 (altivec_vmrghw): Likewise.
8015 (altivec_vmrglb): Likewise.
8016 (altivec_vmrglh): Likewise.
8017 (altivec_vmrglw): Likewise.
8018 (altivec_vspltb): Add missing uses.
8019 (altivec_vsplth): Likewise.
8020 (altivec_vspltw): Likewise.
8021 (altivec_vspltsf): Likewise.
8023 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8026 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8027 frame related instructions.
8029 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8031 PR rtl-optimization/59959
8032 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8033 any reload of register whose subreg is invalid.
8035 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8037 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8038 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8039 Add missing return type - void.
8041 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8043 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8044 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8045 remove element index adjustment for endian (now handled in vsx.md
8047 (altivec_expand_vec_perm_const): Use
8048 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8049 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8050 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8051 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8052 define_expand and a new define_insn *altivec_vspltb_internal;
8053 adjust for -maltivec=be on a little endian target.
8054 (altivec_vspltb_direct): New.
8055 (altivec_vsplth): Divide into a define_expand and a new
8056 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8057 little endian target.
8058 (altivec_vsplth_direct): New.
8059 (altivec_vspltw): Divide into a define_expand and a new
8060 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8061 little endian target.
8062 (altivec_vspltw_direct): New.
8063 (altivec_vspltsf): Divide into a define_expand and a new
8064 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8065 a little endian target.
8067 2014-01-30 Richard Biener <rguenther@suse.de>
8069 PR tree-optimization/59993
8070 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8071 can propagate form the earlier stmt and avoid the transform
8072 when the intermediate result is needed.
8074 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8076 * README.Portability: Fix typo.
8078 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8080 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8081 comparison_operator with ordered_comparison_operator.
8083 2014-01-30 Nick Clifton <nickc@redhat.com>
8085 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8086 Rename to mn10300_store_multiple_regs.
8087 * config/mn10300/mn10300.c: Likewise.
8088 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8089 store_multiple_regs.
8090 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8091 Call mn10300_store_multiple_regs.
8093 2014-01-30 Nick Clifton <nickc@redhat.com>
8094 DJ Delorie <dj@redhat.com>
8096 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8097 %fp 2 to keep registers after it properly word-aligned.
8098 (rl78_alloc_physical_registers_umul): Handle the case where both
8099 input operands are the same.
8101 2014-01-30 Richard Biener <rguenther@suse.de>
8103 PR tree-optimization/59903
8104 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8107 2014-01-30 Jason Merrill <jason@redhat.com>
8110 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8113 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8115 2014-01-30 Richard Biener <rguenther@suse.de>
8117 PR tree-optimization/59951
8118 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8120 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8123 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8124 SFmode to DFmode case.
8126 2014-01-29 DJ Delorie <dj@redhat.com>
8128 * config/msp430/msp430.opt (-minrt): New.
8129 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8131 (ENDFILE_SPEC): Likewise.
8133 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8135 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8136 (estimate_function_body_sizes): Use it.
8138 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8141 * dwarf2out.c (is_cxx_auto): New.
8142 (is_base_type): Use it.
8143 (gen_type_die_with_usage): Likewise.
8145 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8147 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8148 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8149 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8150 -maltivec=be with LE targets.
8151 (vsx_mergeh_<mode>): Likewise.
8152 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8153 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8154 (altivec_vmrghb): Replace with define_expand and new
8155 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8156 (altivec_vmrghb_direct): New define_insn.
8157 (altivec_vmrghh): Replace with define_expand and new
8158 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8159 (altivec_vmrghh_direct): New define_insn.
8160 (altivec_vmrghw): Replace with define_expand and new
8161 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8162 (altivec_vmrghw_direct): New define_insn.
8163 (*altivec_vmrghsf): Adjust for endianness.
8164 (altivec_vmrglb): Replace with define_expand and new
8165 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8166 (altivec_vmrglb_direct): New define_insn.
8167 (altivec_vmrglh): Replace with define_expand and new
8168 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8169 (altivec_vmrglh_direct): New define_insn.
8170 (altivec_vmrglw): Replace with define_expand and new
8171 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8172 (altivec_vmrglw_direct): New define_insn.
8173 (*altivec_vmrglsf): Adjust for endianness.
8174 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8175 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8176 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8177 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8178 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8179 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8180 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8181 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8183 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8185 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8186 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8189 2014-01-29 Richard Biener <rguenther@suse.de>
8191 PR tree-optimization/58742
8192 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8193 associate_pointerplus_align.
8194 (associate_pointerplus_diff): New function.
8195 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8196 and associate_pointerplus_diff.
8198 2014-01-29 Richard Biener <rguenther@suse.de>
8200 * lto-streamer.h (LTO_major_version): Bump to 3.
8201 (LTO_minor_version): Reset to 0.
8203 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8205 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8206 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8207 (arm_file_start): Generate correct asm header for armv7ve.
8208 * config/arm/bpabi.h: Add multilib support for armv7ve.
8209 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8210 and cortex-a15 to armv7ve.
8211 * config/arm/t-aprofile: Add multilib support for armv7ve.
8212 * doc/invoke.texi: Document -march=armv7ve.
8214 2014-01-29 Richard Biener <rguenther@suse.de>
8216 PR tree-optimization/58742
8217 * tree-ssa-forwprop.c (associate_plusminus): Return true
8218 if we changed sth, defer EH cleanup to ...
8219 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8220 (simplify_mult): New function.
8222 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8225 PR tree-optimization/59920
8226 * tree.c (build_common_builtin_nodes): Remove
8227 __builtin_setjmp_dispatcher initialization.
8228 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8229 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8230 instead of gsi_after_labels + manually skipping debug stmts.
8231 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8232 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8233 * tree-inline.c (copy_edges_for_bb): Remove
8234 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8235 argument. Ignore computed_goto_p stmts. Don't call
8236 make_abnormal_goto_edges. If a call might need abnormal edges
8237 for non-local gotos, see if it already has an edge to
8238 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8239 with true argument, don't do anything then, otherwise add
8240 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8241 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8243 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8244 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8245 (lower_stmt): Don't set data->calls_builtin_setjmp.
8246 (lower_builtin_setjmp): Adjust comment.
8247 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8248 * tree-cfg.c (found_computed_goto): Remove.
8249 (factor_computed_gotos): Remove.
8250 (make_goto_expr_edges): Return bool, true for computed gotos.
8251 Don't call make_abnormal_goto_edges.
8252 (build_gimple_cfg): Don't set found_computed_goto, don't call
8253 factor_computed_gotos.
8254 (computed_goto_p): No longer static.
8255 (make_blocks): Don't set found_computed_goto.
8256 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8257 (make_edges): If make_goto_expr_edges returns true, push bb
8258 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8259 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8260 vector. Record mapping between bbs and OpenMP regions if there
8261 are any, adjust make_gimple_omp_edges caller. Call
8262 handle_abnormal_edges.
8263 (make_abnormal_goto_edges): Remove.
8264 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8265 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8266 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8267 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8268 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8269 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8270 filling *region also set *region_idx to (*region)->entry->index.
8273 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8274 For REGs set ORIGINAL_REGNO.
8276 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8278 * doc/md.texi: Mention that a target shouldn't implement
8279 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8282 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8284 PR tree-optimization/59594
8285 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8286 a copy of the datarefs vector rather than the vector itself.
8288 2014-01-28 Jason Merrill <jason@redhat.com>
8291 * dwarf2out.c (auto_die): New static.
8292 (gen_type_die_with_usage): Handle C++1y 'auto'.
8293 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8296 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8299 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8300 (SPEC_X32): Likewise.
8301 (SPEC_64): Likewise.
8302 * config/i386/i386.c (ix86_option_override_internal): Turn off
8303 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8305 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8306 * config/i386/i386.h (TARGET_16BIT): New macro.
8307 (TARGET_16BIT_P): Likewise.
8308 * config/i386/i386.opt: Add m16.
8309 * doc/invoke.texi: Document -m16.
8311 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8313 PR preprocessor/59935
8314 * input.c (location_get_source_line): Bail out on when line number
8315 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8317 2014-01-28 Richard Biener <rguenther@suse.de>
8319 PR tree-optimization/58742
8320 * tree-ssa-forwprop.c (associate_plusminus): Handle
8321 pointer subtraction of the form (T)(P + A) - (T)P.
8323 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8325 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8328 2014-01-28 Richard Biener <rguenther@suse.de>
8331 2014-01-28 Richard Biener <rguenther@suse.de>
8333 PR rtl-optimization/45364
8334 PR rtl-optimization/59890
8335 * var-tracking.c (local_get_addr_clear_given_value): Handle
8336 already cleared slot.
8337 (val_reset): Handle not allocated local_get_addr_cache.
8338 (vt_find_locations): Use post-order on the inverted CFG.
8340 2014-01-28 Richard Biener <rguenther@suse.de>
8342 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8344 2014-01-28 Richard Biener <rguenther@suse.de>
8346 PR rtl-optimization/45364
8347 PR rtl-optimization/59890
8348 * var-tracking.c (local_get_addr_clear_given_value): Handle
8349 already cleared slot.
8350 (val_reset): Handle not allocated local_get_addr_cache.
8351 (vt_find_locations): Use post-order on the inverted CFG.
8353 2014-01-28 Alan Modra <amodra@gmail.com>
8355 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8356 * configure.ac <recursive call for build != host>: Define
8357 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8358 and LD_FOR_BUILD too.
8359 * configure: Regenerate.
8361 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8363 * config/i386/i386.c (get_builtin_code_for_version): Separate
8364 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8365 Broadwell from Haswell.
8367 2014-01-27 Steve Ellcey <sellcey@mips.com>
8369 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8370 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8371 * config/mips/mips.c (mips_option_override): Change setting
8373 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8374 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8375 Change from Mask to Var.
8377 2014-01-27 Jeff Law <law@redhat.com>
8379 * ipa-inline.c (inline_small_functions): Fix typo.
8381 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8383 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8384 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8385 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8386 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8387 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8388 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8389 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8390 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8391 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8392 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8393 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8394 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8395 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8396 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8397 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8398 (_mm512_storeu_epi64): Ditto.
8399 (_mm512_cmpge_epi32_mask): Ditto.
8400 (_mm512_cmpge_epu32_mask): Ditto.
8401 (_mm512_cmpge_epi64_mask): Ditto.
8402 (_mm512_cmpge_epu64_mask): Ditto.
8403 (_mm512_cmple_epi32_mask): Ditto.
8404 (_mm512_cmple_epu32_mask): Ditto.
8405 (_mm512_cmple_epi64_mask): Ditto.
8406 (_mm512_cmple_epu64_mask): Ditto.
8407 (_mm512_cmplt_epi32_mask): Ditto.
8408 (_mm512_cmplt_epu32_mask): Ditto.
8409 (_mm512_cmplt_epi64_mask): Ditto.
8410 (_mm512_cmplt_epu64_mask): Ditto.
8411 (_mm512_cmpneq_epi32_mask): Ditto.
8412 (_mm512_cmpneq_epu32_mask): Ditto.
8413 (_mm512_cmpneq_epi64_mask): Ditto.
8414 (_mm512_cmpneq_epu64_mask): Ditto.
8415 (_mm512_expand_pd): Ditto.
8416 (_mm512_expand_ps): Ditto.
8417 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8418 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8419 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8420 * config/i386/i386.c (ix86_builtins): Add
8421 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8422 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8423 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8424 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8425 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8426 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8427 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8428 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8429 IX86_BUILTIN_PMOVUSQW512_MEM.
8430 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8431 __builtin_ia32_pmovsqd512mem_mask,
8432 __builtin_ia32_pmovqd512mem_mask,
8433 __builtin_ia32_pmovusqw512mem_mask,
8434 __builtin_ia32_pmovsqw512mem_mask,
8435 __builtin_ia32_pmovqw512mem_mask,
8436 __builtin_ia32_pmovusdw512mem_mask,
8437 __builtin_ia32_pmovsdw512mem_mask,
8438 __builtin_ia32_pmovdw512mem_mask,
8439 __builtin_ia32_pmovqb512mem_mask,
8440 __builtin_ia32_pmovusqb512mem_mask,
8441 __builtin_ia32_pmovsqb512mem_mask,
8442 __builtin_ia32_pmovusdb512mem_mask,
8443 __builtin_ia32_pmovsdb512mem_mask,
8444 __builtin_ia32_pmovdb512mem_mask.
8445 (bdesc_args): Add __builtin_ia32_expanddf512,
8446 __builtin_ia32_expandsf512.
8447 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8448 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8449 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8450 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8451 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8452 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8453 (avx512f_<code>v8div16qi2_mask_store): This.
8454 (avx512f_expand<mode>): New.
8456 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8458 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8460 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8461 (_mm512_prefetch_i32scatter_pd): Ditto.
8462 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8463 (_mm512_prefetch_i64scatter_pd): Ditto.
8464 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8465 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8466 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8467 (_mm512_prefetch_i32scatter_ps): Ditto.
8468 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8469 (_mm512_prefetch_i64scatter_ps): Ditto.
8470 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8471 * config/i386/i386-builtin-types.def: Define
8472 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8473 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8474 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8475 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8476 IX86_BUILTIN_SCATTERPFQPD.
8477 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8478 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8479 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8480 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8481 __builtin_ia32_scatterpfqps.
8482 (ix86_expand_builtin): Expand new built-ins.
8483 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8484 fix memory access data type.
8485 (*avx512pf_gatherpf<mode>_mask): Ditto.
8486 (*avx512pf_gatherpf<mode>): Ditto.
8487 (avx512pf_scatterpf<mode>): Ditto.
8488 (*avx512pf_scatterpf<mode>_mask): Ditto.
8489 (*avx512pf_scatterpf<mode>): Ditto.
8490 (GATHER_SCATTER_SF_MEM_MODE): New.
8491 (avx512pf_gatherpf<mode>df): Ditto.
8492 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8493 (*avx512pf_scatterpf<mode>df): Ditto.
8495 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8498 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8499 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8502 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8504 * common/config/arm/arm-common.c
8505 (arm_rewrite_mcpu): Handle multiple names.
8507 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8509 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8511 * gimple-builder.h (create_gimple_tmp): Delete.
8513 2014-01-27 Christian Bruel <christian.bruel@st.com>
8515 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8518 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8520 * config/pa/pa.md (call): Generate indirect long calls to non-local
8521 functions when outputing 32-bit code.
8522 (call_value): Likewise except for special call to buggy powf function.
8524 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8525 portable runtime and PIC indirect calls.
8526 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8527 and PIC call sequences. Use ldo instead of blr to set return register
8528 in PIC call sequence.
8530 2014-01-25 Walter Lee <walt@tilera.com>
8532 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8533 avoid clobbering a live register.
8535 2014-01-25 Walter Lee <walt@tilera.com>
8537 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8538 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8539 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8540 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8542 2014-01-25 Walter Lee <walt@tilera.com>
8544 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8545 arguments on even registers.
8546 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8548 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8549 (BIGGEST_ALIGNMENT): Ditto.
8550 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8552 2014-01-25 Walter Lee <walt@tilera.com>
8554 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8555 insns before bundling.
8556 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8558 2014-01-25 Walter Lee <walt@tilera.com>
8560 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8561 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8562 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8564 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8566 * config/mips/constraints.md (kl): Delete.
8567 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8568 define expands, using...
8569 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8570 instructions for MIPS16.
8571 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8572 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8574 2014-01-25 Walter Lee <walt@tilera.com>
8576 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8580 2014-01-25 Walter Lee <walt@tilera.com>
8582 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8583 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8585 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8587 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8590 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8592 * print-rtl.c (in_call_function_usage): New var.
8593 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8594 EXPR_LIST mode as mode and not as reg note name.
8597 * cfgloopmanip.c (copy_loop_info): If
8598 loop->warned_aggressive_loop_optimizations, make sure
8599 the flag is set in target loop too.
8601 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8603 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8605 * builtins.def: Likewise.
8606 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8607 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8608 * ira.c (ira_setup_eliminable_regset): Likewise.
8609 * omp-low.c (gate_expand_omp): Likewise.
8610 (execute_lower_omp): Likewise.
8611 (diagnose_sb_0): Likewise.
8612 (gate_diagnose_omp_blocks): Likewise.
8613 (simd_clone_clauses_extract): Likewise.
8616 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8618 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8619 correction for little endian...
8620 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8623 2014-01-24 Jeff Law <law@redhat.com>
8625 PR tree-optimization/59919
8626 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8627 for non-returning calls.
8629 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8631 * common/config/aarch64/aarch64-common.c
8632 (aarch64_rewrite_mcpu): Handle multiple names.
8633 * config/aarch64/aarch64.h
8634 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8636 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8638 * input.c (add_file_to_cache_tab): Handle the case where fopen
8641 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8644 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8645 from push operand if code of push isn't PRE_DEC.
8647 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8650 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8651 -mquad-memory-atomic. Update -mquad-memory documentation to say
8652 it is only used for non-atomic loads/stores.
8654 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8655 -mquad-memory or -mquad-memory-atomic switches.
8657 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8658 -mquad-memory-atomic to ISA 2.07 support.
8660 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8661 to separate support of normal quad word memory operations (ldq, stq)
8662 from the atomic quad word memory operations.
8664 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8665 support to separate non-atomic quad word operations from atomic
8666 quad word operations. Disable non-atomic quad word operations in
8667 little endian mode so that we don't have to swap words after the
8668 load and before the store.
8669 (quad_load_store_p): Add comment about atomic quad word support.
8670 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8671 options printed with -mdebug=reg.
8673 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8674 -mquad-memory-atomic as the test for whether we have quad word
8675 atomic instructions.
8676 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8677 or -mp8-vector are used, allow byte/half-word atomic operations.
8679 * config/rs6000/sync.md (load_lockedti): Insure that the address
8680 is a proper indexed or indirect address for the lqarx instruction.
8681 On little endian systems, swap the hi/lo registers after the lqarx
8683 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8684 insure the address is valid for the lqarx instruction.
8685 (store_conditionalti): Insure that the address is a proper indexed
8686 or indirect address for the stqcrx. instruction. On little endian
8687 systems, swap the hi/lo registers before doing the stqcrx.
8689 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8690 insure the address is valid for the stqcrx. instruction.
8692 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8693 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8694 type of quad memory support is available.
8696 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8699 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8700 there is a danger of looping.
8702 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8704 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8705 force flag_ira_loop_pressure if set via command line.
8707 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8709 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8710 (ashr_simd): New builtin handling DI mode.
8711 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8712 (aarch64_sshr_simddi): New match pattern.
8713 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8714 (vshrd_n_s64): Likewise.
8715 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8717 2014-01-23 Nick Clifton <nickc@redhat.com>
8719 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8720 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8721 favour of mcu specific scripts.
8722 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8725 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8726 Alex Velenko <Alex.Velenko@arm.com>
8728 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8729 (vaddv_s16): Likewise.
8730 (vaddv_s32): Likewise.
8731 (vaddv_u8): Likewise.
8732 (vaddv_u16): Likewise.
8733 (vaddv_u32): Likewise.
8734 (vaddvq_s8): Likewise.
8735 (vaddvq_s16): Likewise.
8736 (vaddvq_s32): Likewise.
8737 (vaddvq_s64): Likewise.
8738 (vaddvq_u8): Likewise.
8739 (vaddvq_u16): Likewise.
8740 (vaddvq_u32): Likewise.
8741 (vaddvq_u64): Likewise.
8742 (vaddv_f32): Likewise.
8743 (vaddvq_f32): Likewise.
8744 (vaddvq_f64): Likewise.
8745 (vmaxv_f32): Likewise.
8746 (vmaxv_s8): Likewise.
8747 (vmaxv_s16): Likewise.
8748 (vmaxv_s32): Likewise.
8749 (vmaxv_u8): Likewise.
8750 (vmaxv_u16): Likewise.
8751 (vmaxv_u32): Likewise.
8752 (vmaxvq_f32): Likewise.
8753 (vmaxvq_f64): Likewise.
8754 (vmaxvq_s8): Likewise.
8755 (vmaxvq_s16): Likewise.
8756 (vmaxvq_s32): Likewise.
8757 (vmaxvq_u8): Likewise.
8758 (vmaxvq_u16): Likewise.
8759 (vmaxvq_u32): Likewise.
8760 (vmaxnmv_f32): Likewise.
8761 (vmaxnmvq_f32): Likewise.
8762 (vmaxnmvq_f64): Likewise.
8763 (vminv_f32): Likewise.
8764 (vminv_s8): Likewise.
8765 (vminv_s16): Likewise.
8766 (vminv_s32): Likewise.
8767 (vminv_u8): Likewise.
8768 (vminv_u16): Likewise.
8769 (vminv_u32): Likewise.
8770 (vminvq_f32): Likewise.
8771 (vminvq_f64): Likewise.
8772 (vminvq_s8): Likewise.
8773 (vminvq_s16): Likewise.
8774 (vminvq_s32): Likewise.
8775 (vminvq_u8): Likewise.
8776 (vminvq_u16): Likewise.
8777 (vminvq_u32): Likewise.
8778 (vminnmv_f32): Likewise.
8779 (vminnmvq_f32): Likewise.
8780 (vminnmvq_f64): Likewise.
8782 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8784 * config/aarch64/aarch64-simd.md
8785 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8786 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8787 (*aarch64_mul3_elt<mode>): Likewise.
8788 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8789 (*aarch64_mul3_elt_to_64v2df): Likewise.
8790 (*aarch64_mla_elt<mode>): Likewise.
8791 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8792 (*aarch64_mls_elt<mode>): Likewise.
8793 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8794 (*aarch64_fma4_elt<mode>): Likewise.
8795 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8796 (*aarch64_fma4_elt_to_64v2df): Likewise.
8797 (*aarch64_fnma4_elt<mode>): Likewise.
8798 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8799 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8800 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8801 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8802 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8803 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8804 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8805 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8806 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8808 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8810 * config/aarch64/aarch64-simd.md
8811 (aarch64_be_checked_get_lane<mode>): New define_expand.
8812 * config/aarch64/aarch64-simd-builtins.def
8813 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8814 New builtin definition.
8815 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8816 Use new safe be builtin.
8818 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8820 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8822 (aarch64_be_st1<mode>): Likewise.
8823 (aarch_ld1<VALL:mode>): Define_expand modified.
8824 (aarch_st1<VALL:mode>): Likewise.
8825 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8826 (UNSPEC_ST1): Likewise.
8828 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8830 * config/microblaze/microblaze.md: Add trap insn and attribute
8832 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8834 PR preprocessor/58580
8835 * input.h (location_get_source_line): Take an additional line_size
8837 (void diagnostics_file_cache_fini): Declare new function.
8838 * input.c (struct fcache): New type.
8839 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8840 New static constants.
8841 (diagnostic_file_cache_init, total_lines_num)
8842 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8843 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8844 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8845 (get_next_line, read_next_line, goto_next_line, read_line_num):
8846 New static function definitions.
8847 (diagnostic_file_cache_fini): New function.
8848 (location_get_source_line): Take an additional output line_len
8849 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8851 * diagnostic.c (diagnostic_finish): Call
8852 diagnostic_file_cache_fini.
8853 (adjust_line): Take an additional input parameter for the length
8854 of the line, rather than calculating it with strlen.
8855 (diagnostic_show_locus): Adjust the use of
8856 location_get_source_line and adjust_line with respect to their new
8857 signature. While displaying a line now, do not stop at the first
8858 null byte. Rather, display the zero byte as a space and keep
8859 going until we reach the size of the line.
8860 * Makefile.in: Add vec.o to OBJS-libcommon
8862 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8863 Ilya Tocar <ilya.tocar@intel.com>
8865 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8866 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8867 (__builtin_ia32_kmov16): Ditto.
8868 * config/i386/i386.md (UNSPEC_KMOV): New.
8871 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8873 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8874 (_mm512_storeu_si512): Ditto.
8876 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8879 * rtl.h (get_referenced_operands): Declare.
8880 * recog.c (get_referenced_operands): New function.
8881 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8882 operands have been referenced when recording LO_SUM references.
8884 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8886 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8888 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8890 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8891 Enable for generic and recent AMD targets.
8893 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8895 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8896 ARG_SIZE note when adjustment was eliminated.
8898 2014-01-22 Jeff Law <law@redhat.com>
8900 PR tree-optimization/59597
8901 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8902 in file. Accept new argument REGISTERING and use it to modify
8903 dump output appropriately.
8904 (register_jump_thread): Corresponding changes.
8905 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8906 thread paths involving joiner blocks. Add code to dump cancelled
8907 jump threading paths.
8909 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8911 PR rtl-optimization/59477
8912 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8913 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8915 2014-01-22 Tom Tromey <tromey@redhat.com>
8917 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8919 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8921 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8923 PR rtl-optimization/59896
8924 * lra-constraints.c (process_alt_operands): Check unused note for
8925 matched operands of insn with no output reloads.
8927 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8929 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8930 (mips_move_from_gpr_cost): Likewise.
8932 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8934 PR rtl-optimization/59858
8935 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8936 ira_class_hard_regs_num.
8937 (process_alt_operands): Increase reject for dying matched operand.
8939 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8942 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8943 smaller than size, perform several stores or loads and stores
8944 at dst + count - size to store or copy all of size bytes, rather
8945 than just last modesize bytes.
8947 2014-01-20 DJ Delorie <dj@redhat.com>
8949 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8950 that CLOBBERs are REGs before propogating their values.
8952 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8955 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8956 (cgraph_inline_failed_type): New function.
8957 * cgraph.h (DEFCIFCODE): Add type.
8958 (cgraph_inline_failed_type_t): New enum.
8959 (cgraph_inline_failed_type): New prototype.
8960 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8961 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8962 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8963 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8964 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8965 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8966 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8967 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8968 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8969 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8970 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8971 OPTIMIZATION_MISMATCH.
8972 * tree-inline.c (expand_call_inline): Emit errors during
8973 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8975 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8978 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8979 mode attribute in insn output.
8981 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8983 * output.h (output_constant): Delete.
8984 * varasm.c (output_constant): Make private.
8986 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8988 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8990 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8993 * tree.h (fold_builtin_strcat): New prototype.
8994 * builtins.c (fold_builtin_strcat): No longer static. Add len
8995 argument, if non-NULL, don't call c_strlen. Optimize
8996 directly into __builtin_memcpy instead of __builtin_strcpy.
8997 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8998 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9000 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9002 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9003 for SImode_address_operand operands, having only a REG argument.
9005 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9007 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9008 loader name using mbig-endian.
9009 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9011 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9013 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9017 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9019 * config/aarch64/aarch64-protos.h
9020 (aarch64_cannot_change_mode_class_ptr): Declare.
9021 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9022 aarch64_cannot_change_mode_class_ptr): New.
9023 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9024 backend hook aarch64_cannot_change_mode_class.
9026 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9028 * common/config/aarch64/aarch64-common.c
9029 (aarch64_handle_option): Don't handle any option order logic here.
9030 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9031 selected_cpu, warn on architecture version mismatch.
9032 (aarch64_override_options): Fix parsing order for option strings.
9034 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9035 Iain Sandoe <iain@codesourcery.com>
9038 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9039 warning. Amend comment to reflect current functionality.
9041 2014-01-20 Richard Biener <rguenther@suse.de>
9044 * builtins.c (fold_builtin_strcat): Remove case better handled
9045 by tree-ssa-strlen.c.
9047 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9049 * config/aarch64/aarch64.opt
9050 (mcpu, march, mtune): Make case-insensitive.
9052 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9055 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9056 if operands[1] is a REG or ZERO_EXTEND of a REG.
9058 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9060 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9062 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9064 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9065 long non-pic millicode calls.
9067 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9069 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9071 2014-01-19 Kito Cheng <kito@0xlab.org>
9073 * builtins.c (expand_movstr): Check movstr expand done or fail.
9075 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9076 H.J. Lu <hongjiu.lu@intel.com>
9079 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9080 to DImode for zero-extended addresses.
9082 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9084 PR rtl-optimization/57763
9085 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9086 on the new indirect jump_insn and increment LABEL_NUSES (label).
9088 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9092 * config.gcc (x86_archs): New variable.
9093 (x86_64_archs): Likewise.
9094 (x86_cpus): Likewise.
9095 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9096 --with-arch/--with-cpu= options.
9097 Support --with-arch=/--with-cpu={nehalem,westmere,
9098 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9100 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9102 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9103 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9105 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9107 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9109 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9112 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9113 clear cpp_get_options (parse_in)->warn_unused_macros for
9114 ix86_target_macros_internal with cpp_define.
9116 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9118 * jump.c (delete_related_insns): Keep (use (insn))s.
9119 * reorg.c (redundant_insn): Check for barriers too.
9121 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9123 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9125 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9127 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9128 call to $$dyncall when TARGET_LONG_CALLS is true.
9130 2014-01-17 Jeff Law <law@redhat.com>
9132 * ree.c (combine_set_extension): Temporarily disable test for
9133 changing number of hard registers.
9135 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9138 * ipa-inline-analysis.c (inline_free_summary):
9139 Do not free summary of aliases.
9141 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9144 * gimplify.c (gimplify_expr): Use create_tmp_var
9145 instead of create_tmp_var_raw. If cond doesn't have
9146 integral type, don't add the IFN_ANNOTATE builtin at all.
9148 2014-01-17 Martin Jambor <mjambor@suse.cz>
9151 * ipa-cp.c (prev_edge_clone): New variable.
9152 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9153 Also resize prev_edge_clone vector.
9154 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9155 (ipcp_edge_removal_hook): New function.
9156 (ipcp_driver): Register ipcp_edge_removal_hook.
9158 2014-01-17 Andrew Pinski <apinski@cavium.com>
9159 Steve Ellcey <sellcey@mips.com>
9162 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9165 2014-01-17 Jeff Law <law@redhat.com>
9168 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9169 so that pass_ccp runs first.
9171 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9173 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9174 (ix86_adjust_cost): Use !TARGET_XXX.
9175 (do_reorder_for_imul): Likewise.
9176 (swap_top_of_ready_list): Likewise.
9177 (ix86_sched_reorder): Likewise.
9179 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9181 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9182 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9183 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9184 (intel_memset): New. Duplicate slm_memset.
9185 (intel_cost): New. Duplicate slm_cost.
9186 (m_INTEL): New macro.
9187 (processor_target_table): Add "intel".
9188 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9189 with PROCESSOR_INTEL for "intel".
9190 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9191 PROCESSOR_SILVERMONT.
9192 (ix86_issue_rate): Likewise.
9193 (ix86_adjust_cost): Likewise.
9194 (ia32_multipass_dfa_lookahead): Likewise.
9195 (swap_top_of_ready_list): Likewise.
9196 (ix86_sched_reorder): Likewise.
9197 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9198 instead of TARGET_OPT_AGU.
9199 * config/i386/i386.h (TARGET_INTEL): New.
9200 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9201 (processor_type): Add PROCESSOR_INTEL.
9202 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9203 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9205 2014-01-17 Marek Polacek <polacek@redhat.com>
9208 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9211 2014-01-17 Richard Biener <rguenther@suse.de>
9213 PR tree-optimization/46590
9214 * opts.c (default_options_table): Add entries for
9215 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9216 all enabled at -O1 but not for -Og.
9217 * common.opt (fbranch-count-reg): Remove Init(1).
9218 (fmove-loop-invariants): Likewise.
9219 (ftree-pta): Likewise.
9221 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9223 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9224 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9225 decls to at least the GCC 4.8 used alignments.
9228 * tree-nested.c (convert_nonlocal_reference_stmt,
9229 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9230 of GIMPLE_BIND stmts, adjust associated decls.
9232 2014-01-17 Richard Biener <rguenther@suse.de>
9234 PR tree-optimization/46590
9235 * vec.h (vec<>::bseach): New member function implementing
9236 binary search according to C89 bsearch.
9237 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9238 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9239 bitmap pointer again. Make accesses_in_loop a flat array.
9240 (mem_ref_obstack): New global.
9241 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9242 (mark_ref_stored): Likewise.
9243 (ref_indep_loop_p_2): Likewise.
9244 (set_ref_stored_in_loop): New helper function.
9245 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9246 (memref_free): Adjust.
9247 (record_mem_ref_loc): Simplify.
9248 (gather_mem_refs_stmt): Adjust.
9249 (sort_locs_in_loop_postorder_cmp): New function.
9250 (analyze_memory_references): Sort accesses_in_loop after
9251 loop postorder number.
9252 (find_ref_loc_in_loop_cmp): New function.
9253 (for_all_locs_in_loop): Find relevant cluster of locs in
9254 accesses_in_loop and iterate without recursion.
9255 (execute_sm): Avoid uninit warning.
9256 (struct ref_always_accessed): Simplify.
9257 (ref_always_accessed::operator ()): Likewise.
9258 (ref_always_accessed_p): Likewise.
9259 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9260 loop postorder numbers here.
9261 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9264 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9267 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9268 on decls for which assemble_alias has been called.
9270 2014-01-17 Nick Clifton <nickc@redhat.com>
9272 * config/msp430/msp430.opt: (mcpu): New option.
9273 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9274 (msp430_option_override): Parse target_cpu. If the MCU name
9275 matches a generic string, clear target_mcu.
9276 (msp430_attr): Allow numeric interrupt values up to 63.
9277 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9278 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9280 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9282 * config/msp430/msp430.md (popm): Use %J rather than %I.
9283 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9284 (addhi_cy_i): Use immediate_operand for operand 2.
9285 * doc/invoke.texi: Document -mcpu option.
9287 2014-01-17 Richard Biener <rguenther@suse.de>
9289 PR rtl-optimization/38518
9290 * df.h (df_analyze_loop): Declare.
9291 * df-core.c: Include cfgloop.h.
9292 (df_analyze_1): Split out main part of df_analyze.
9293 (df_analyze): Adjust.
9294 (loop_inverted_post_order_compute): New function.
9295 (loop_post_order_compute): Likewise.
9296 (df_analyze_loop): New function avoiding whole-function
9298 * loop-invariant.c (find_defs): Use df_analyze_loop.
9299 (find_invariants): Adjust.
9300 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9302 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9304 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9305 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9307 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9309 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9310 traversal when removing references.
9312 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9315 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9317 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9320 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9321 pushing a reload for an autoinc when we had previously reloaded an
9322 inner part of the address.
9324 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9326 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9328 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9329 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9330 when not giving up or versioning for alias only because of
9332 (vect_analyze_data_ref_dependences): Set to true.
9333 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9335 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9336 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9340 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9343 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9344 operand 0 predicate for gathers, use a new pseudo as subtarget.
9346 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9349 * lra-constraints.c (process_alt_operands): Add printing debug info.
9350 Check absence of input/output reloads for matched operands too.
9352 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9354 PR rtl-optimization/59835
9355 * ira.c (ira_init_register_move_cost): Increase cost for
9358 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9360 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9362 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9365 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9366 non-register objects. Use gen_(high/low)part more consistently.
9369 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9372 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9373 endian support, remove tests for WORDS_BIG_ENDIAN.
9374 (p8_mfvsrd_3_<mode>): Likewise.
9375 (reload_gpr_from_vsx<mode>): Likewise.
9376 (reload_gpr_from_vsxsf): Likewise.
9377 (p8_mfvsrd_4_disf): Likewise.
9379 2014-01-16 Richard Biener <rguenther@suse.de>
9381 PR rtl-optimization/46590
9382 * lcm.c (compute_antinout_edge): Use postorder iteration.
9383 (compute_laterin): Use inverted postorder iteration.
9385 2014-01-16 Nick Clifton <nickc@redhat.com>
9388 * varasm.c (output_constant): Return the number of bytes actually
9390 (output_constructor_array_range): Update the field size with the
9391 number of bytes emitted by output_constant.
9392 (output_constructor_regular_field): Likewise. Also do not
9393 complain if the total number of bytes emitted is now greater
9394 than the expected fieldpos.
9395 * output.h (output_constant): Update prototype and descriptive comment.
9397 2014-01-16 Marek Polacek <polacek@redhat.com>
9400 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9401 it is error_mark_node.
9403 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9405 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9406 VALID_AVX256_REG_OR_OI_MODE.
9408 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9410 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9411 current procedure should be profiled.
9413 2014-01-15 Andrew Pinski <apinski@cavium.com>
9415 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9416 of moving from/to the STACK_REG register class.
9418 2014-01-15 Richard Henderson <rth@redhat.com>
9421 * reginfo.c (global_regs_decl): Globalize.
9422 * rtl.h (global_regs_decl): Declare.
9423 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9424 reserved via global_regs.
9426 2014-01-15 Teresa Johnson <tejohnson@google.com>
9428 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9430 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9432 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9433 and vmulosh rather than call gen_vec_widen_smult_*.
9434 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9435 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9436 (vec_widen_smult_even_v16qi): Likewise.
9437 (vec_widen_umult_even_v8hi): Likewise.
9438 (vec_widen_smult_even_v8hi): Likewise.
9439 (vec_widen_umult_odd_v16qi): Likewise.
9440 (vec_widen_smult_odd_v16qi): Likewise.
9441 (vec_widen_umult_odd_v8hi): Likewise.
9442 (vec_widen_smult_odd_v8hi): Likewise.
9443 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9444 vmuloub rather than call gen_vec_widen_umult_*.
9445 (vec_widen_umult_lo_v16qi): Likewise.
9446 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9447 vmulosb rather than call gen_vec_widen_smult_*.
9448 (vec_widen_smult_lo_v16qi): Likewise.
9449 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9450 rather than call gen_vec_widen_umult_*.
9451 (vec_widen_umult_lo_v8hi): Likewise.
9452 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9453 rather than call gen_vec_widen_smult_*.
9454 (vec_widen_smult_lo_v8hi): Likewise.
9456 2014-01-15 Jeff Law <law@redhat.com>
9458 PR tree-optimization/59747
9459 * ree.c (find_and_remove_re): Properly handle case where a second
9460 eliminated extension requires widening a copy created for elimination
9461 of a prior extension.
9462 (combine_set_extension): Ensure that the number of hard regs needed
9463 for a destination register does not change when we widen it.
9465 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9467 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9468 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9469 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9470 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9471 (avr-*-rtems*): Likewise.
9472 (bfin*-rtems*): Likewise.
9473 (moxie-*-rtems*): Likewise.
9474 (h8300-*-rtems*): Likewise.
9475 (i[34567]86-*-rtems*): Likewise.
9476 (lm32-*-rtems*): Likewise.
9477 (m32r-*-rtems*): Likewise.
9478 (m68k-*-rtems*): Likewise.
9479 (microblaze*-*-rtems*): Likewise.
9480 (mips*-*-rtems*): Likewise.
9481 (powerpc-*-rtems*): Likewise.
9482 (sh-*-rtems*): Likewise.
9483 (sparc-*-rtems*): Likewise.
9484 (sparc64-*-rtems*): Likewise.
9485 (v850-*-rtems*): Likewise.
9486 (m32c-*-rtems*): Likewise.
9488 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9490 PR rtl-optimization/59511
9491 * ira.c (ira_init_register_move_cost): Use memory costs for some
9492 cases of register move cost calculations.
9493 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9494 instead of BB frequency.
9495 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9496 * lra-assigns.c (find_hard_regno_for): Ditto.
9498 2014-01-15 Richard Biener <rguenther@suse.de>
9500 PR tree-optimization/59822
9501 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9502 (vectorizable_load): Use it to hoist defs of uses of invariant
9503 loads out of the loop.
9505 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9506 Kugan Vivekanandarajah <kuganv@linaro.org>
9509 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9512 2014-01-15 Richard Biener <rguenther@suse.de>
9514 PR rtl-optimization/59802
9515 * lcm.c (compute_available): Use inverted postorder to seed
9516 the initial worklist.
9518 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9521 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9522 ADDR_REGS for invalid symrefs in non-PIC code.
9524 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9527 * builtins.c (determine_block_size): Initialize *probable_max_size
9528 even if len_rtx is CONST_INT.
9530 2014-01-14 Andrew Pinski <apinski@cavium.com>
9532 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9533 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9534 (cortexa53_tunings): Likewise.
9535 (aarch64_sched_issue_rate): New function.
9536 (TARGET_SCHED_ISSUE_RATE): Define.
9538 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9540 * ira-costs.c (find_costs_and_classes): Add missed
9541 ira_init_register_move_cost_if_necessary.
9543 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9546 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9548 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9551 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9552 to indicate if type is used for function return value. Warn ABI
9553 change if the vector mode isn't available for function return value.
9554 (ix86_function_arg_advance): Pass false to type_natural_mode.
9555 (ix86_function_arg): Likewise.
9556 (ix86_gimplify_va_arg): Likewise.
9557 (function_arg_32): Don't warn ABI change.
9558 (ix86_function_value): Pass true to type_natural_mode.
9559 (ix86_return_in_memory): Likewise.
9560 (ix86_struct_value_rtx): Removed.
9561 (TARGET_STRUCT_VALUE_RTX): Likewise.
9563 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9565 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9566 converting a conditional jump into a conditional return.
9568 2014-01-14 Richard Biener <rguenther@suse.de>
9570 PR tree-optimization/58921
9571 PR tree-optimization/59006
9572 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9573 hoisting invariant stmts.
9574 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9575 invariant loads on the preheader edge if possible.
9577 2014-01-14 Joey Ye <joey.ye@arm.com>
9579 * doc/plugin.texi (Building GCC plugins): Update to C++.
9581 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9583 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9584 (_mm_rcp28_round_ss): Ditto.
9585 (_mm_rsqrt28_round_sd): Ditto.
9586 (_mm_rsqrt28_round_ss): Ditto.
9587 (_mm_rcp28_sd): Ditto.
9588 (_mm_rcp28_ss): Ditto.
9589 (_mm_rsqrt28_sd): Ditto.
9590 (_mm_rsqrt28_ss): Ditto.
9591 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9592 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9593 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9594 (IX86_BUILTIN_RCP28SD): Ditto.
9595 (IX86_BUILTIN_RCP28SS): Ditto.
9596 (IX86_BUILTIN_RSQRT28SD): Ditto.
9597 (IX86_BUILTIN_RSQRT28SS): Ditto.
9598 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9599 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9600 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9601 (ix86_expand_special_args_builtin): Expand new FTYPE.
9602 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9603 (srcp14<mode>): Make insn unary.
9604 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9605 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9606 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9607 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9608 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9609 Fix rounding: make it SAE only.
9610 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9612 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9614 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9615 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9616 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9617 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9618 (round_saeonly_mask_scalar_operand4): Ditto.
9619 (round_saeonly_mask_scalar_op3): Ditto.
9620 (round_saeonly_mask_scalar_op4): Ditto.
9622 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9624 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9625 Implement -maltivec=be for vec_insert and vec_extract.
9627 2014-01-10 DJ Delorie <dj@redhat.com>
9629 * config/msp430/msp430.md (call_internal): Don't allow memory
9630 references with SP as the base register.
9631 (call_value_internal): Likewise.
9632 * config/msp430/constraints.md (Yc): New. For memory references
9633 that don't use SP as a base register.
9635 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9636 "an integer without a # prefix"
9637 * config/msp430/msp430.md (epilogue_helper): Use it.
9639 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9642 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9643 AVX512F gather builtins.
9644 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9645 on gather decls with INTEGER_TYPE masktype.
9646 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9647 directly into the builtin rather than hoisting it before loop.
9649 PR tree-optimization/59387
9650 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9651 (scev_const_prop): If folded_casts and type has undefined overflow,
9652 use force_gimple_operand instead of force_gimple_operand_gsi and
9653 for each added stmt if it is assign with
9654 arith_code_with_undefined_signed_overflow, call
9655 rewrite_to_defined_overflow.
9656 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9657 gimple-fold.h instead.
9658 (arith_code_with_undefined_signed_overflow,
9659 rewrite_to_defined_overflow): Moved to ...
9660 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9661 rewrite_to_defined_overflow): ... here. No longer static.
9662 Include gimplify-me.h.
9663 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9664 rewrite_to_defined_overflow): New prototypes.
9666 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9668 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9670 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9672 * builtins.c (get_object_alignment_2): Minor tweak.
9673 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9675 2014-01-13 Christian Bruel <christian.bruel@st.com>
9677 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9678 optimized non constant lengths.
9680 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9683 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9684 load as __atomic_load_N if possible.
9686 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9688 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9690 (rs6000_expand_builtin): Adjust call.
9692 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9695 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9696 * config/rs6000/rs6000.c: Include target-globals.h.
9697 (rs6000_set_current_function): Instead of doing target_reinit
9698 unconditionally, use save_target_globals_default_opts and
9699 restore_target_globals.
9701 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9703 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9704 (rs6000_expand_builtin): Handle mffs and mtfsf.
9705 (rs6000_init_builtins): Define mffs and mtfsf.
9706 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9707 (rs6000_mffs): New pattern.
9708 (rs6000_mtfsf): New pattern.
9710 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9712 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9713 Start narrowing with START. Apply candidate-use pair
9714 and check overall cost in narrowing.
9715 (iv_ca_prune): Pass new argument.
9717 2014-01-10 Jeff Law <law@redhat.com>
9720 * ree.c (combine_reaching_defs): Ensure the defining statement
9721 occurs before the extension when optimizing extensions with
9722 different source and destination hard registers.
9724 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9727 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9728 vtables into the type inheritance graph.
9730 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9732 PR rtl-optimization/59754
9733 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9734 modes in the REGNO != REGNO case.
9736 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9738 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9740 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9742 PR tree-optimization/59745
9743 * tree-predcom.c (tree_predictive_commoning_loop): Call
9744 free_affine_expand_cache if giving up because components is NULL.
9746 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9747 GC in payload of target_globals struct instead of allocating them on
9748 the heap and the larger structs separately using GC.
9749 * target-globals.h (struct target_globals): Make regs, hard_regs,
9750 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9751 of GTY((skip)) and change type to void *.
9752 (reset_target_globals): Cast loads from those fields to corresponding
9755 2014-01-10 Steve Ellcey <sellcey@mips.com>
9758 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9759 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9760 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9762 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9765 * aarch64-modes.def (CC_Zmode): New flags mode.
9766 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9767 represents an equality.
9768 (aarch64_get_condition_code): Handle CC_Zmode.
9769 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9771 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9773 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9774 extraction in good case.
9776 2014-01-10 Richard Biener <rguenther@suse.de>
9778 PR tree-optimization/59374
9779 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9780 checking after SLP discovery. Mark stmts not participating
9781 in any SLP instance properly.
9783 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9785 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9786 when handling a SET rtx.
9788 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9790 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9791 (cortex-a57): Likewise.
9792 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9794 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9796 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9797 non-iwmmxt builtins.
9799 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9803 * ipa-devirt.c record_target_from_binfo): Take as argument
9804 stack of binfos and lookup matching one for virtual inheritance.
9805 (possible_polymorphic_call_targets_1): Update.
9807 2014-01-10 Huacai Chen <chenhc@lemote.com>
9809 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9810 kernel strings for Loongson-2E/2F/3A.
9812 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9815 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9816 is_gimple_call before calling gimple_call_internal_p.
9818 2014-01-09 Steve Ellcey <sellcey@mips.com>
9820 * Makefile.in (TREE_FLOW_H): Remove.
9821 (TREE_SSA_H): Add file names from tree-flow.h.
9822 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9823 * tree.h: Remove tree-flow.h reference.
9824 * hash-table.h: Remove tree-flow.h reference.
9825 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9826 reference with tree-ssa-loop.h.
9828 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9830 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9831 default element-order behavior for -maltivec.
9832 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9833 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9834 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9835 when targeting big endian, at least for now.
9836 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9838 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9841 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9842 var satisfies use_register_for_decl, just take into account type
9843 alignment, rather than decl alignment.
9845 PR tree-optimization/59622
9846 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9847 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9848 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9849 Don't devirtualize for inplace at all. For targets.length () == 1,
9850 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9852 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9854 * config/i386/i386.md (cpu): Remove the unused btver1.
9856 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9858 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9860 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9863 * tree-core.h (struct target_globals): New forward declaration.
9864 (struct tree_target_option): Add globals field.
9865 * tree.h (TREE_TARGET_GLOBALS): Define.
9866 (prepare_target_option_nodes_for_pch): New prototype.
9867 * target-globals.h (struct target_globals): Define even if
9869 * tree.c (prepare_target_option_node_for_pch,
9870 prepare_target_option_nodes_for_pch): New functions.
9871 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9872 * config/i386/i386.c: Include target-globals.h.
9873 (ix86_set_current_function): Instead of doing target_reinit
9874 unconditionally, use save_target_globals_default_opts and
9875 restore_target_globals.
9877 2014-01-09 Richard Biener <rguenther@suse.de>
9879 PR tree-optimization/59715
9880 * tree-cfg.h (split_critical_edges): Declare.
9881 * tree-cfg.c (split_critical_edges): Export.
9882 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9884 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9886 * cfgexpand.c (expand_stack_vars): Optionally disable
9887 asan stack protection.
9888 (expand_used_vars): Likewise.
9889 (partition_stack_vars): Likewise.
9890 * asan.c (asan_emit_stack_protection): Optionally disable
9891 after return stack usage.
9892 (instrument_derefs): Optionally disable memory access instrumentation.
9893 (instrument_builtin_call): Likewise.
9894 (instrument_strlen_call): Likewise.
9895 (asan_protect_global): Optionally disable global variables protection.
9896 * doc/invoke.texi: Added doc for new options.
9897 * params.def: Added new options.
9898 * params.h: Likewise.
9900 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9902 PR rtl-optimization/59724
9903 * ifcvt.c (cond_exec_process_if_block): Don't call
9904 flow_find_head_matching_sequence with 0 longest_match.
9905 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9906 non-active insns if !stop_after.
9907 (try_head_merge_bb): Revert 2014-01-07 changes.
9909 2014-01-08 Jeff Law <law@redhat.com>
9911 * ree.c (get_sub_rtx): New function, extracted from...
9912 (merge_def_and_ext): Here.
9913 (combine_reaching_defs): Use get_sub_rtx.
9915 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9917 * cgraph.h (varpool_variable_node): Do not choke on null node.
9919 2014-01-08 Catherine Moore <clm@codesourcery.com>
9921 * config/mips/mips.md (simple_return): Attempt to use JRC
9923 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9925 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9927 PR rtl-optimization/59137
9928 * reorg.c (steal_delay_list_from_target): Call update_block for
9930 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9932 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9934 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9935 two duplicate entries.
9937 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9940 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9942 * config/mips/mips.c (mips_truncated_op_cost): New function.
9943 (mips_rtx_costs): Adjust test for BADDU.
9944 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9946 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9948 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9949 (*baddu_si): ...this new pattern.
9951 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9954 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9956 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9959 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9961 (expand_expr, expand_normal): Adjust.
9962 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9963 inner_reference_p. Use inner_reference_p to expand inner references.
9964 (store_expr): Adjust.
9965 * cfgexpand.c (expand_call_stmt): Adjust.
9967 2014-01-08 Rong Xu <xur@google.com>
9969 * gcov-io.c (gcov_var): Move from gcov-io.h.
9970 (gcov_position): Ditto.
9971 (gcov_is_error): Ditto.
9972 (gcov_rewrite): Ditto.
9973 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9974 only part to libgcc/libgcov.h.
9976 2014-01-08 Marek Polacek <polacek@redhat.com>
9979 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9981 2014-01-08 Marek Polacek <polacek@redhat.com>
9984 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9986 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9988 PR rtl-optimization/59649
9989 * stor-layout.c (get_mode_bounds): For BImode return
9990 0 and STORE_FLAG_VALUE.
9992 2014-01-08 Richard Biener <rguenther@suse.de>
9995 * gimple.h (is_gimple_builtin_call): Remove.
9996 (gimple_builtin_call_types_compatible_p): New.
9997 (gimple_call_builtin_p): New overload.
9998 * gimple.c (is_gimple_builtin_call): Remove.
9999 (validate_call): Rename to ...
10000 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10001 check return types.
10002 (validate_type): New static function.
10003 (gimple_call_builtin_p): New overload and adjust.
10004 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10005 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10006 (gimple_fold_stmt_to_constant_1): Likewise.
10007 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10009 2014-01-08 Richard Biener <rguenther@suse.de>
10011 PR middle-end/59471
10012 * gimplify.c (gimplify_expr): Gimplify register-register type
10013 VIEW_CONVERT_EXPRs to separate stmts.
10015 2014-01-07 Jeff Law <law@redhat.com>
10017 PR middle-end/53623
10018 * ree.c (combine_set_extension): Handle case where source
10019 and destination registers in an extension insn are different.
10020 (combine_reaching_defs): Allow source and destination registers
10021 in extension to be different under limited circumstances.
10022 (add_removable_extension): Remove restriction that the
10023 source and destination registers in the extension are the same.
10024 (find_and_remove_re): Emit a copy from the extension's
10025 destination to its source after the defining insn if
10026 the source and destination registers are different.
10028 PR middle-end/59285
10029 * ifcvt.c (merge_if_block): If we are merging a block with more than
10030 one successor with a block with no successors, remove any BARRIER
10031 after the second block.
10033 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10035 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10037 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10040 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10041 for 14-bit register offsets when INT14_OK_STRICT is false.
10043 2014-01-07 Roland Stigge <stigge@antcom.de>
10044 Michael Meissner <meissner@linux.vnet.ibm.com>
10047 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10048 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10050 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10052 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10055 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10057 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10058 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10059 rtx is const0_rtx or not.
10061 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10064 * target-globals.c (save_target_globals): Remove this_fn_optab
10066 * toplev.c: Include optabs.h.
10067 (target_reinit): Temporarily restore the global options if another
10068 set of options are in force.
10070 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10072 PR rtl-optimization/58668
10073 * cfgcleanup.c (flow_find_cross_jump): Don't count
10074 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10075 to determine what is counted.
10076 (flow_find_head_matching_sequence): Use active_insn_p to determine
10078 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10080 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10081 determine what is counted.
10083 PR tree-optimization/59643
10084 * tree-predcom.c (split_data_refs_to_components): If one dr is
10085 read and one write, determine_offset fails and the write isn't
10086 in the bad component, just put the read into the bad component.
10088 2014-01-07 Mike Stump <mikestump@comcast.net>
10089 Jakub Jelinek <jakub@redhat.com>
10092 * tree-core.h (struct tree_optimization_option): Change optabs
10093 type from unsigned char * to void *.
10094 * optabs.c (init_tree_optimization_optabs): Adjust
10095 TREE_OPTIMIZATION_OPTABS initialization.
10097 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10100 * config/i386/i386.h (struct machine_function): Add
10101 no_drap_save_restore field.
10102 * config/i386/i386.c (ix86_save_reg): Use
10103 !cfun->machine->no_drap_save_restore instead of
10104 crtl->stack_realign_needed.
10105 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10106 this function clears frame_pointer_needed. Set
10107 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10108 and DRAP reg is needed.
10110 2014-01-06 Marek Polacek <polacek@redhat.com>
10113 * doc/implement-c.texi: Mention that other integer types are
10114 permitted as bit-field types in strictly conforming mode.
10116 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10118 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10119 is newly allocated.
10121 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10123 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10125 2014-01-06 Martin Jambor <mjambor@suse.cz>
10128 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10130 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10132 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10136 * var-tracking.c (add_stores): Preserve the value of the source even if
10137 we don't record the store.
10139 2014-01-06 Terry Guo <terry.guo@arm.com>
10141 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10143 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10146 * config/darwin.c (darwin_function_section): Adjust return values to
10147 correspond to optimisation changes made in r206070.
10149 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10151 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10152 from prefetch_block tune setting.
10153 (nocona_cost): Correct size of prefetch block to 64.
10155 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10157 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10158 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10159 used to save the static chain register in the computation of the offset
10160 from which the FP registers need to be restored.
10162 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10164 PR tree-optimization/59519
10165 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10166 ICE if get_current_def (current_new_name) is already non-NULL, as long
10167 as it is a phi result of some other phi in *new_exit_bb that has
10170 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10171 or vmovdqu* for misaligned_operand.
10172 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10173 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10174 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10175 aligned_mem for AVX512F masked aligned load and store builtins and for
10176 non-temporal moves.
10178 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10180 PR tree-optimization/59651
10181 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10182 Address range for negative step should be added by TYPE_SIZE_UNIT.
10184 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10186 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10187 overlapping registers also for registers other than the stack pointer.
10189 2014-01-03 Marek Polacek <polacek@redhat.com>
10192 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10195 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10198 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10201 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10202 (push splitter): Use <P:MODE_SIZE> instead of
10203 GET_MODE_SIZE (<P:MODE>mode).
10204 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10205 (mov -1, reg peephole2): Likewise.
10206 * config/i386/sse.md (*mov<mode>_internal,
10207 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10208 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10209 *<code><mode>3, *andnot<mode>3<mask_name>,
10210 <mask_codefor><code><mode>3<mask_name>): Likewise.
10211 * config/i386/subst.md (mask_mode512bit_condition,
10212 sd_mask_mode512bit_condition): Likewise.
10214 2014-01-02 Xinliang David Li <davidxl@google.com>
10216 PR tree-optimization/59303
10217 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10218 (dump_predicates): Better output format.
10219 (pred_equal_p): New function.
10220 (is_neq_relop_p): Ditto.
10221 (is_neq_zero_form_p): Ditto.
10222 (pred_expr_equal_p): Ditto.
10223 (pred_neg_p): Ditto.
10224 (simplify_pred): Ditto.
10225 (simplify_preds_2): Ditto.
10226 (simplify_preds_3): Ditto.
10227 (simplify_preds_4): Ditto.
10228 (simplify_preds): Ditto.
10229 (push_pred): Ditto.
10230 (push_to_worklist): Ditto.
10231 (get_pred_info_from_cmp): Ditto.
10232 (is_degenerated_phi): Ditto.
10233 (normalize_one_pred_1): Ditto.
10234 (normalize_one_pred): Ditto.
10235 (normalize_one_pred_chain): Ditto.
10236 (normalize_preds): Ditto.
10237 (normalize_cond_1): Remove function.
10238 (normalize_cond): Ditto.
10239 (is_gcond_subset_of): Ditto.
10240 (is_subset_of_any): Ditto.
10241 (is_or_set_subset_of): Ditto.
10242 (is_and_set_subset_of): Ditto.
10243 (is_norm_cond_subset_of): Ditto.
10244 (pred_chain_length_cmp): Ditto.
10245 (convert_control_dep_chain_into_preds): Type change.
10246 (find_predicates): Ditto.
10247 (find_def_preds): Ditto.
10248 (destroy_predicates_vecs): Ditto.
10249 (find_matching_predicates_in_rest_chains): Ditto.
10250 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10251 (is_pred_expr_subset): Ditto.
10252 (is_pred_chain_subset_of): Ditto.
10253 (is_included_in): Ditto.
10254 (is_superset_of): Ditto.
10256 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10258 Update copyright years.
10260 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10262 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10263 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10264 config/arc/arc.md, config/arc/arc.opt,
10265 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10266 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10267 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10268 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10269 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10270 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10271 vtable-verify.c, vtable-verify.h: Use the standard form for the
10274 2014-01-02 Tobias Burnus <burnus@net-b.de>
10276 * gcc.c (process_command): Update copyright notice dates.
10277 * gcov-dump.c: Ditto.
10279 * doc/cpp.texi: Bump @copying's copyright year.
10280 * doc/cppinternals.texi: Ditto.
10281 * doc/gcc.texi: Ditto.
10282 * doc/gccint.texi: Ditto.
10283 * doc/gcov.texi: Ditto.
10284 * doc/install.texi: Ditto.
10285 * doc/invoke.texi: Ditto.
10287 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10289 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10291 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10293 * config/i386/sse.md (*mov<mode>_internal): Guard
10294 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10296 PR rtl-optimization/59647
10297 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10298 new_rtx into UNSIGNED_FLOAT rtxes.
10300 Copyright (C) 2014 Free Software Foundation, Inc.
10302 Copying and distribution of this file, with or without modification,
10303 are permitted in any medium without royalty provided the copyright
10304 notice and this notice are preserved.