1 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2 Diego Novillo <dnovillo@google.com>
4 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6 (vld32wh_insn, vld32wl_insn): Delete commented-out old
7 versions of these patterns.
9 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
10 (__builtin_arc_aligned): Likewise.
12 * config/arc/arc.md: Expand adc_0 comment stating the intended
13 purpose and why it isn't ready.
14 Replace commented out call_value_via_label_mixed with a
15 plain comment about bl_s.
17 * config/arc/arc.c (stdio.h): Don't include directly.
18 (arc_frame_info): Remove GTY marker.
19 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
20 Remove if (1) condition.
21 (arc_encode_section_info): Fix comment.
23 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
25 * config/arc/arc.c (arc_conditional_register_usage):
26 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
27 Also set reg_alloc_order for DMA config regs.
29 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
30 Jeremy Bennett <jeremy.bennett@embecosm.com>
32 * doc/install.texi (--with-cpu): Mention ARC.
33 (arc-*-elf32): New paragraph.
34 (arc-linux-uclibc): Likewise.
35 * doc/md.texi (Machine Constraints): Add ARC part.
36 * doc/invoke.texi: (menu): Add ARC Options.
37 (Machine Dependent Options) <ARC Options>: Add synopsis.
38 (node ARC Options): Add.
39 * doc/extend.texi (long_call / short_call attribute): Add ARC.
40 (ARC Built-in Functions): New section defining
41 generic ARC built-in functions.
42 (ARC SIMD Built-in Functions): New section defining SIMD specific
44 (Declaring Attributes of Functions): Extended
45 description of short_call and long_call attributes for ARC and
48 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
49 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
50 Joern Rennecke <joern.rennecke@embecosm.com>
51 Muhammad Khurram Riaz <khurram.riaz@arc.com>
52 Brendan Kehoe <brendan@zen.org>
53 Michael Eager <eager@eagercon.com>
54 Simon Cook <simon.cook@embecosm.com>
55 Jeremy Bennett <jeremy.bennett@embecosm.com>
57 * config/arc, common/config/arc: New directories.
59 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
60 Brendan Kehoe <brendan@zen.org>
61 Simon Cook <simon.cook@embecosm.com>
63 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
65 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
67 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
68 * tree-ssa-coalesce.h: New. Move prototype to here.
69 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
70 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
71 (gimple_can_coalesce_p): Move to...
72 * gimple.c (gimple_can_coalesce_p): Here.
74 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
76 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
77 (dump_decl_set): Move to gimple.c.
78 * gimple.h: Don't include tree-ssa-operands.h.
79 (dump_decl_set): Add prototype.
80 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
82 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
83 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
84 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
85 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
86 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
88 (dump_decl_set): Relocate here.
89 * gimple-ssa.h: New file.
90 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
91 Relocate from gimple.h.
92 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
93 * tree-ssa-operands.c (swap_ssa_operands): Rename from
94 swap_tree_operands and remove non-ssa path.
95 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
96 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
98 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
99 vect_is_simple_reduction_1): Use swap_ssa_operands.
100 * tree-flow.h: Move various prototypes to tree-phinodes.h.
101 (enum need_phi_state): Move to tree-into-ssa.c.
102 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
103 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
104 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
105 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
106 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
107 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
108 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
109 num_imm_uses): Move to ssa-iterators.h.
110 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
111 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
113 (op_iter_done, op_iter_next_def, op_iter_next_tree,
114 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
115 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
116 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
117 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
118 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
119 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
120 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
121 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
122 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
123 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
124 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
125 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
126 (set_phi_nodes): Move to tree-phinodes.h.
127 * tree-ssa-operands.h (enum ssa_op_iter_type,
128 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
129 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
130 (dump_decl_set): Remove prototype.
131 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
132 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
133 (set_phi_nodes): Relocate from tree-flow-inline.h.
134 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
136 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
137 include list. Temporarily add gimple.h to include list.
138 * ssa-iterators.h: New file.
139 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
140 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
141 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
142 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
143 Relocate from tree-ssa-operands.h.
144 (delink_imm_use, link_imm_use_to_list, link_imm_use,
145 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
146 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
147 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
148 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
149 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
150 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
151 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
152 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
153 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
154 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
155 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
156 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
157 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
158 Relocate from tree-flow-inline.h.
159 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
161 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
164 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
165 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
166 Insert '\t' to output template.
167 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
168 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
169 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
170 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
172 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
174 * doc/install.texi (Host/target specific installation notes for GCC):
175 Put @anchor before @heading.
176 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
177 Use @email for email addresses.
179 2013-10-01 Jeff Law <law@redhat.com>
181 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
182 a vec. Only delete the path if we create one without successfully
183 registering a jump thread.
184 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
186 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
187 (paths): New vector of jump threading paths.
188 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
189 (THREAD_PATH): New accessor macro for the entire thread path.
190 (lookup_redirection_data): Get intermediate and final outgoing edge
191 from the thread path.
192 (create_edge_and_update_destination_phis): Copy the threading path.
193 (ssa_fix_duplicate_block_edges): Get edges and block types from the
195 (ssa_redirect_edges): Get edges and block types from the jump threading
196 path. Free the path vector.
197 (thread_block): Get edges from the jump threading path. Look at the
198 entire path to see if we thread to a loop exit. If we cancel a jump
199 thread request, then free the path vector.
200 (thread_single_edge): Get edges and block types from the jump threading
201 path. Free the path vector.
202 (thread_through_loop_header): Get edges and block types from the jump
203 threading path. Free the path vector.
204 (mark_threaded_blocks): Iterate over the vector of paths and store
205 the path on the appropriate edge. Get edges and block types from the
207 (mark_threaded_blocks): Get edges and block types from the jump
208 threading path. Free the path vector.
209 (thread_through_all_blocks): Use the vector of paths rather than
210 a vector of 3-edge sets.
211 (register_jump_thread): Accept pointer to a path vector rather
212 than the path vector itself. Store the path vector for later use.
215 2013-10-01 Jakub Jelinek <jakub@redhat.com>
216 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
219 * config/s390/s390.c (s390_split_branches): Modify check for table
221 (s390_chunkify_start): Rearrange table jump insn check in order to
222 deal with compare and branch insns correctly.
224 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
228 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
229 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
230 define_insn_and_split.
231 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
232 (shiftsi3_compare): New pattern.
234 * config/arm/unspecs.md (UNSPEC_RRX): New.
236 2013-10-01 Alan Modra <amodra@gmail.com>
238 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
239 casing inout operands.
241 2013-10-01 Richard Biener <rguenther@suse.de>
243 PR tree-optimization/58553
244 * tree-loop-distribution.c (struct partition_s): Add niter member.
245 (classify_partition): Populate niter member for the partition
246 and properly identify whether the relevant store happens before
247 or after the loop exit.
248 (generate_memset_builtin): Use niter member from the partition.
249 (generate_memcpy_builtin): Likewise.
251 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
253 * vec.h (vec_prefix, vec): Prefix member names with "m_".
254 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
256 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
258 * basic-block.h (edge_list): Prefix member names with "m_".
259 * context.h (context): Likewise.
260 * domwalk.h (dom_walker): Likewise.
261 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
262 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
263 * hash-table.h (hash_table): Likewise.
264 * machmode.h (bit_field_mode_iterator): Likewise.
265 * pass_manager.h (pass_list): Likewise.
266 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
267 * tree-pass.h (pass_data): Likewise.
268 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
269 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
270 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
271 * asan.c (pass_data_asan): Update accordingly.
272 * cfganal.c (control_dependences::find_control_dependence): Likewise.
273 (control_dependences::control_dependences): Likewise.
274 (control_dependences::~control_dependences): Likewise.
275 (control_dependences::~control_dependences): Likewise.
276 (control_dependences::get_edges_dependent_on): Likewise.
277 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
278 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
279 * context.c (gcc::context::context): Likewise.
280 * cprop.c (pass_rtl_cprop::clone): Likewise.
281 * domwalk.c (dom_walker::walk): Likewise.
282 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
283 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
284 * mode-switching.c (pass_mode_switching::clone): Likewise.
285 * passes.c (opt_pass::opt_pass): Likewise.
286 (pass_manager::pass_manager): Likewise.
287 * predict.c (pass_strip_predict_hints::clone): Likewise.
288 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
289 (pass_split_all_insns::clone): Likewise.
290 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
292 (bit_field_mode_iterator::next_mode): Likewise.
293 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
294 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
295 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
296 * tree-complex.c (pass_lower_complex::clone): Likewise.
297 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
298 * tree-object-size.c (pass_object_sizes::clone): Likewise.
299 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
300 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
301 (pass_fold_builtins::clone): Likewise.
302 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
303 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
304 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
305 (pass_cd_dce::clone): Likewise.
306 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
307 (pass_phi_only_cprop::clone): Likewise.
308 * tree-ssa-dse.c (pass_dse::clone): Likewise.
309 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
310 * tree-ssa-loop.c (pass_lim::clone): Likewise.
311 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
312 * tree-ssa-pre.c (pass_fre::clone): Likewise.
313 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
314 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
315 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
316 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
317 * tree-vrp.c (pass_vrp::clone): Likewise.
318 * tsan.c (pass_tsan::clone): Likewise.
320 2013-09-30 Jakub Jelinek <jakub@redhat.com>
323 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
324 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
327 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
328 optimization, punt if sign_bit_p looked through any zero extension.
330 2013-09-30 Teresa Johnson <tejohnson@google.com>
332 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
333 Update redirected out edge count in joiner case.
334 (ssa_redirect_edges): Common the joiner and non-joiner cases
335 so that joiner case gets profile updates.
337 2013-09-30 Richard Biener <rguenther@suse.de>
339 PR tree-optimization/58554
340 * tree-loop-distribution.c (classify_partition): Require
341 unconditionally executed stores for memcpy and memset recognition.
342 (tree_loop_distribution): Calculate dominance info.
344 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
346 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
347 (NO_PROFILE_COUNTERS): Likewise.
348 (PROFILE_HOOK): Likewise.
349 (FUNCTION_PROFILER): Likewise.
350 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
352 2013-09-30 Iain Sandoe <iain@codesourcery.com>
354 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
355 calls and defines in TARGET_MACHO conditional.
356 (load_macho_picbase_di): Likewise.
357 (reload_macho_picbase): Likewise.
358 (reload_macho_picbase_si): Likewise.
359 (reload_macho_picbase_di): Likewise.
360 (nonlocal_goto_receiver): Likewise.
362 2013-09-30 Nick Clifton <nickc@redhat.com>
364 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
365 that use the MSP430X ISA.
366 (msp430_option_override): Scan -mmcu command line option for any
367 MCU name that supports the MSP430X ISA.
368 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
369 -mmcu options which enable the MSP430X ISA.
371 2013-09-30 Richard Biener <rguenther@suse.de>
374 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
375 before looking for setjmp-like calls.
377 2013-09-29 Iain Sandoe <iain@codesourcery.com>
380 * config/darwin-protos.h (machopic_get_function_picbase): New.
381 * config/darwin.c (machopic_get_function_picbase): New.
382 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
383 label for a new func. (load_macho_picbase_di): Likewise.
384 (reload_macho_picbase): New expand.
385 (reload_macho_picbase_si): New insn.
386 (reload_macho_picbase_di): New insn.
387 (nonlocal_goto_receiver): New define and split.
388 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
389 (unspecv enum): Add UNSPECV_NLGR.
391 2013-09-29 Iain Sandoe <iain@codesourcery.com>
393 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
394 that altivec registers are correctly sized on Darwin.
396 2013-09-29 Iain Sandoe <iain@codesourcery.com>
398 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
399 darwin-driver.o): Use COMPILE and POSTCOMPILE.
400 * config/x-darwin (host-darwin.o): Likewise.
401 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
402 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
403 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
405 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
407 * doc/invoke.texi: Fix usage of @tie{} command.
409 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
411 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
413 2013-09-28 Jan Hubicka <jh@suse.cz>
415 * config/i386/x86-tune.def: Add documentation for each of the options;
418 2013-09-28 Jan Hubicka <jh@suse.cz>
420 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
422 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
423 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
424 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
425 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
426 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
428 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
430 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
431 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
432 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
433 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
434 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
435 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
436 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
437 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
438 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
439 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
440 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
441 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
442 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
443 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
444 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
445 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
446 gimple.h, godump.c, graphite-clast-to-gimple.c,
447 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
448 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
449 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
450 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
451 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
452 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
453 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
454 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
455 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
456 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
457 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
458 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
459 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
460 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
461 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
462 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
463 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
464 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
465 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
466 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
467 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
468 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
469 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
470 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
471 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
472 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
473 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
474 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
475 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
476 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
477 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
478 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
479 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
480 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
481 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
482 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
483 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
484 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
485 whitespace before "(".
487 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
489 * expr.h (extract_bit_field): Remove packedp parameter.
490 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
491 from forward declaration.
492 (store_split_bit_field): Remove packedp arg from calls to
493 extract_fixed_bit_field.
494 (extract_bit_field_1): Remove packedp parameter and packedp
495 argument from recursive calls and calls to extract_fixed_bit_field.
496 (extract_bit_field): Remove packedp parameter and corresponding
497 arg to extract_bit_field_1.
498 (extract_fixed_bit_field): Remove packedp parameter. Remove code
500 (extract_split_bit_field): Remove packedp arg from call to
501 extract_fixed_bit_field.
502 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
503 (copy_blkmode_from_reg): Likewise.
504 (copy_blkmode_to_reg): Likewise.
505 (read_complex_part): Likewise.
506 (store_field): Likewise.
507 (expand_expr_real_1): Likewise.
508 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
509 to extract_bit_field.
510 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
511 call to extract_bit_field.
512 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
513 call to extract_bit_field.
514 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
515 and special packedp behavior from -fstrict-volatile-bitfields
518 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
520 * lra-eliminations.c (init_elim_table): Guard value_p.
522 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
524 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
525 DFmode, DImode, and SFmode in the upper VSX registers based on the
526 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
527 if -mpower8-vector. Combine -mvsx-timode handling with the rest
528 of the VSX register handling.
530 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
532 (zero_extendsidi2_lfiwzx): Add support for loading into the
533 Altivec registers with -mpower8-vector. Use wu/wv constraints to
534 only do VSX memory options on Altivec registers.
535 (extendsidi2_lfiwax): Likewise.
536 (extendsfdf2_fpr): Likewise.
537 (mov<mode>_hardfloat, SF/SD modes): Likewise.
538 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
539 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
540 (movdi_internal64): Likewise.
542 2013-09-27 Xinliang David Li <davidxl@google.com>
544 * opts.c (finish_options): Adjust parameters
545 according to vect cost model.
546 (common_handle_option): Set dynamic vect cost
548 targhooks.c (default_add_stmt_cost): Compute stmt cost
550 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
552 * tree-vectorizer.h (unlimited_cost_model): New function.
553 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
554 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
556 (vect_enhance_data_refs_alignment): Ditto.
557 * flag-types.h: New enum.
558 * common/config/i386/i386-common.c (ix86_option_init_struct):
559 No need to initialize vect_cost_model flag.
560 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
563 2013-09-27 Diego Novillo <dnovillo@google.com>
565 * gimple.h (enum ssa_mode): Remove.
567 2013-09-27 Paulo Matos <pmatos@broadcom.com>
570 * gcc.dg/pr58463.c: New test.
572 2013-09-27 Paulo Matos <pmatos@broadcom.com>
574 * cfgloop.h (number_of_loops): Fix typo in check for null.
576 2013-09-27 Jakub Jelinek <jakub@redhat.com>
579 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
580 are children of outermost saved_cfun's loop, and set it up to
581 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
582 if loop != loop0 and SESE region contains bbs that belong to loop0.
584 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
586 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
587 (binary_scale_code_p, get_base_term, get_index_term): New functions.
588 (set_address_segment, set_address_base, set_address_index)
589 (set_address_disp): Accept the argument unconditionally.
590 (baseness): Remove must_be_base_p and must_be_index_p checks.
591 (decompose_normal_address): Classify as much as possible in the
594 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
596 * cse.c (count_reg_usage): Handle INT_LIST.
597 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
598 * reginfo.c (reg_scan_mark_refs): Likewise.
599 * reload1.c (eliminate_regs_1): Likewise.
601 2013-09-27 Iain Sandoe <iain@codesourcery.com>
604 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
607 2013-09-27 Richard Biener <rguenther@suse.de>
609 PR tree-optimization/58459
610 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
611 restriction not propagating into loops.
613 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
615 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
616 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
617 * doc/tree-ssa.texi (Walking use-def chains): Delete.
619 2013-09-26 Richard Biener <rguenther@suse.de>
621 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
623 2013-09-26 Richard Biener <rguenther@suse.de>
625 * alias.h (component_uses_parent_alias_set): Rename to ...
626 (component_uses_parent_alias_set_from): ... this.
627 * alias.c (component_uses_parent_alias_set): Rename to ...
628 (component_uses_parent_alias_set_from): ... this and return
630 (reference_alias_ptr_type_1): Use the result from
631 component_uses_parent_alias_set_from instead of stripping
632 components one at a time.
633 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
635 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
637 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
638 Move prototypes to...
639 * tree-ssa-ter.h: New File. Move prototypes here.
640 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
641 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
642 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
643 from is_replaceable_p.
644 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
645 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
646 (process_replaceable): Use ter_is_replaceable_p.
647 (find_replaceable_in_bb): Use ter_is_replaceable_p.
648 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
649 newly refactored ssa_is_replaceable_p.
650 * cfgexpand.c: Include tree-outof-ssa.h.
651 * ssaexpand.h: Delete.
653 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
655 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
656 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
658 (create_gimple_tmp): Delete.
659 (get_expr_type, build_assign, build_type_cast): Move to...
660 * gimple-builder.c: New File.
661 (get_expr_type): Relocate from gimple.c.
662 (build_assign, build_type_cast): Change to only create ssanames.
663 * gimple.h: Move prototypes to...
664 * gimple-builder.h: New File. Here.
665 * tree-ssa.h: And here.
666 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
667 count_uses_and_derefs): Relocate from gimple.c.
668 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
669 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
670 * tree-ssa-math-opts (execute_cse_reciprocals): Use
671 gimple_replace_ssa_lhs.
672 * asan.c: Include gimple-builder.h.
673 * Makefile.in: Add gimple-builder.o.
675 2013-09-26 Richard Biener <rguenther@suse.de>
677 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
678 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
679 (loe_visit_block): Use gcc_checking_assert.
680 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
681 gimple_assign_ssa_name_copy_p.
682 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
685 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
687 * config/rs6000/t-rs6000 (rs6000.o): Remove.
688 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
690 2013-09-26 Richard Biener <rguenther@suse.de>
692 PR tree-optimization/58539
693 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
694 the fact that debug statements are not taking part in loop-closed
697 2013-09-26 Nick Clifton <nickc@redhat.com>
699 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
700 time warning message.
701 (msp430_print_operand_raw): Delete unused letter parameter.
702 (TARGET_PRINT_OPERAND_ADDRESS): Define.
703 (msp430_print_operand_address): New function.
704 (msp430_print_operand): Move address printing code from here to
706 * config/msp430/msp430.md (movsipsi2): Add comment in generated
708 (zero_extendpsisi2): Likewise.
709 (extendpsisi2): New pattern.
710 (andneghi3): New pattern.
712 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
714 * config/aarch64/aarch64.opt (mlra): New option.
715 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
716 (TARGET_LRA_P): Define.
718 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
720 * expr.c (expand_assignment): Remove obsolete comment.
722 2013-09-25 Jeff Law <law@redhat.com>
724 * tree-flow.h (thread_through_all_blocks): Prototype moved into
725 tree-ssa-threadupdate.h.
726 (register_jump_thread): Similarly.
727 * tree-ssa-threadupdate.h: New header file.
728 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
729 * tree-vrp.c: Likewise.
730 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
731 (thread_around_empty_blocks): Change type of path vector argument to
732 an edge,type pair from just an edge. Initialize both elements when
733 appending to a jump threading path. Tweak references to elements
735 (thread_across_edge): Similarly. Release memory for the elements
737 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
738 (dump_jump_thread_path): New function broken out from
739 register_jump_thread.
740 (register_jump_thread): Use dump_jump_thread_path. Change type of
741 path vector entries. Search the path for NULL edges and dump
742 the path if one is found. Tweak the conversion of path to 3-edge
743 form to use the block copy type information embedded in the path.
745 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
747 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
749 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
750 Vladimir Makarov <vmakarov@redhat.com>
752 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
753 from the least significant bit.
754 (strip_address_mutations): Add bitfield operations handling.
755 (must_be_index_p): Add shifting and rotate operations handling.
756 (set_address_base): Use must_be_base_p predicate.
757 (set_address_index): Use must_be_index_p predicate.
759 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
760 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
761 Sergey Lega <sergey.s.lega@intel.com>
762 Anna Tikhonova <anna.tikhonova@intel.com>
763 Ilya Tocar <ilya.tocar@intel.com>
764 Andrey Turetskiy <andrey.turetskiy@intel.com>
765 Ilya Verbin <ilya.verbin@intel.com>
766 Kirill Yukhin <kirill.yukhin@intel.com>
767 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
769 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
771 (ix86_expand_vector_move_misalign): Support new unaligned load and
772 stores and use new names.
773 (CODE_FOR_sse2_storedqu): Rename to ...
774 (CODE_FOR_sse2_storedquv16qi): ... this.
775 (CODE_FOR_sse2_loaddqu): Rename to ...
776 (CODE_FOR_sse2_loaddquv16qi): ... this.
777 (CODE_FOR_avx_loaddqu256): Rename to ...
778 (CODE_FOR_avx_loaddquv32qi): ... this.
779 (CODE_FOR_avx_storedqu256): Rename to ...
780 (CODE_FOR_avx_storedquv32qi): ... this.
781 * config/i386/i386.md (fpint_logic): New.
782 * config/i386/sse.md (VMOVE): Extend for AVX512.
786 (VI_UNALIGNED_LOADSTORE): Ditto.
787 (sse2_avx_avx512f): Ditto.
788 (sse2_avx2): Extend for AVX512.
789 (sse4_1_avx2): Ditto.
791 (sse): Extend for AVX512.
794 (avxsizesuffix): Ditto.
795 (sseintvecmode): Ditto.
797 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
798 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
799 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
800 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
801 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
802 (<sse2_avx_avx512f>_storedqu<mode): ... this.
803 (<sse>_movnt<mode>): Replace constraint "x" with "v".
804 (STORENT_MODE): Extend for AVX512.
805 (*absneg<mode>2): Replace constraint "x" with "v".
806 (*mul<mode>3): Ditto.
807 (*ieee_smin<mode>3): Ditto.
808 (*ieee_smax<mode>3): Ditto.
809 (avx_cmp<mode>3): Replace VF with VF_128_256.
810 (*<sse>_maskcmp<mode>3_comm): Ditto.
811 (<sse>_maskcmp<mode>3): Ditto.
812 (<sse>_andnot<mode>3): Extend for AVX512.
813 (<code><mode>3, anylogic): Replace VF with VF_128_256.
814 (<code><mode>3, fpint_logic): New.
815 (*<code><mode>3): Extend for AVX512.
816 (avx512flogicsuff): New.
817 (avx512f_<logic><mode>): Ditto.
818 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
820 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
821 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
822 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
823 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
824 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
825 (xop_vpermil2<mode>3): Ditto.
826 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
827 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
828 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
829 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
831 2013-09-25 Tom Tromey <tromey@redhat.com>
833 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
834 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
835 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
836 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
837 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
838 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
839 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
840 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
841 (GRAPHITE_HTAB_H): Remove.
843 2013-09-25 Tom Tromey <tromey@redhat.com>
845 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
847 2013-09-25 Tom Tromey <tromey@redhat.com>
849 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
851 2013-09-25 Tom Tromey <tromey@redhat.com>
853 * config/i386/t-i386 (i386.o): Remove.
854 (i386-c.o): Use COMPILE and POSTCOMPILE.
856 2013-09-25 Tom Tromey <tromey@redhat.com>
858 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
860 2013-09-25 Tom Tromey <tromey@redhat.com>
862 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
863 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
864 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
865 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
866 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
867 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
868 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
869 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
870 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
871 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
872 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
873 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
874 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
875 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
876 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
877 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
878 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
879 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
880 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
881 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
882 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
883 (tree-ssa-pre.o, tree-ssa-sccvn.o)
884 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
885 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
886 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
887 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
888 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
889 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
890 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
891 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
892 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
893 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
894 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
895 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
896 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
897 (graphite-clast-to-gimple.o, graphite-dependences.o)
898 (graphite-interchange.o, graphite-poly.o)
899 (graphite-scop-detection.o, graphite-sese-to-poly.o)
900 (graphite-optimize-isl.o, tree-vect-loop.o)
901 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
902 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
903 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
904 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
905 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
906 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
907 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
908 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
909 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
910 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
911 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
912 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
913 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
914 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
915 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
916 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
917 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
918 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
919 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
920 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
921 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
922 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
923 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
924 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
925 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
926 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
927 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
928 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
929 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
930 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
931 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
932 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
933 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
934 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
935 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
936 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
937 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
938 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
939 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
940 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
941 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
942 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
943 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
944 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
945 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
946 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
947 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
948 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
949 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
950 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
951 (gcov-dump.o): Remove.
952 (default-c.o): Use COMPILE and POSTCOMPILE.
953 (CFLAGS-gcc.o): New variable.
954 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
956 2013-09-25 Tom Tromey <tromey@redhat.com>
958 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
959 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
960 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
961 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
962 (c-family/c-pragma.o, c-family/c-pretty-print.o)
963 (c-family/c-semantics.o, c-family/c-ada-spec.o)
964 (c-family/array-notation-common.o, c-family/stub-objc.o)
965 (c-family/c-ubsan.o): Remove.
967 2013-09-25 Tom Tromey <tromey@redhat.com>
969 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
971 2013-09-25 Tom Tromey <tromey@redhat.com>
973 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
974 to add -DENABLE_SHARED_LIBGCC.
975 (gcc.o): Don't use subshell.
977 2013-09-25 Tom Tromey <tromey@redhat.com>
979 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
980 * configure.ac: Don't invoke AM_PROG_CC_C_O.
981 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
982 * configure, config.in: Rebuild.
984 2013-09-25 Tom Tromey <tromey@redhat.com>
986 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
987 (COMPILE, POSTCOMPILE): New variables.
988 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
989 (DEPFILES): New variable.
991 * configure.ac: Add checks for dependency checking.
992 * configure, aclocal.m4: Regenerate.
994 2013-09-25 Tom Tromey <tromey@redhat.com>
996 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
997 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
999 2013-09-25 Tom Tromey <tromey@redhat.com>
1001 * Makefile.in (generated_files): Add options.h,
1002 target-hooks-def.h, insn-opinit.h,
1003 common/common-target-hooks-def.h, pass-instances.def,
1004 c-family/c-target-hooks-def.h.
1006 2013-09-25 Jeff Law <law@redhat.com>
1008 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
1009 than foo[foo.length () - 1] to access last member in a vec.
1010 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
1012 2013-09-25 Richard Biener <rguenther@suse.de>
1015 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1017 2013-09-25 Jan Hubicka <jh@suse.cz>
1019 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1022 2013-09-25 Marek Polacek <polacek@redhat.com>
1025 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1026 when determining the type name.
1028 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1030 * config/sh/sh.md: Fix formatting.
1032 2013-09-24 Xinliang David Li <davidxl@google.com>
1034 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1035 max peel iterations parameter.
1036 * param.def: New parameter.
1037 * doc/invoke.texi: Document New parameter.
1039 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1041 * gimple-pretty-print.c: Various whitespace tweaks.
1042 * tree-core.h: Likewise.
1043 * tree-pretty-print.c: Likewise.
1044 * tree-ssa-alias.c: Likewise.
1045 * tree-ssa-copy.c: Likewise.
1046 * tree-ssanames.c: Likewise.
1047 * tree-ssanames.h: Likewise.
1048 * tree-vrp.c: Likewise.
1050 2013-09-24 Alan Modra <amodra@gmail.com>
1054 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1055 for output operands that disallow regs. Don't use EXPAND_WRITE on
1058 2013-09-24 Richard Biener <rguenther@suse.de>
1061 * tree.c (reference_alias_ptr_type): Move ...
1062 * alias.c (reference_alias_ptr_type): ... here and implement
1063 in terms of the new reference_alias_ptr_type_1.
1064 (ref_all_alias_ptr_type_p): New helper.
1065 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1066 use ref_all_alias_ptr_type_p.
1067 (get_deref_alias_set): Add flag_strict_aliasing check here.
1068 (reference_alias_ptr_type_1): New function, split out from ...
1069 (get_alias_set): ... here.
1070 (alias_ptr_types_compatible_p): New function.
1071 * alias.h (reference_alias_ptr_type): Declare.
1072 (alias_ptr_types_compatible_p): Likewise.
1073 * tree.h (reference_alias_ptr_type): Remove.
1074 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1075 to compare MEM_REF alias types.
1077 2013-09-24 Richard Biener <rguenther@suse.de>
1079 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1081 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1083 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1084 reload helper function arrays into a single array reg_addr.
1085 (reload_fpr_gpr): Likewise.
1086 (reload_gpr_vsx): Likewise.
1087 (reload_vsx_gpr): Likewise.
1088 (struct rs6000_reg_addr): Likewise.
1089 (reg_addr): Likewise.
1090 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1091 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1092 (rs6000_init_hard_regno_mode_ok): Likewise.
1093 (rs6000_secondary_reload_direct_move): Likewise.
1094 (rs6000_secondary_reload): Likewise.
1096 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1097 constraints: wu, ww, and wy. Repurpose wv constraint added during
1098 power8 changes. Put wg constraint in alphabetical order.
1100 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1101 for future work to add ISA 2.07 VSX single precision support.
1102 (-mvsx-scalar-double): Change default from -1 to 1, update
1103 documentation comment.
1104 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1105 (-mupper-regs-df): New debug switch to control whether DF values
1106 can go in the traditional Altivec registers.
1107 (-mupper-regs-sf): New debug switch to control whether SF values
1108 can go in the traditional Altivec registers.
1110 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1112 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1113 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1114 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1115 (rs6000_debug_legitimate_address_p): Print if we are running
1116 before, during, or after reload.
1117 (rs6000_secondary_reload): Add a comment.
1118 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1120 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1121 constraints. Update documentation string.
1122 (wd constraint): Likewise.
1123 (wf constraint): Likewise.
1124 (wg constraint): Likewise.
1125 (wn constraint): Likewise.
1126 (ws constraint): Likewise.
1127 (wt constraint): Likewise.
1128 (wx constraint): Likewise.
1129 (wz constraint): Likewise.
1130 (wu constraint): New constraint for ISA 2.07 SFmode scalar
1132 (ww constraint): Likewise.
1133 (wy constraint): Likewise.
1134 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
1135 the previous submissions.
1136 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
1138 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
1140 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
1141 (REG_BR_PROB): Say that the probability is stored in an int_list.
1142 * reg-notes.def: Update commentary to mention INT_LIST.
1143 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
1144 (INT_LIST): New rtx.
1145 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
1146 * rtlanal.c (int_reg_note_p): New function.
1147 (alloc_reg_note): Assert that the note does not have an int argument.
1148 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
1149 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
1150 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
1151 rather than an INSN_LIST. Handle INT_LIST.
1152 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
1153 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
1154 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
1155 Manipulate them as ints rather than rtxes.
1156 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
1157 * regmove.c (copy_src_to_dest): Likewise.
1158 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
1160 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
1161 into the cases that need it.
1162 * config/arm/arm.c (arm_unwind_emit): Likewise.
1164 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
1165 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
1166 * loop-doloop.c (add_test, doloop_modify): Likewise.
1167 * loop-unswitch.c (compare_and_jump_seq): Likewise.
1168 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
1169 * predict.c (combine_predictions_for_insn): Likewise.
1170 * print-rtl.c (print_rtx): Handle INT_LIST.
1171 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
1172 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
1173 * config/arm/arm.c (emit_unlikely_jump): Likewise.
1174 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
1175 (ix86_split_fp_branch, predict_jump): Likewise.
1176 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
1177 * config/sh/sh.c (expand_cbranchsi4): Likewise.
1178 * config/spu/spu.c (ea_load_store_inline): Likewise.
1180 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
1181 value of a REG_BR_PROB note.
1182 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
1183 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
1184 * emit-rtl.c (try_split): Likewise.
1185 * predict.c (br_prob_note_reliable_p): Likewise.
1186 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
1187 * reorg.c (mostly_true_jump): Likewise.
1188 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
1189 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
1190 * config/i386/i386.c (ix86_print_operand): Likewise.
1191 * config/ia64/ia64.c (ia64_print_operand): Likewise.
1192 * config/mmix/mmix.c (mmix_print_operand): Likewise.
1193 * config/rs6000/rs6000.c (output_cbranch): Likewise.
1194 * config/s390/s390.c (s390_expand_tbegin): Likewise.
1195 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
1196 * config/sparc/sparc.c (output_cbranch): Likewise.
1197 * config/spu/spu.c (get_branch_target): Likewise.
1198 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
1199 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
1201 2013-09-23 Jan Hubicka <jh@suse.cz>
1203 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
1205 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
1206 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
1208 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
1209 (gimple_fold_call): Dump inconsistent devirtualizations; add
1210 sanity check for type based devirtualizations.
1211 * ipa-prop.c: Include ipa-utils.h
1212 (ipa_intraprocedural_devirtualization): Add sanity check.
1213 (try_make_edge_direct_virtual_call): Likewise.
1215 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
1217 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
1218 assignment statements.
1220 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1222 * gimple-pretty-print.c (dump_ssaname_info): New function.
1223 (dump_gimple_phi): Call it.
1224 (pp_gimple_stmt_1): Likewise.
1225 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
1226 (range_info_def): Declare.
1227 * tree-pretty-print.c (pp_double_int): New function.
1228 (dump_generic_node): Call it.
1229 * tree-pretty-print.h (pp_double_int): Declare.
1230 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
1231 * tree-ssanames.h (range_info_def): New structure.
1232 (value_range_type): Move definition here.
1233 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
1235 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
1237 (set_range_info): New function.
1238 (get_range_info): Likewise.
1239 (duplicate_ssa_name_range_info): Likewise.
1240 (duplicate_ssa_name_fn): Check pointer type and call
1241 duplicate_ssa_name_range_info.
1242 * tree-ssa-copy.c (fini_copy_prop): Likewise.
1243 * tree-vrp.c (value_range_type): Remove definition, now in
1245 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
1246 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
1247 (SSA_NAME_RANGE_INFO): New macro.
1249 2013-09-23 Richard Biener <rguenther@suse.de>
1251 PR tree-optimization/58464
1252 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1253 (phi_trans_add): Change to add conditionally on being not present.
1254 (phi_translate_1): Remove recursion detection here.
1255 (phi_translate): Pre-seed the cache with NULL to catch
1256 recursion here in a more generic way.
1257 (bitmap_find_leader): Adjust comment.
1258 (get_representative_for): Dump value-numbers.
1259 (create_expression_by_pieces): Likewise.
1260 (insert_into_preds_of_block): Likewise.
1262 2013-09-23 Christian Bruel <christian.bruel@st.com>
1265 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1266 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1268 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
1271 2013-09-20 Renlin Li <renlin.li@arm.com>
1273 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1275 (aarch64_expand_epilogue): Likewise.
1276 (aarch64_legitimize_reload_address): Likewise.
1278 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
1280 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
1281 NULL_TREE before pushing them onto the vector. Likewise for labels.
1283 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
1285 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
1286 comparison operators when -fno-trapping-math is in effect.
1287 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
1288 comparison operators in TFmode and assert that unsupported operators
1290 (ia64_print_operand): Likewise.
1292 2013-09-21 Jan Hubicka <jh@suse.cz>
1294 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
1295 (sse_typeless_stores): Enable for core
1296 (sse_load0_by_pxor): Likewise.
1297 (four_jump_limit): Disable for core.
1298 (pad_returns): Likewise.
1299 (avoid_vector_decode): Likewise.
1300 (fuse_cmp_and_branch): Enable for cores.
1301 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
1303 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1306 * config/pa/pa.c (pa_option_override): Disable auto increment and
1307 decrement instructions until reload is completed.
1309 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
1310 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
1311 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
1313 2013-09-20 DJ Delorie <dj@redhat.com>
1314 Nick Clifton <nickc@redhat.com>
1316 * config/rl78/rl78.c: Various whitespace and comment tweaks.
1317 (need_to_save): Save bank 0 on interrupts.
1318 (characterize_address): Strip far address wrappers.
1319 (rl78_as_legitimate_address): Likewise.
1320 (transcode_memory_rtx): Likewise.
1321 (rl78_peep_movhi_p): Disable this peephole after devirt.
1322 (rl78_propogate_register_origins): Forget all origins when a
1324 * config/rl78/rl78-virt.md: Various whitespace tweaks.
1325 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
1327 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
1328 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
1329 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
1331 * config/rl78/constraints.md: Various whitespace and paren tweaks.
1333 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1335 * config/pa/pa.md: In "scc" insn patterns, change output template to
1336 handle const0_rtx in reg_or_0_operand operands.
1338 2013-09-20 Martin Husemann <martin@NetBSD.org>
1341 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
1342 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
1344 2013-09-20 Richard Biener <rguenther@suse.de>
1347 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
1348 remembering SSA name version and block index.
1349 (new_scev_info_str): Adjust.
1350 (hash_scev_info): Likewise. Also hash the block index.
1351 (eq_scev_info): Adjust.
1352 (find_var_scev_info): Likewise.
1353 (struct instantiate_cache_entry): Remove.
1354 (struct instantiate_cache_type): Use a htab to map name, block
1356 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1357 (get_instantiated_value_entry): Likewise.
1358 (hash_idx_scev_info, eq_idx_scev_info): New functions.
1359 (instantiate_scev_name): Adjust.
1361 2013-09-20 Jeff Law <law@redhat.com>
1363 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
1365 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
1367 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
1368 Call aarch64_simd_expand_args to update op[argc].
1370 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1372 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
1375 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1377 * gengtype.c (file_rules): Added rule for *.cc files.
1378 (get_output_file_with_visibility): Give fatal message when no
1381 2013-09-20 Renlin Li <renlin.li@arm.com>
1383 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
1384 (aarch64_expand_epilogue): Likewise.
1385 (aarch64_legitimize_reload_address): Likewise.
1387 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
1390 * expr.c (expand_assignment): Remove misalignp code path.
1392 2013-09-20 Marek Polacek <polacek@redhat.com>
1395 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
1396 TYPE_PRECISION. Add asserts.
1398 2013-09-20 Richard Biener <rguenther@suse.de>
1400 PR tree-optimization/58453
1401 * tree-loop-distribution.c (distribute_loop): Apply the cost
1402 model for -ftree-loop-distribute-patterns, too.
1404 2013-09-20 Richard Biener <rguenther@suse.de>
1407 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
1408 make type comparison less strict.
1410 2013-09-20 Alan Modra <amodra@gmail.com>
1412 * configure: Regenerate.
1413 * aclocal.m4: Regenerate.
1415 2013-09-20 Marek Polacek <polacek@redhat.com>
1418 * doc/extend.texi: Document that attribute used is meant to be used
1419 on variables with static storage duration.
1421 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1423 PR tree-optimization/58472
1424 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
1425 simd_lane_access set inv_p = false.
1426 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
1427 the simduid magic VAR_DECL.
1429 2013-09-19 Jan Hubicka <jh@suse.cz>
1431 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
1433 2013-09-17 Jeff Law <law@redhat.com>
1435 * tree-ssa-dom.c (record_temporary_equivalences): New function
1436 split out of dom_opt_dom_walker::after_dom_children.
1437 (dom_opt_dom_walker::thread_across_edge): Move common code
1438 in here from dom_opt_dom_walker::after_dom_children.
1439 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
1441 2013-09-19 Jan Hubicka <jh@suse.cz>
1443 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
1444 (TARGET_GENERIC): Use PROCESOR_GENERIC
1445 (enum processor_type): Unify generic32 and 64.
1446 * i386.md (cpu): Likewise.
1447 * x86-tune.def (use_leave): Enable for generic32.
1448 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
1449 * athlon.md: Change generic64 to generic in all occurences.
1450 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
1451 (ix86_target_macros_internal): Likewise.
1452 * driver-i386.c (host_detect_local_cpu): Likewise.
1453 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
1455 (generic_memcpy, generic_memset, generic_cost): This one.
1456 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
1457 (m_GENERIC32, m_GENERIC64): Remove.
1458 (m_GENERIC): Turn into one flag.
1459 (processor_target): Unify generic tunnings.
1460 (ix86_option_override_internal): Replace generic32/64 by generic.
1461 (ix86_issue_rate): Likewise.
1462 (ix86_adjust_cost): Likewise.
1464 2013-09-19 Jan Hubicka <jh@suse.cz>
1466 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
1467 of speculative flag.
1469 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1471 * omp-low.c (expand_omp_sections): Always pass len - 1 to
1472 GOMP_sections_start, even if !exit_reachable.
1474 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1476 * lra-constraints.c (need_for_all_save_p): Use macro
1477 HARD_REGNO_CALL_PART_CLOBBERED.
1478 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
1479 set up pseudo conflict hard regs.
1481 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1484 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
1487 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1489 PR rtl-optimization/58438
1490 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
1491 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
1492 from previous subpasses.
1494 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
1496 * arm.c (arm_get_frame_offsets): Validate architecture supports
1497 LDRD/STRD before accepting the tuning preference.
1498 (arm_expand_prologue): Likewise.
1499 (arm_expand_epilogue): Likewise.
1501 2013-09-18 Richard Biener <rguenther@suse.de>
1503 PR tree-optimization/58417
1504 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
1505 have chrecs with symbols defined in the loop as operands.
1506 (chrec_fold_multiply): Likewise.
1507 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
1508 parameters before folding binary operations.
1509 (struct instantiate_cache_entry_hasher): Remove.
1510 (struct instantiate_cache_type): Use a pointer-map.
1511 (instantiate_cache_type::instantiate_cache_type): New function.
1512 (instantiate_cache_type::get): Likewise.
1513 (instantiate_cache_type::set): Likewise.
1514 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1515 (get_instantiated_value_entry): Likewise.
1516 (global_cache): New global.
1517 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
1518 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
1519 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
1520 (instantiate_scev_name): Adjust.
1521 (instantiate_scev): Construct global instead of local cache.
1522 (resolve_mixers): Likewise.
1524 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
1525 Paolo Carlini <paolo.carlini@oracle.com>
1528 * doc/implement-cxx.texi: Fix references to the C++ standards.
1530 2013-09-18 Jakub Jelinek <jakub@redhat.com>
1532 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
1533 * tree-vect-data-refs.c (vect_analyze_data_refs): For
1534 simd_lane_access drs, update also DR_ALIGNED_TO.
1536 2013-09-18 Marek Polacek <polacek@redhat.com>
1539 * doc/extend.texi: Document no_sanitize_undefined attribute.
1540 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1541 no_sanitize_undefined attribute.
1543 2013-09-18 Nick Clifton <nickc@redhat.com>
1545 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1546 (ASM_DECLARE_FUNCTION_NAME): Define.
1548 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1550 * compare-elim.c (find_comparison_dom_walker): New class
1551 (find_comparisons_in_bb): Rename to
1552 find_comparison_dom_walker::before_dom_children
1553 (find_comparisons): Adjust
1554 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1556 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1557 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1558 (init_walk_dominator_tree): Remove declaration.
1559 (fini_walk_dominator_tree): Remove declaration.
1560 * fwprop.c (single_def_use_dom_walker): New class
1561 (single_def_use_enter_block): Convert to
1562 single_def_use_dom_walker::before_dom_children.
1563 (single_def_use_leave_block): Convert to
1564 single_def_use_dom_walker::after_dom_children.
1565 (build_single_def_use_links): Adjust.
1566 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1568 (find_candidates_in_block): Convert to
1569 find_candidates_dom_walker::before_dom_children.
1570 (execute_strength_reduction): Adjust.
1571 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1572 (sese_dom_walker): New class.
1573 (sese_dom_walker::sese_dom_walker): New constructor.
1574 (sese_dom_walker::~sese_dom_walker): New destructor.
1575 (build_sese_conditions_before): Convert to
1576 sese_dom_walker::before_dom_children.
1577 (build_sese_conditions_after): Convert to
1578 sese_dom_walker::after_dom_children.
1579 (build_poly_scop): Adjust
1580 * tree-into-ssa.c (rewrite_dom_walker): New class
1581 (rewrite_enter_block): Convert to
1582 rewrite_dom_walker::before_dom_children.
1583 (rewrite_leave_block): Convert to
1584 rewrite_dom_walker::after_dom_children.
1585 (rewrite_update_dom_walker): New class.
1586 (rewrite_update_enter_block): Convert to
1587 rewrite_update_dom_walker::before_dom_children.
1588 (rewrite_update_leave_block): Convert to
1589 rewrite_update_dom_walker::after_dom_children.
1590 (rewrite_blocks, rewrite_into_ssa): Adjust.
1591 (mark_def_dom_walker): New class.
1592 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1593 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1594 (mark_def_sites_blocks): Convert to
1595 mark_def_dom_walker::before_dom_children.
1596 (mark_def_site_blocks): Remove.
1597 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1598 (tree_ssa_dominator_optimize): Adjust.
1599 (dom_thread_across_edge): Convert to method
1600 dom_opt_dom_walker::thread_across_edge.
1601 (dom_opt_enter_block): Convert to member function
1602 dom_opt_dom_walker::before_dom_children.
1603 (dom_opt_leave_block): Convert to member function
1604 dom_opt_dom_walker::after_dom_children.
1605 * tree-ssa-dse.c (dse_dom_walker): New class.
1606 (dse_enter_block): Convert to member function
1607 dse_dom_walker::before_dom_children.
1608 (tree_ssa_dse): Adjust.
1609 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1610 (determine_invariantness_stmt): Convert to method
1611 invariantness_dom_walker::before_dom_children.
1612 (determine_invariantness): Remove
1613 (move_computations_dom_walker): New class.
1614 (move_computations_stmt): Convert to method
1615 move_computations_dom_walker::before_dom_children.
1616 (move_computations, tree_ssa_lim): Adjust.
1617 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1618 (nt_init_block): Make method
1619 notrappping_dom_walker::before_dom_children.
1620 (nt_fini_block): Make
1621 method nontrapping_dom_walker::after_dom_children.
1622 (get_non_trapping): Adjust.
1623 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1624 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1625 (eliminate_leave_block): Make method.
1626 eliminate_dom_walker::after_dom_children.
1628 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1629 (strlen_enter_block): Make method
1630 strlen_dom_walker::before_dom_children.
1631 (strlen_leave_block): Make
1632 method strlen_dom_walker::after_dom_children.
1633 (tree_ssa_strlen): Adjust.
1634 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1635 (tree_ssa_uncprop): Adjust.
1636 (uncprop_leave_block): Make method
1637 uncprop_dom_walker::after_dom_children.
1638 (uncprop_leave_block): Make method
1639 uncprop_dom_walker::before_dom_children.
1641 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1643 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
1644 branch in same basic block. Check both src and dest of the move insn.
1646 2013-09-17 Nick Clifton <nickc@redhat.com>
1648 * config/rl78/rl78-real.md (bf): New pattern.
1650 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1651 (rl78_print_operand): Do not put a # before a %B.
1652 * config/rl78/rl78.opt: Tweak doc strings.
1654 2013-09-17 DJ Delorie <dj@redhat.com>
1656 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1657 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1658 (rl78_option_override): Likewise, if -mallregs.
1659 (is_virtual_register): Likewise.
1660 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1661 (REGNO_OK_FOR_BASE_P): Likewise.
1662 * config/rl78/rl78.opt (-mallregs): New.
1664 2013-09-17 Nick Clifton <nickc@redhat.com>
1666 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1667 For interrupt functions: save all call clobbered registers if the
1668 interrupt handler is not a leaf function.
1669 (rl78_expand_prologue): Always recompute the frame information.
1670 For interrupt functions: only select bank 0 if one of the bank 0
1671 registers is going to be psuhed.
1673 2013-09-17 DJ Delorie <dj@redhat.com>
1675 * config/rl78/constraints.md: For each W* constraint, rename to C*
1676 and create a W* constraint that checks for an optional ES: prefix
1678 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1679 (es_addr): New. Used to wrap far addresses.
1680 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1681 (rl78_es_base): New.
1682 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1683 wrapped far addresses.
1684 (rl78_print_operand_1): Unwrap far addresses before processing.
1685 (rl78_lo16): Wrap far addresses in unspecs.
1686 (rl78_es_addr): New.
1687 (rl78_es_base): New.
1688 (insn_ok_now): Check for not-yet-wrapped far addresses.
1689 (transcode_memory_rtx): Properly re-wrap far addresses.
1691 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1693 * config/sparc/t-rtems: Add leon3 multilibs.
1695 2013-09-17 Cong Hou <congh@google.com>
1697 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1698 when checking the dot production pattern. The type of rhs operand
1699 of multiply is now checked correctly.
1701 2013-09-17 Jeff Law <law@redhat.com>
1703 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1704 edge implied equivalences into successor phis.
1705 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1706 here from tree-ssa-threadedge.c.
1707 (mark_threaded_blocks): When threading through a joiner, if both
1708 successors of the joiner's clone reach the same block, verify the
1709 PHI arguments are equal. If not, cancel the jump threading request.
1710 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1711 tree-ssa-threadupdate.c
1712 (thread_across_edge): Don't check PHI argument equality when
1713 threading through joiner block here.
1715 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1717 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1718 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1719 (warn_uninit, warn_uninitialized_vars,
1720 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1721 Move to tree-ssa-uninit.c.
1722 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1723 (has_undefined_value_p): New. Pass dependant parts of
1724 ssa_undefined_value_p.
1725 (uninit_undefined_value_p): Use has_undefined_value_p.
1726 (warn_uninit, warn_uninitialized_vars,
1727 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1728 Move from tree-ssa.c.
1729 * tree-ssa.h: Adjust prototypes.
1731 2013-09-17 Jan Hubicka <jh@suse.cz>
1734 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1735 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1736 FUNCTION_NOT_OPTIMIZED.
1737 * ipa-inline-analysis.c (compute_inline_parameters): Function
1738 not optimized is not inlinable unless it is alwaysinline.
1739 (inline_analyze_function): Force calls in not optimized
1740 function not inlinable.
1742 2013-09-17 Jan Hubicka <jh@suse.cz>
1745 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1747 * ipa.c (function_and_variable_visibility): Likewise.
1748 * ipa-profile.c (ipa_profile): Likewise.
1750 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1753 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1754 attribute, and don't inline the resolver in this case.
1756 2013-09-17 Teresa Johnson <tejohnson@google.com>
1758 * coverage.c (get_coverage_counts): Add missing newline.
1760 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1762 PR tree-optimization/58088
1763 * fold-const.c (mask_with_trailing_zeros): New function.
1764 (fold_binary_loc): Make sure we don't recurse infinitely
1765 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1766 Use mask_with_trailing_zeros where appropriate.
1768 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1770 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1771 of 'prev' var to get better distance estimation.
1773 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1775 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1776 * tree-inline.c (is_parameter_of): New predicate.
1777 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1778 a parameter has been remapped.
1779 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1780 (optimize_inline_calls): Initialize transform_parameter.
1781 (copy_gimple_seq_and_replace_locals): Likewise.
1782 (tree_function_versioning): Likewise.
1783 (maybe_inline_call_in_expr): Likewise.
1785 2013-09-17 Nick Clifton <nickc@redhat.com>
1787 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1788 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1790 (is_attr_func): New function.
1791 (msp430_is_interrupt_func): New function.
1792 (is_naked_func): New function.
1793 (is_reentrant_func): New function.
1794 (is_critical_func): New function.
1795 (msp430_start_function): Add annotations for function attributes.
1796 (msp430_attr): New function.
1797 (msp430_attribute_table): New.
1798 (msp430_function_section): New function.
1799 (TARGET_ASM_FUNCTION_SECTION): Define.
1800 (msp430_builtin): New enum.
1801 (msp430_init_builtins): New function.
1802 (msp430_builtin_devl): New function.
1803 (msp430_expand_builtin): New function.
1804 (TARGET_INIT_BUILTINS): Define.
1805 (TARGET_EXPAND_BUILTINS): Define.
1806 (TARGET_BUILTIN_DECL): Define.
1807 (msp430_expand_prologue): Add support for naked, interrupt,
1808 critical and reentrant functions.
1809 (msp430_expand_epilogue): Likewise.
1810 (msp430_print_operand): Handle 'O' character.
1811 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1813 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1814 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1815 (pushm): Use a 'n' rather than an 'i' constraint.
1816 (msp_return): Add generation of the interrupt return instruction.
1817 (disable_interrupts): New pattern.
1818 (enable_interrupts): New pattern.
1819 (push_intr_state): New pattern.
1820 (pop_intr_state): New pattern.
1821 (bic_SR): New pattern.
1822 (bis_SR): New pattern.
1823 * doc/extend.texi: Document MSP430 function attributes and builtin
1826 2013-09-17 Richard Biener <rguenther@suse.de>
1828 PR tree-optimization/58432
1829 * tree-loop-distribution.c (tree_loop_distribution): Also
1830 scan PHIs for outside loop uses and seed a partition from them.
1832 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1834 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1835 (restructure_reference): Call backtrace_base_for_ref.
1837 2013-09-17 Alan Modra <amodra@gmail.com>
1840 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
1843 2013-09-16 DJ Delorie <dj@redhat.com>
1845 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1846 vregs location for RL78/G10.
1847 (rl78_expand_prologue): Avoid SEL on G10.
1848 (rl78_expand_epilogue): Likewise.
1849 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1850 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1851 __RL78_G10__ when appropriate.
1852 (ASM_SPEC): Pass -mg10 along to the assembler.
1853 * config/rl78/rl78.md (sel_rb): Disable for G10.
1854 * config/rl78/rl78.opt: Add -mg10 option.
1855 * config/rl78/t-rl78: Add -mg10 multilib.
1857 2013-09-16 Xinliang David Li <davidxl@google.com>
1859 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1860 * omp-low.c (omp_max_vf): Ditto.
1861 (expand_omp_simd): Ditto.
1862 * tree-vectorizer.c (vectorize_loops): Ditto.
1863 (gate_vect_slp): Ditto.
1864 (gate_increase_alignment): Ditto.
1865 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1866 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1867 (gate_tree_vectorize): Name change.
1868 (tree_vectorize): Ditto.
1869 (pass_vectorize::gate): Call new function.
1870 (pass_vectorize::execute): Ditto.
1871 * opts.c: O3 default setting change.
1872 (finish_options): Check new flag.
1873 * doc/invoke.texi: Document new flags.
1874 * common.opt: New flags.
1876 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1878 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1879 (Condition Code): Don't reference it.
1880 * doc/tm.texi: Regenerate.
1882 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1885 * lra-constraints.c (undo_optional_reloads): Consider all optional
1886 reload even if it did not get a hard reg.
1888 2013-09-16 Teresa Johnson <tejohnson@google.com>
1890 * dumpfile.c (dump_loc): Remove newline emission.
1891 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1892 emission to dump_printf_loc calls where missing.
1893 (vect_mark_for_runtime_alias_test): Ditto.
1894 (vect_analyze_data_ref_dependence): Ditto.
1895 (vect_analyze_data_ref_dependences): Ditto.
1896 (vect_slp_analyze_data_ref_dependence): Ditto.
1897 (vect_slp_analyze_data_ref_dependences): Ditto.
1898 (vect_compute_data_ref_alignment): Ditto.
1899 (vect_update_misalignment_for_peel): Ditto.
1900 (vect_verify_datarefs_alignment): Ditto.
1901 (vector_alignment_reachable_p): Ditto.
1902 (vect_get_data_access_cost): Ditto.
1903 (vect_enhance_data_refs_alignment): Ditto.
1904 (vect_find_same_alignment_drs): Ditto.
1905 (vect_analyze_data_refs_alignment): Ditto.
1906 (vect_analyze_group_access): Ditto.
1907 (vect_analyze_data_ref_access): Ditto.
1908 (vect_analyze_data_ref_accesses): Ditto.
1909 (vect_prune_runtime_alias_test_list): Ditto.
1910 (vect_analyze_data_refs): Ditto.
1911 (vect_create_addr_base_for_vector_ref): Ditto.
1912 (vect_create_data_ref_ptr): Ditto.
1913 (vect_grouped_store_supported): Ditto.
1914 (vect_grouped_load_supported): Ditto.
1915 * value-prof.c (check_counter): Ditto.
1916 (check_ic_target): Ditto.
1917 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1918 (vect_recog_widen_mult_pattern): Ditto.
1919 (vect_recog_widen_sum_pattern): Ditto.
1920 (vect_recog_over_widening_pattern): Ditto.
1921 (vect_recog_widen_shift_pattern): Ditto.
1922 (vect_recog_rotate_pattern): Ditto.
1923 (vect_recog_vector_vector_shift_pattern): Ditto.
1924 (vect_recog_divmod_pattern): Ditto.
1925 (vect_recog_mixed_size_cond_pattern): Ditto.
1926 (vect_recog_bool_pattern): Ditto.
1927 (vect_pattern_recog_1): Ditto.
1928 (vect_pattern_recog): Ditto.
1929 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1930 (vect_is_simple_iv_evolution): Ditto.
1931 (vect_analyze_scalar_cycles_1): Ditto.
1932 (vect_get_loop_niters): Ditto.
1933 (vect_analyze_loop_1): Ditto.
1934 (vect_analyze_loop_form): Ditto.
1935 (vect_analyze_loop_operations): Ditto.
1936 (vect_analyze_loop_2): Ditto.
1937 (vect_analyze_loop): Ditto.
1938 (report_vect_op): Ditto.
1939 (vect_is_slp_reduction): Ditto.
1940 (vect_is_simple_reduction_1): Ditto.
1941 (vect_get_known_peeling_cost): Ditto.
1942 (vect_estimate_min_profitable_iters): Ditto.
1943 (vect_model_reduction_cost): Ditto.
1944 (vect_model_induction_cost): Ditto.
1945 (get_initial_def_for_induction): Ditto.
1946 (vect_create_epilog_for_reduction): Ditto.
1947 (vectorizable_reduction): Ditto.
1948 (vectorizable_induction): Ditto.
1949 (vectorizable_live_operation): Ditto.
1950 (vect_loop_kill_debug_uses): Ditto.
1951 (vect_transform_loop): Ditto.
1952 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1953 (vect_stmt_relevant_p): Ditto.
1954 (process_use): Ditto.
1955 (vect_mark_stmts_to_be_vectorized): Ditto.
1956 (vect_model_simple_cost): Ditto.
1957 (vect_model_promotion_demotion_cost): Ditto.
1958 (vect_model_store_cost): Ditto.
1959 (vect_get_store_cost): Ditto.
1960 (vect_model_load_cost): Ditto.
1961 (vect_get_load_cost): Ditto.
1962 (vect_init_vector_1): Ditto.
1963 (vect_get_vec_def_for_operand): Ditto.
1964 (vect_finish_stmt_generation): Ditto.
1965 (vectorizable_call): Ditto.
1966 (vectorizable_conversion): Ditto.
1967 (vectorizable_assignment): Ditto.
1968 (vectorizable_shift): Ditto.
1969 (vectorizable_operation): Ditto.
1970 (vectorizable_store): Ditto.
1971 (vectorizable_load): Ditto.
1972 (vectorizable_condition): Ditto.
1973 (vect_analyze_stmt): Ditto.
1974 (vect_transform_stmt): Ditto.
1975 (vect_is_simple_use): Ditto.
1976 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1977 (vect_can_advance_ivs_p): Ditto.
1978 (vect_update_ivs_after_vectorizer): Ditto.
1979 (vect_do_peeling_for_loop_bound): Ditto.
1980 (vect_gen_niters_for_prolog_loop): Ditto.
1981 (vect_update_inits_of_drs): Ditto.
1982 (vect_create_cond_for_alias_checks): Ditto.
1983 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1984 (vect_build_slp_tree_1): Ditto.
1985 (vect_supported_load_permutation_p): Ditto.
1986 (vect_analyze_slp_instance): Ditto.
1987 (vect_analyze_slp): Ditto.
1988 (vect_make_slp_decision): Ditto.
1989 (vect_detect_hybrid_slp): Ditto.
1990 (vect_bb_vectorization_profitable_p): Ditto.
1991 (vect_slp_analyze_bb_1): Ditto.
1992 (vect_update_slp_costs_according_to_vf): Ditto.
1993 (vect_get_mask_element): Ditto.
1994 (vect_transform_slp_perm_load): Ditto.
1995 (vect_schedule_slp_instance): Ditto.
1996 (vect_schedule_slp): Ditto.
1997 (vect_slp_transform_bb): Ditto.
1998 * profile.c (read_profile_edge_counts): Ditto.
1999 (compute_branch_probabilities): Ditto.
2000 * coverage.c (get_coverage_counts): Ditto.
2002 2013-09-16 Diego Novillo <dnovillo@google.com>
2004 * tree-core.h: Add missing comment lines from refactoring of tree.h.
2006 2013-09-16 Jan Hubicka <jh@suse.cz>
2008 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
2009 abstract functions; for static functions check the presence of body.
2011 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2013 * config/aarch64/aarch64-simd-builtins.def (fma): New.
2014 * config/aarch64/aarch64-simd.md
2015 (aarch64_mla_elt<mode>): New.
2016 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2017 (aarch64_mls_elt<mode>): Likewise.
2018 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2019 (aarch64_fma4_elt<mode>): Likewise.
2020 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2021 (aarch64_fma4_elt_to_128v2df): Likewise.
2022 (aarch64_fma4_elt_to_64df): Likewise.
2023 (fnma<mode>4): Likewise.
2024 (aarch64_fnma4_elt<mode>): Likewise.
2025 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2026 (aarch64_fnma4_elt_to_128v2df): Likewise.
2027 (aarch64_fnma4_elt_to_64df): Likewise.
2028 * config/aarch64/iterators.md (VDQSF): New.
2029 * config/aarch64/arm_neon.h
2030 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2031 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2033 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2035 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2036 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2037 (aarch64_mul3_elt_to_128df): Likewise.
2038 (aarch64_mul3_elt_to_64v2df): Likewise.
2039 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2041 (VMUL_CHANGE_NLANES) Likewise.
2044 * config/aarch64/arm_neon.h
2045 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2047 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2049 * config/aarch64/arm_neon.h
2050 (vcvtx_high_f32_f64): Fix parameters.
2052 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2053 Uros Bizjak <ubizjak@gmail.com>
2055 * config/alpha.c: Include tree-ssanames.h.
2057 2013-09-16 Richard Biener <rguenther@suse.de>
2059 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2060 (dot_rdg_1): Handle control_dd.
2061 (create_edge_for_control_dependence): New function.
2062 (create_rdg_edges): Add control dependences if asked for.
2063 (build_rdg): Likewise.
2064 (generate_loops_for_partition): If there are not necessary
2065 control stmts remove all their dependencies.
2066 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2067 (distribute_loop): Pass on control dependences.
2068 (tree_loop_distribution): Compute control dependences and remove
2069 restriction on number of loop nodes.
2071 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2073 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2076 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2078 * cse.c (try_const_anchors): Punt on CC modes.
2080 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2082 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2084 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2087 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2090 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2093 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2095 (darwin_objc1_section): Likewise.
2096 (darwin_file_end): Emit Image Info section when required.
2098 2013-09-14 Jan Hubicka <jh@suse.cz>
2100 * tree-into-ssa.c (gate_into_ssa): New.
2101 (pass_data_build_ssa): Use it.
2102 * cgraph.h (expand_thunk): Update prototype.
2103 * cgraphunit.c (analyze_function): Expand thunks early.
2104 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2105 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2106 set lowered flag; do not add new function.
2107 (assemble_thunks_and_aliases): Update.
2108 * tree-ssa.c (gate_init_datastructures): New gate.
2109 (pass_data_init_datastructures): Use it.
2111 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2114 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2115 xmm register set described in the psABI.
2117 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2119 * dwarf2out.c (should_emit_struct_debug): Add check
2120 for type_decl variable is not NULL.
2122 2013-09-13 Jacek Caban <jacek@codeweavers.com>
2124 * config.gcc: Use new winnt-c.c target hooks
2125 * config/t-winnt: New file
2126 * config/winnt-c.c: New file
2127 * doc/tm.texi.in: Document new hook
2128 * doc/tm.texi: Regenerated
2130 2013-09-13 Jan Hubicka <jh@suse.cz>
2133 * ipa-inline.c (check_callers): New function.
2134 (check_caller_edge): Remove.
2135 (want_inline_function_to_all_callers_p): Also permit alises that are
2137 (inline_to_all_callers): Terminate the walk when devirtualization
2138 introduce new calls.
2140 2013-09-13 Jan Hubicka <jh@suse.cz>
2142 * ipa-inline-analysis.c (struct growth_data): Add node.
2143 (do_estimate_growth_1): Fix detection of recursion.
2145 2013-09-13 Jakub Jelinek <jakub@redhat.com>
2147 PR tree-optimization/58392
2148 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
2149 to avoid shadowing of outer loop variable. If
2150 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
2151 replace_by_duplicate_decl simduid of loops that have it set and
2152 set dest_cfun->has_simduid_loops and/or
2153 dest_cfun->has_force_vect_loops.
2154 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
2155 instead of maybe_lookup_decl.
2156 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
2157 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
2158 call. Copy over force_vect and copy and remap simduid. Set
2159 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
2160 (copy_cfg_body): Remove blocks_to_copy argument. Use
2161 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
2162 caller. Don't set cfun->has_simduid_loops and/or
2163 cfun->has_force_vect_loops here.
2164 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
2166 (expand_call_inline, tree_function_versioning): Adjust copy_body
2169 2013-09-13 Martin Jambor <mjambor@suse.cz>
2172 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
2173 the assert if the edge was a speculative one.
2175 2013-09-13 Richard Biener <rguenther@suse.de>
2177 * tree-data-ref.h (known_dependences_p): Move here ...
2178 * tree-loop-distribution.c (known_dependences_p): ... from here.
2179 (dump_rdg_component, debug_rdg_component): Remove.
2181 (generate_loops_for_partition): Use gimple_uid instead of
2182 relying on matching stmt visit order.
2183 (rdg_build_partitions): Take starting stmt vector.
2184 (ldist_gen): Merge into ...
2185 (distribute_loop): ... this function. Do not compute starting
2187 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
2189 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2191 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
2192 Set type attribute correctly. Set predicable_short_it attribute.
2193 (cmpsi_shiftsi): Remove %? from output template.
2195 2013-09-13 Richard Biener <rguenther@suse.de>
2197 * tree-loop-distribution.c (struct rdg_component,
2198 rdg_defs_used_in_other_loops_p, free_rdg_components,
2199 rdg_build_components): Remove.
2200 (stmts_from_loop): Do not record virtual PHIs.
2201 (generate_loops_for_partition): Skip virtual PHIs.
2202 (build_rdg_partition_for_component): Rename to ...
2203 (build_rdg_partition_for_vertex): ... this and adjust.
2204 (rdg_build_partitions): Take a vector of starting vertices
2205 instead of components. Remove unnecessary leftover handling.
2206 (ldist_gen): Do not build components or record other stores.
2207 (distribute_loop): Do not distribute loops containing stmts
2210 2013-09-13 Christian Bruel <christian.bruel@st.com>
2213 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
2215 2013-09-13 Kai Tietz <ktietz@redhat.com>
2217 * config.gcc: Separate cases for mingw and cygwin targets,
2218 and add 64-bit cygwin target case.
2220 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
2221 dll-export inline-functions.
2222 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
2225 2013-09-13 Jeff Law <law@redhat.com>
2229 2013-09-06 Jeff Law <law@redhat.com>
2231 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2232 edge implied equivalences into successor phis.
2234 2013-09-12 DJ Delorie <dj@redhat.com>
2236 * config/rl78/rl78-virt.md: Change from | to \; for asm line
2239 2013-09-12 Brooks Moses <bmoses@google.com>
2242 * Makefile.in: Do not install driver binaries in $(target)/bin.
2244 2013-09-12 DJ Delorie <dj@redhat.com>
2246 * config/rl78/rl78.opt (mrelax): New.
2247 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
2248 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2250 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2252 (rl78_expand_epilogue): Likewise.
2254 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2257 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2258 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2259 pointer elimination with using remove_reg_equal_offset_note.
2261 2013-09-12 DJ Delorie <dj@redhat.com>
2263 * config/msp430/: New port.
2264 * config.gcc (msp430): Added.
2265 * doc/invoke.texi: Document MSP430 options.
2266 * doc/install.texi: Document msp430-elf
2267 * doc/md.texi: Document msp430-elf
2268 * doc/contrib.texi: Document msp430-elf
2270 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
2273 2013-09-12 Martin Jambor <mjambor@suse.cz>
2276 * ipa-prop.c (remove_described_reference): Give up if the edge in the
2277 reference descriptor is NULL.
2278 (ipa_edge_removal_hook): If owning a reference descriptor, set its
2281 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2283 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
2284 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
2285 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
2286 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
2287 make_temp_ssa_name): move to tree-ssanames.h
2288 * tree-ssa-alias.h: Move prototype.
2289 * tree-ssa.h: Include tree-ssanames.h.
2290 * tree-ssanames.c (FREE_SSANAMES): Move to here.
2291 * tree-ssanames.h: New. Move items from tree-flow*.h
2292 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
2294 2013-09-12 Richard Biener <rguenther@suse.de>
2296 PR tree-optimization/58404
2297 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
2298 propagate non-invariant addresses into dereferences wrapped
2299 in component references.
2301 2013-09-12 Richard Biener <rguenther@suse.de>
2303 PR tree-optimization/58402
2304 * passes.def: Move pass_late_warn_uninitialized later.
2306 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2308 * tree-ssa.h: New. Move content from tree-flow.h and
2310 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
2311 Move prototypes belonging to tree-ssa.c.
2312 * tree-flow-inline.h (redirect_edge_var_map_def,
2313 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
2315 * gimple.h: Adjust prototypes.
2316 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
2318 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
2319 * tree.h: Move prototype to tree-ssa.h.
2320 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
2321 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
2322 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
2323 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
2324 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
2325 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
2326 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
2327 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
2328 graphite-blocking.c, graphite-clast-to-gimple.c,
2329 graphite-dependences.c, graphite-interchange.c,
2330 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
2331 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
2332 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
2333 ipa-reference.c, ipa-split.c, ipa-utils.c,
2334 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
2335 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
2336 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
2337 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2338 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
2339 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
2340 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
2341 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
2342 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
2343 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
2344 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
2345 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2346 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
2347 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
2348 value-prof.c, var-tracking.c,
2349 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
2351 2013-09-12 Richard Biener <rguenther@suse.de>
2353 PR tree-optimization/58396
2354 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
2355 (build_rdg): Take a loop-nest parameter, fix memleaks.
2356 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
2358 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2360 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
2363 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
2365 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
2366 The FMA instruction names should have a 'v' prefix.
2368 2013-09-12 Richard Biener <rguenther@suse.de>
2370 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
2371 (dot_rdg): Use popen instead of system in optional code.
2372 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
2373 (already_processed_vertex_p): Adjust.
2374 (has_anti_or_output_dependence, predecessor_has_mem_write,
2375 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
2376 rdg_flag_uses): Remove.
2377 (rdg_flag_vertex): Simplify.
2378 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
2380 (build_rdg_partition_for_component): Process the first vertex
2381 of a component only.
2382 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
2384 2013-09-12 Alan Modra <amodra@gmail.com>
2386 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
2388 2013-09-11 DJ Delorie <dj@redhat.com>
2389 Nick Clifton <nickc@redhat.com>
2391 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
2392 (rl78_stack_based_mem): New.
2393 * config/rl78/constraints.md (Iv08): New.
2402 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
2404 (movsi): Change from expand to insn-and-split.
2405 (ashrsi3): Clobber AX.
2409 * config/rl78/rl78.md (CC_REG): Fix.
2410 (addsi3): Allow memory and immediate operands.
2411 (addsi3_internal): Split into...
2412 (addsi3_internal_virt): ...new, and ...
2413 (addsi3_internal_real): ...new.
2415 (subsi3_internal_virt): New.
2416 (subsi3_internal_real): New.
2417 (mulsi3): Add memory operand.
2418 (mulsi3_rl78): Likewise.
2419 (mulsi3_g13): Likewise.
2420 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
2421 (cbranchqi4_real): Add more constraint options.
2422 (cbranchhi4_real): Expand pattern.
2423 (cbranchhi4_real_signed): New.
2424 (cbranchhi4_real_inverted): New.
2425 (cbranchsi4_real_lt): New.
2426 (cbranchsi4_real_ge): New.
2427 (cbranchsi4_real_signed): New.
2428 (cbranchsi4_real): New.
2430 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
2432 (lshrsi3_virt): Likewise.
2433 (ashlsi3_virt): Likewise.
2434 (cbranchqi4_virt_signed): New.
2435 (cbranchhi4_virt_signed): New.
2436 (cbranchsi4_virt): New.
2437 * config/rl78/rl78.c: Whitespace fixes throughout.
2438 (move_elim_pass): New.
2439 (pass_data_rl78_move_elim): New.
2440 (pass_rl78_move_elim): New.
2441 (make_pass_rl78_move_elim): New.
2442 (rl78_devirt_info): Run devirt earlier.
2443 (rl78_move_elim_info): New.
2444 (rl78_asm_file_start): Register it.
2445 (rl78_split_movsi): New.
2446 (rl78_as_legitimate_address): Allow virtual base registers when
2448 (rl78_addr_space_convert): Remove spurious debug stuff.
2449 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
2450 (rl78_print_operand): More cases for not printing '#'.
2451 (rl78_expand_compare): Remove most of the logic.
2452 (content_memory): New.
2453 (clear_content_memory): New.
2454 (get_content_index): New.
2455 (get_content_name): New.
2456 (display_content_memory): New.
2457 (update_content): New.
2458 (record_content): New.
2459 (already_contains): New.
2460 (insn_ok_now): Re-recog insns with virtual registers.
2461 (add_postponed_content_update): New.
2462 (process_postponed_content_update): New.
2463 (gen_and_emit_move): New.
2464 (transcode_memory_rtx): Record new location content.
2465 Use gen_and_emit_move.
2466 (force_into_acc): New.
2467 (move_to_acc): Use gen_and_emit_move.
2468 (move_from_acc): Likewise.
2469 (move_acc_to_reg): Likewise.
2470 (move_to_x): Likewise.
2471 (move_to_hl): Likewise.
2472 (move_to_de): Likewise.
2473 (rl78_alloc_physical_registers_op1): Record location content.
2474 (has_constraint): New.
2475 (rl78_alloc_physical_registers_op2): Record location content.
2477 (rl78_alloc_physical_registers_ro1): Likewise.
2478 (rl78_alloc_physical_registers_cmp): Likewise.
2479 (rl78_alloc_physical_registers_umul): Likewise.
2480 (rl78_alloc_address_registers_macax): New.
2481 (rl78_alloc_physical_registers): Initialize and set location
2482 content memory as needed.
2483 (rl78_reorg): Make sure split2 is called.
2484 (rl78_rtx_costs): New.
2486 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
2488 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
2489 for (not (neg ...)) and (neg (not ...)) cases.
2491 2013-09-11 Richard Biener <rguenther@suse.de>
2494 * passes.def: Split critical edges before late uninit warning passes.
2495 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
2497 2013-09-11 Jakub Jelinek <jakub@redhat.com>
2499 PR tree-optimization/58385
2500 * fold-const.c (build_range_check): If both low and high are NULL,
2501 use omit_one_operand_loc to preserve exp side-effects.
2503 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2505 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
2507 2013-09-11 Richard Biener <rguenther@suse.de>
2509 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
2510 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
2511 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
2512 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
2513 stmts_from_loop, known_dependences_p, build_empty_rdg,
2514 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
2515 * tree-loop-distribution.c: ... here.
2516 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
2517 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
2518 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
2519 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
2520 * tree-loop-distribution.c: ... here.
2521 * tree-loop-distribution.c: Include gimple-pretty-print.h.
2522 (struct partition_s): Add loops member.
2523 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
2524 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
2525 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
2527 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2528 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2529 Sergey Lega <sergey.s.lega@intel.com>
2530 Anna Tikhonova <anna.tikhonova@intel.com>
2531 Ilya Tocar <ilya.tocar@intel.com>
2532 Andrey Turetskiy <andrey.turetskiy@intel.com>
2533 Ilya Verbin <ilya.verbin@intel.com>
2534 Kirill Yukhin <kirill.yukhin@intel.com>
2535 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2537 * config/i386/constraints.md (k): New.
2539 * config/i386/i386.c (const regclass_map): Add new mask registers.
2540 (dbx_register_map): Ditto.
2541 (dbx64_register_map): Ditto.
2542 (svr4_dbx_register_map): Ditto.
2543 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2545 (ix86_preferred_reload_class): Disable constants for mask registers.
2546 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2547 (ix86_hard_regno_mode_ok): Support new mask registers.
2548 (x86_order_regs_for_local_alloc): Ditto.
2549 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2550 (FIXED_REGISTERS): Add new mask registers.
2551 (CALL_USED_REGISTERS): Ditto.
2552 (REG_ALLOC_ORDER): Ditto.
2553 (VALID_MASK_REG_MODE): New.
2554 (FIRST_MASK_REG): Ditto.
2555 (LAST_MASK_REG): Ditto.
2556 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2557 (MAYBE_MASK_CLASS_P): New.
2558 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2559 (REG_CLASS_CONTENTS): Ditto.
2560 (MASK_REGNO_P): New.
2561 (ANY_MASK_REG_P): Ditto.
2562 (HI_REGISTER_NAMES): Add new mask registers.
2563 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2564 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2566 (attribute "type"): Add mskmov, msklog.
2567 (attribute "length_immediate"): Support them.
2568 (attribute "memory"): Ditto.
2569 (attribute "prefix_0f"): Ditto.
2570 (*movhi_internal): Support new mask registers.
2571 (*movqi_internal): Ditto.
2572 (define_split): Split out clobber pattern is a logic
2573 insn on mask registers.
2574 (*k<logic><mode>): New.
2575 (*andhi_1): Extend to support mask regs.
2576 (*andqi_1): Extend to support mask regs.
2578 (define_split): Split and-not to and and not if operands
2580 (*<code><mode>_1): Separate HI mode to new pattern...
2581 (*<code>hi_1): This.
2582 (*<code>qi_1): Extend to support mask regs.
2584 (kortestzhi): Ditto.
2585 (kortestchi): Ditto.
2587 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2588 (*one_cmplhi2_1): ...Here, now with mask registers support.
2589 (*one_cmplqi2_1): Support new mask registers.
2590 (HI/QImode arithmetics splitter): Don't split if mask registers
2592 (HI/QImode not splitter): Ditto.
2593 * config/i386/predicated.md (mask_reg_operand): New.
2594 (general_reg_operand): Ditto.
2596 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2598 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2599 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2601 2013-09-10 Jeff Law <law@redhat.com>
2603 PR tree-optimization/58380
2604 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2605 of threading through a buried loop header.
2607 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2608 return value for single successor case.
2610 2013-09-10 Jan Hubicka <jh@suse.cz>
2612 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2614 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2617 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2618 support conditional execution.
2619 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2621 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2623 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2625 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2626 Check destination too to check move insn.
2627 (undo_optional_reloads): Add check that the original peudo did not
2628 changed its allocation and the optional reload was inherited on last
2629 inheritance pass. Break loop after deciding to keep optional reload.
2630 (lra_undo_inheritance): Add check that inherited pseudo still in
2633 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2635 * config/aarch64/aarch64.md (generic_sched): New.
2636 * config/aarch64/aarch64-generic.md (load): Make conditional
2637 on generic_sched attribute.
2638 (nonload): Likewise.
2640 2013-09-10 Jan Hubicka <jh@suse.cz>
2642 * lto-cgraph.c: Include ipa-utils.h.
2643 (compute_ltrans_boundary): Also add possible targets into the boundary.
2645 2013-09-10 Jan Hubicka <jh@suse.cz>
2647 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2648 VAR_DECL of vtable rather than full expression.
2650 2013-09-10 Jan Hubicka <jh@suse.cz>
2651 Paolo Carlini <paolo.carlini@oracle.com>
2653 * cgraphunit.c (analyze_functions): Save input_location, set it
2654 to UNKNOWN_LOCATION and restore it at the end.
2656 2013-09-10 Martin Jambor <mjambor@suse.cz>
2658 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2659 represented by a thunk.
2661 2013-09-10 Jeff Law <law@redhat.com>
2663 PR tree-optimization/58343
2664 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2665 jump threading requests through loop headers buried in the middle
2666 of a jump threading path.
2668 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2669 in return value/type.
2671 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2673 PR rtl-optimization/58365
2674 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2675 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2678 2013-09-10 Richard Biener <rguenther@suse.de>
2680 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2681 * tree-data-ref.c (create_rdg_vertices): Collect all data
2682 references, signal failure to the caller, use data-ref API.
2683 (build_rdg): Compute data references only once. Maintain lifetime
2684 of data references and data dependences from within RDG.
2685 (free_rdg): Free dependence relations.
2686 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2687 inventing extra dependences.
2688 (distribute_loop): Update for RDG API changes.
2690 2013-09-10 Kai Tietz <ktietz@redhat.com>
2692 * doc/invoke.texi (fms-extensions): Document changed
2693 behavior for ms-abi targets.
2694 * config/i386/i386.c (ix86_option_override_internal):
2695 Set default value of option -fms-extension for ms-abi targets.
2697 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2699 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2701 2013-09-10 Alan Modra <amodra@gmail.com>
2704 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2706 2013-09-10 Alan Modra <amodra@gmail.com>
2708 * config/rs6000/predicates.md (add_cint_operand): New.
2709 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2710 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2711 using add_cint_operand.
2712 (largetoc_high_plus_aix): Likewise.
2714 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2716 PR tree-optimization/58364
2717 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2718 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2719 the current range can't be an unconditional true or false.
2721 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2723 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2725 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2727 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2729 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2732 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2734 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2736 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2737 NO_REGS for immediate that can't be moved directly into FP_REGS.
2739 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2741 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2742 comparison with negated operand.
2743 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2746 2013-09-09 Richard Biener <rguenther@suse.de>
2749 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2750 of a subloop record all its block as affecting loop-closed SSA form.
2752 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2754 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2755 of an rtx/bitpos pair.
2756 (store_fixed_bit_field): Update accordingly.
2758 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2760 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2762 * builtins.c (expand_errno_check): Likewise.
2763 * dwarf2cfi.c (init_return_column_size): Likewise.
2764 * except.c (sjlj_mark_call_sites): Likewise.
2765 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2766 * lra-constraints.c (emit_inc): Likewise.
2767 * ree.c (combine_set_extension): Likewise.
2768 * regmove.c (fixup_match_2): Likewise.
2769 * reload1.c (inc_for_reload): Likewise.
2771 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2773 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2774 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2775 Use gen_int_mode with the mode of the associated simplify_* call.
2776 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2777 * expmed.c (expand_shift_1): Likewise.
2778 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2779 * loop-iv.c (iv_number_of_iterations): Likewise.
2780 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2781 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2783 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2785 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2786 of the associated expand_* call.
2787 (asan_emit_stack_protection): Likewise.
2788 * builtins.c (round_trampoline_addr): Likewise.
2789 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2790 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2791 (emit_store_flag): Likewise.
2792 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2793 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2795 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2796 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2797 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2798 * modulo-sched.c (generate_prolog_epilog): Likewise.
2799 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2800 (expand_ctz, expand_ffs, expand_unop): Likewise.
2802 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2804 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2805 of the associated gen_rtx_* call.
2806 * caller-save.c (init_caller_save): Likewise.
2807 * combine.c (find_split_point, make_extraction): Likewise.
2808 (make_compound_operation): Likewise.
2809 * dwarf2out.c (mem_loc_descriptor): Likewise.
2810 * explow.c (plus_constant, probe_stack_range): Likewise.
2811 * expmed.c (expand_mult_const): Likewise.
2812 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2813 * reload1.c (init_reload): Likewise.
2814 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2815 * var-tracking.c (adjust_mems): Likewise.
2816 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2817 rather than gen_rtx_fmt_ee.
2819 2013-09-09 Jan Hubicka <jh@suse.cz>
2822 * value-prof.c (gimple_ic): Copy also abnormal edges.
2824 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2826 * asan.c (asan_shadow_cst): Use gen_int_mode.
2828 2013-09-08 Jan Hubicka <jh@suse.cz>
2830 * ipa-profile.c: Add toplevel comment.
2831 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2832 (contains_hot_call_p): New function.
2833 (ipa_propagate_frequency): Set frequencies based on counts when
2835 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2836 profile; do not tamper with profile after inlining if it is read.
2838 2013-09-08 Jan Hubicka <jh@suse.cz>
2840 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2843 2013-09-08 Jan Hubicka <jh@suse.cz>
2845 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2848 2013-09-08 Jeff Law <law@redhat.com>
2851 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2854 2013-09-08 Andi Kleen <ak@linux.intel.com>
2856 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2858 2013-09-08 Jan Hubicka <jh@suse.cz>
2860 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2862 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2864 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2865 for non-debug insns.
2866 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2867 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2868 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2871 2013-09-08 Jan Hubicka <jh@suse.cz>
2873 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2874 targets and devirtualize to BUILT_IN_UNREACHABLE.
2875 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2876 * ipa.c (walk_polymorphic_call_targets): New function.
2877 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2878 functions; use the new timevar.
2879 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2880 was removed from the program.
2881 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2882 not consider it in the walk when its vtable is dead.
2883 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2885 (devirt_variable_node_removal_hook): New function.
2886 (possible_polymorphic_call_targets): Also register
2887 devirt_variable_node_removal_hook.
2888 (ipa_devirt): Do not do non-speculative devirtualization.
2889 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2891 2013-09-08 Jan Hubicka <jh@suse.cz>
2893 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2894 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2895 varpool_remove_variable_insertion_hook): Declare.
2896 * varpool.c (varpool_node_hook_list): New structure.
2897 (first_varpool_node_removal_hook,
2898 first_varpool_variable_insertion_hook): New variables.
2899 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2900 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2901 varpool_remove_variable_insertion_hook,
2902 varpool_call_variable_insertion_hooks): New functions.
2903 (varpool_remove_node): Use it.
2905 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2908 * diagnostic.c (diagnostic_build_prefix): When s.file is
2909 "<built-in>" don't output line and column numbers.
2911 2013-09-06 Jan Hubicka <jh@suse.cz>
2913 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2914 * lto-streamer-out.c: Stream thunks, too.
2915 * lto-streamer-in.c (input_function): Pop cfun here
2916 (lto_read_body): Instead of here.
2918 2013-09-06 Caroline Tice <cmtice@google.com>
2920 * doc/install.texi: Add documentation for the --enable-vtable-verify
2921 and the --disable-libvtv configure options.
2923 2013-09-06 Jeff Law <law@redhat.com>
2925 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2926 edge implied equivalences into successor phis.
2928 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2930 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2932 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2934 * resource.c (mark_target_live_regs): Compute resources taking
2935 into account if a call is predicated or not.
2937 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2939 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2940 the compiler in the function names.
2942 2013-09-06 Jan Hubicka <jh@suse.cz>
2945 * ipa-inline.c (has_caller_p): New function.
2946 (want_inline_function_to_all_callers_p): Use it.
2947 (sum_callers, inline_to_all_callers): Break out from ...
2948 (ipa_inline): ... here.
2950 2013-09-06 Jan Hubicka <jh@suse.cz>
2952 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2953 only when AVX is enabled.
2955 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2957 * config/aarch64/aarch64.md
2958 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2959 is fpsimd_<load/store>2.
2960 (load_pair<mode>): Likewise.
2961 (store_pair<mode>): Likewise.
2963 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2965 * config/arm/types.md (type): Add "mrs" type.
2966 * config/aarch64/aarch64.md
2967 (aarch64_load_tp_hard): Make type "mrs".
2969 (load_tp_hard): Make type "mrs".
2970 * config/arm/cortex-a15.md: Update with new attributes.
2971 * config/arm/cortex-a5.md: Update with new attributes.
2972 * config/arm/cortex-a53.md: Update with new attributes.
2973 * config/arm/cortex-a7.md: Update with new attributes.
2974 * config/arm/cortex-a8.md: Update with new attributes.
2975 * config/arm/cortex-a9.md: Update with new attributes.
2976 * config/arm/cortex-m4.md: Update with new attributes.
2977 * config/arm/cortex-r4.md: Update with new attributes.
2978 * config/arm/fa526.md: Update with new attributes.
2979 * config/arm/fa606te.md: Update with new attributes.
2980 * config/arm/fa626te.md: Update with new attributes.
2981 * config/arm/fa726te.md: Update with new attributes.
2983 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2985 * config/aarch64/aarch64.md
2986 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2987 (*movtf_aarch64): Likewise.
2989 (thumb1_movdi_insn): Use "multiple" for type where more than one
2990 instruction is used for a move.
2991 (*arm32_movhf): Likewise.
2992 (*thumb_movdf_insn): Likewise.
2994 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2996 * config/arm/types.md (type): Rename fcpys to fmov.
2998 (*arm_movsi_vfp): Rename type fcpys as fmov.
2999 (*thumb2_movsi_vfp): Likewise
3000 (*movhf_vfp_neon): Likewise
3001 (*movhf_vfp): Likewise
3002 (*movsf_vfp): Likewise
3003 (*thumb2_movsf_vfp): Likewise
3004 (*movsfcc_vfp): Likewise
3005 (*thumb2_movsfcc_vfp): Likewise
3006 * config/aarch64/aarch64-simd.md
3007 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
3008 * config/aarch64/aarch64.md
3009 (*movsi_aarch64): Replace type mov_reg with fmovs.
3010 (*movdi_aarch64): Likewise
3011 (*movsf_aarch64): Likewise
3012 (*movdf_aarch64): Likewise
3014 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3015 * config/arm/iwmmxt.md
3016 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3017 * config/arm/arm1020e.md: Update with new attributes.
3018 * config/arm/cortex-a15-neon.md: Update with new attributes.
3019 * config/arm/cortex-a5.md: Update with new attributes.
3020 * config/arm/cortex-a53.md: Update with new attributes.
3021 * config/arm/cortex-a7.md: Update with new attributes.
3022 * config/arm/cortex-a8-neon.md: Update with new attributes.
3023 * config/arm/cortex-a9.md: Update with new attributes.
3024 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3025 * config/arm/cortex-r4f.md: Update with new attributes.
3026 * config/arm/marvell-pj4.md: Update with new attributes.
3027 * config/arm/vfp11.md: Update with new attributes.
3029 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3031 * config/aarch64/aarch64.md
3032 (*madd<mode>): Fix type attribute.
3033 (*maddsi_uxtw): Likewise.
3034 (*msub<mode>): Likewise.
3035 (*msubsi_uxtw): Likewise.
3036 (<su_optab>maddsidi4): Likewise.
3037 (<su_optab>msubsidi4): Likewise.
3039 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3041 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3042 * config/arm/arm.md (core_cycles): Remove fdiv.
3043 * config/arm/vfp.md:
3044 (*sqrtsf2_vfp): Update for attribute changes.
3045 (*sqrtdf2_vfp): Likewise.
3046 * config/aarch64/aarch64.md:
3047 (sqrt<mode>2): Update for attribute changes.
3048 * config/arm/arm1020e.md: Update with new attributes.
3049 * config/arm/cortex-a15-neon.md: Update with new attributes.
3050 * config/arm/cortex-a5.md: Update with new attributes.
3051 * config/arm/cortex-a53.md: Update with new attributes.
3052 * config/arm/cortex-a7.md: Update with new attributes.
3053 * config/arm/cortex-a8-neon.md: Update with new attributes.
3054 * config/arm/cortex-a9.md: Update with new attributes.
3055 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3056 * config/arm/cortex-r4f.md: Update with new attributes.
3057 * config/arm/marvell-pj4.md: Update with new attributes.
3058 * config/arm/vfp11.md: Update with new attributes.
3060 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3062 * config/arm/types.md
3063 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3064 * config/aarch64/aarch64.md
3065 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3067 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3068 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3069 (float<GPI:mode><GPF:mode>2): Likewise.
3071 (*truncsisf2_vfp): Update with new attributes.
3072 (*truncsidf2_vfp): Likewise.
3073 (fixuns_truncsfsi2): Likewise.
3074 (fixuns_truncdfsi2): Likewise.
3075 (*floatsisf2_vfp): Likewise.
3076 (*floatsidf2_vfp): Likewise.
3077 (floatunssisf2): Likewise.
3078 (floatunssidf2): Likewise.
3079 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3080 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3081 * config/arm/arm1020e.md: Update with new attributes.
3082 * config/arm/cortex-a15-neon.md: Update with new attributes.
3083 * config/arm/cortex-a5.md: Update with new attributes.
3084 * config/arm/cortex-a53.md: Update with new attributes.
3085 * config/arm/cortex-a7.md: Update with new attributes.
3086 * config/arm/cortex-a8-neon.md: Update with new attributes.
3087 * config/arm/cortex-a9.md: Update with new attributes.
3088 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3089 * config/arm/cortex-r4f.md: Update with new attributes.
3090 * config/arm/marvell-pj4.md: Update with new attributes.
3091 * config/arm/vfp11.md: Update with new attributes.
3093 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3095 * config/aarch64/arm_neon.h
3096 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3097 (vqtbx<1,2,3,4><q>_s8): Likewise.
3099 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3101 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3102 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3103 (add<mode>3): Add type attribute.
3104 (add<mode>3): Likewise.
3105 (usadd<mode>3): Likewise.
3106 (ssadd<mode>3): Likewise.
3107 (sub<mode>3): Likewise.
3108 (sub<mode>3): Likewise.
3109 (ussub<mode>3): Likewise.
3110 (sssub<mode>3): Likewise.
3111 (ssmulsa3): Likewise.
3112 (usmulusa3): Likewise.
3113 (arm_usatsihi): Likewise.
3115 (*movdi_vfp): Add types for all instructions.
3116 (*movdi_vfp_cortexa8): Likewise.
3117 (*movhf_vfp_neon): Likewise.
3118 (*movhf_vfp): Likewise.
3119 (*movdf_vfp): Likewise.
3120 (*thumb2_movdf_vfp): Likewise.
3121 (*thumb2_movdfcc_vfp): Likewise.
3122 * config/arm/arm.md: Add type attribute to all insn patterns.
3123 (*thumb1_adddi3): Add type attribute.
3124 (*arm_adddi3): Likewise.
3125 (*adddi_sesidi_di): Likewise.
3126 (*adddi_zesidi_di): Likewise.
3127 (*thumb1_addsi3): Likewise.
3128 (addsi3_compare0): Likewise.
3129 (*addsi3_compare0_scratch): Likewise.
3130 (*compare_negsi_si): Likewise.
3131 (cmpsi2_addneg): Likewise.
3132 (*addsi3_carryin_<optab>): Likewise.
3133 (*addsi3_carryin_alt2_<optab>): Likewise.
3134 (*addsi3_carryin_clobercc_<optab>): Likewise.
3135 (*subsi3_carryin): Likewise.
3136 (*subsi3_carryin_const): Likewise.
3137 (*subsi3_carryin_compare): Likewise.
3138 (*subsi3_carryin_compare_const): Likewise.
3139 (*arm_subdi3): Likewise.
3140 (*thumb_subdi3): Likewise.
3141 (*subdi_di_zesidi): Likewise.
3142 (*subdi_di_sesidi): Likewise.
3143 (*subdi_zesidi_di): Likewise.
3144 (*subdi_sesidi_di): Likewise.
3145 (*subdi_zesidi_ze): Likewise.
3146 (thumb1_subsi3_insn): Likewise.
3147 (*arm_subsi3_insn): Likewise.
3148 (*anddi3_insn): Likewise.
3149 (*anddi_zesidi_di): Likewise.
3150 (*anddi_sesdi_di): Likewise.
3151 (*ne_zeroextracts): Likewise.
3152 (*ne_zeroextracts): Likewise.
3153 (*ite_ne_zeroextr): Likewise.
3154 (*ite_ne_zeroextr): Likewise.
3155 (*anddi_notdi_di): Likewise.
3156 (*anddi_notzesidi): Likewise.
3157 (*anddi_notsesidi): Likewise.
3158 (andsi_notsi_si): Likewise.
3159 (thumb1_bicsi3): Likewise.
3160 (*iordi3_insn): Likewise.
3161 (*iordi_zesidi_di): Likewise.
3162 (*iordi_sesidi_di): Likewise.
3163 (*thumb1_iorsi3_insn): Likewise.
3164 (*xordi3_insn): Likewise.
3165 (*xordi_zesidi_di): Likewise.
3166 (*xordi_sesidi_di): Likewise.
3167 (*arm_xorsi3): Likewise.
3168 (*andsi_iorsi3_no): Likewise.
3169 (*smax_0): Likewise.
3170 (*smax_m1): Likewise.
3171 (*arm_smax_insn): Likewise.
3172 (*smin_0): Likewise.
3173 (*arm_smin_insn): Likewise.
3174 (*arm_umaxsi3): Likewise.
3175 (*arm_uminsi3): Likewise.
3176 (*minmax_arithsi): Likewise.
3177 (*minmax_arithsi_): Likewise.
3178 (*satsi_<SAT:code>): Likewise.
3179 (arm_ashldi3_1bit): Likewise.
3180 (arm_ashrdi3_1bit): Likewise.
3181 (arm_lshrdi3_1bit): Likewise.
3182 (*arm_negdi2): Likewise.
3183 (*thumb1_negdi2): Likewise.
3184 (*arm_negsi2): Likewise.
3185 (*thumb1_negsi2): Likewise.
3186 (*negdi_extendsid): Likewise.
3187 (*negdi_zero_extend): Likewise.
3188 (*arm_abssi2): Likewise.
3189 (*thumb1_abssi2): Likewise.
3190 (*arm_neg_abssi2): Likewise.
3191 (*thumb1_neg_abss): Likewise.
3192 (one_cmpldi2): Likewise.
3193 (extend<mode>di2): Likewise.
3194 (*compareqi_eq0): Likewise.
3195 (*arm_extendhisi2addsi): Likewise.
3196 (*arm_movdi): Likewise.
3197 (*thumb1_movdi_insn): Likewise.
3198 (*arm_movt): Likewise.
3199 (*thumb1_movsi_insn): Likewise.
3200 (pic_add_dot_plus_four): Likewise.
3201 (pic_add_dot_plus_eight): Likewise.
3202 (tls_load_dot_plus_eight): Likewise.
3203 (*thumb1_movhi_insn): Likewise.
3204 (*thumb1_movsf_insn): Likewise.
3205 (*movdf_soft_insn): Likewise.
3206 (*thumb_movdf_insn): Likewise.
3207 (cbranchsi4_insn): Likewise.
3208 (cbranchsi4_scratch): Likewise.
3209 (*negated_cbranchsi4): Likewise.
3210 (*tbit_cbranch): Likewise.
3211 (*tlobits_cbranch): Likewise.
3212 (*tstsi3_cbranch): Likewise.
3213 (*cbranchne_decr1): Likewise.
3214 (*addsi3_cbranch): Likewise.
3215 (*addsi3_cbranch_scratch): Likewise.
3216 (*arm_cmpdi_insn): Likewise.
3217 (*arm_cmpdi_unsig): Likewise.
3218 (*arm_cmpdi_zero): Likewise.
3219 (*thumb_cmpdi_zero): Likewise.
3220 (*deleted_compare): Likewise.
3221 (*mov_scc): Likewise.
3222 (*mov_negscc): Likewise.
3223 (*mov_notscc): Likewise.
3224 (*cstoresi_eq0_thumb1_insn): Likewise.
3225 (cstoresi_nltu_thumb1): Likewise.
3226 (cstoresi_ltu_thu): Likewise.
3227 (thumb1_addsi3_addgeu): Likewise.
3228 (*arm_jump): Likewise.
3229 (*thumb_jump): Likewise.
3230 (*check_arch2): Likewise.
3231 (arm_casesi_internal): Likewise.
3232 (thumb1_casesi_dispatch): Likewise.
3233 (*arm_indirect_jump): Likewise.
3234 (*thumb1_indirect_jump): Likewise.
3236 (*and_scc): Likewise.
3237 (*ior_scc): Likewise.
3238 (*compare_scc): Likewise.
3239 (*cond_move): Likewise.
3240 (*cond_arith): Likewise.
3241 (*cond_sub): Likewise.
3242 (*cmp_ite0): Likewise.
3243 (*cmp_ite1): Likewise.
3244 (*cmp_and): Likewise.
3245 (*cmp_ior): Likewise.
3246 (*ior_scc_scc): Likewise.
3247 (*ior_scc_scc_cmp): Likewise.
3248 (*and_scc_scc): Likewise.
3249 (*and_scc_scc_cmp): Likewise.
3250 (*and_scc_scc_nod): Likewise.
3251 (*negscc): Likewise.
3252 (movcond_addsi): Likewise.
3253 (movcond): Likewise.
3254 (*ifcompare_plus_move): Likewise.
3255 (*if_plus_move): Likewise.
3256 (*ifcompare_move_plus): Likewise.
3257 (*if_move_plus): Likewise.
3258 (*ifcompare_arith_arith): Likewise.
3259 (*if_arith_arith): Likewise.
3260 (*ifcompare_arith_move): Likewise.
3261 (*if_arith_move): Likewise.
3262 (*ifcompare_move_arith): Likewise.
3263 (*if_move_arith): Likewise.
3264 (*ifcompare_move_not): Likewise.
3265 (*if_move_not): Likewise.
3266 (*ifcompare_not_move): Likewise.
3267 (*if_not_move): Likewise.
3268 (*ifcompare_shift_move): Likewise.
3269 (*if_shift_move): Likewise.
3270 (*ifcompare_move_shift): Likewise.
3271 (*if_move_shift): Likewise.
3272 (*ifcompare_shift_shift): Likewise.
3273 (*ifcompare_not_arith): Likewise.
3274 (*ifcompare_arith_not): Likewise.
3275 (*if_arith_not): Likewise.
3276 (*ifcompare_neg_move): Likewise.
3277 (*if_neg_move): Likewise.
3278 (*ifcompare_move_neg): Likewise.
3279 (*if_move_neg): Likewise.
3280 (prologue_thumb1_interwork): Likewise.
3281 (*cond_move_not): Likewise.
3282 (*sign_extract_onebit): Likewise.
3283 (*not_signextract_onebit): Likewise.
3284 (stack_tie): Likewise.
3285 (align_4): Likewise.
3286 (align_8): Likewise.
3287 (consttable_end): Likewise.
3288 (consttable_1): Likewise.
3289 (consttable_2): Likewise.
3290 (consttable_4): Likewise.
3291 (consttable_8): Likewise.
3292 (consttable_16): Likewise.
3293 (*thumb1_tablejump): Likewise.
3294 (prefetch): Likewise.
3295 (force_register_use): Likewise.
3296 (thumb_eh_return): Likewise.
3297 (load_tp_hard): Likewise.
3298 (load_tp_soft): Likewise.
3299 (tlscall): Likewise.
3300 (*arm_movtas_ze): Likewise.
3301 (*arm_rev): Likewise.
3302 (*arm_revsh): Likewise.
3303 (*arm_rev16): Likewise.
3304 * config/arm/thumb2.md
3305 (*thumb2_smaxsi3): Likewise.
3306 (*thumb2_sminsi3): Likewise.
3307 (*thumb32_umaxsi3): Likewise.
3308 (*thumb2_uminsi3): Likewise.
3309 (*thumb2_negdi2): Likewise.
3310 (*thumb2_abssi2): Likewise.
3311 (*thumb2_neg_abss): Likewise.
3312 (*thumb2_movsi_insn): Likewise.
3313 (tls_load_dot_plus_four): Likewise.
3314 (*thumb2_movhi_insn): Likewise.
3315 (*thumb2_mov_scc): Likewise.
3316 (*thumb2_mov_negs): Likewise.
3317 (*thumb2_mov_negs): Likewise.
3318 (*thumb2_mov_nots): Likewise.
3319 (*thumb2_mov_nots): Likewise.
3320 (*thumb2_movsicc_): Likewise.
3321 (*thumb2_movsfcc_soft_insn): Likewise.
3322 (*thumb2_indirect_jump): Likewise.
3323 (*thumb2_and_scc): Likewise.
3324 (*thumb2_ior_scc): Likewise.
3325 (*thumb2_ior_scc_strict_it): Likewise.
3326 (*thumb2_cond_move): Likewise.
3327 (*thumb2_cond_arith): Likewise.
3328 (*thumb2_cond_ari): Likewise.
3329 (*thumb2_cond_sub): Likewise.
3330 (*thumb2_negscc): Likewise.
3331 (*thumb2_movcond): Likewise.
3332 (thumb2_casesi_internal): Likewise.
3333 (thumb2_casesi_internal_pic): Likewise.
3334 (*thumb2_alusi3_short): Likewise.
3335 (*thumb2_mov<mode>_shortim): Likewise.
3336 (*thumb2_addsi_short): Likewise.
3337 (*thumb2_subsi_short): Likewise.
3338 (thumb2_addsi3_compare0): Likewise.
3339 (*thumb2_cbz): Likewise.
3340 (*thumb2_cbnz): Likewise.
3341 (*thumb2_one_cmplsi2_short): Likewise.
3342 (*thumb2_negsi2_short): Likewise.
3343 (*orsi_notsi_si): Likewise.
3344 * config/arm/arm1020e.md: Update with new attributes.
3345 * config/arm/arm1026ejs.md: Update with new attributes.
3346 * config/arm/arm1136jfs.md: Update with new attributes.
3347 * config/arm/arm926ejs.md: Update with new attributes.
3348 * config/arm/cortex-a15.md: Update with new attributes.
3349 * config/arm/cortex-a5.md: Update with new attributes.
3350 * config/arm/cortex-a53.md: Update with new attributes.
3351 * config/arm/cortex-a7.md: Update with new attributes.
3352 * config/arm/cortex-a8.md: Update with new attributes.
3353 * config/arm/cortex-a9.md: Update with new attributes.
3354 * config/arm/cortex-m4.md: Update with new attributes.
3355 * config/arm/cortex-r4.md: Update with new attributes.
3356 * config/arm/fa526.md: Update with new attributes.
3357 * config/arm/fa606te.md: Update with new attributes.
3358 * config/arm/fa626te.md: Update with new attributes.
3359 * config/arm/fa726te.md: Update with new attributes.
3361 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3363 * config/aarch64/aarch64-simd.md
3364 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
3365 <vwx> iterator to ensure correct register choice.
3366 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3367 (aarch64_sqdmull_n<mode>): Likewise.
3368 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3369 * config/aarch64/arm_neon.h
3370 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
3371 (vml<as><q>_n_<su>16): Likewise.
3372 (vml<as>l_high_lane<q>_<su>16): Likewise.
3373 (vml<as>l_high_n_<su>16): Likewise.
3374 (vml<as>l_lane<q>_<su>16): Likewise.
3375 (vml<as>l_n_<su>16): Likewise.
3376 (vmul<q>_lane<q>_<su>16): Likewise.
3377 (vmul<q>_n_<su>16): Likewise.
3378 (vmull_lane<q>_<su>16): Likewise.
3379 (vmull_n_<su>16): Likewise.
3380 (vmull_high_lane<q>_<su>16): Likewise.
3381 (vmull_high_n_<su>16): Likewise.
3382 (vqrdmulh<q>_n_s16): Likewise.
3384 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
3386 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
3387 have the correct lane parameter.
3389 2013-09-06 Richard Biener <rguenther@suse.de>
3391 * cfganal.c (control_dependences::~control_dependences):
3392 Properly free all of the vector.
3394 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
3397 * config/i386/i386.c (ix86_conditional_register_usage):
3398 Proper initialize extended SSE registers.
3400 2013-09-06 Jan Hubicka <jh@suse.cz>
3402 PR tree-optimization/58311
3403 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
3405 2013-09-06 Jan Hubicka <jh@suse.cz>
3407 * Makefile.in (tree-sra.o): Update dependencies.
3408 * tree-sra.c: Include ipa-utils.h
3409 (scan_function): Use recursive_call_p.
3410 (has_caller_p): New function.
3411 (cgraph_for_node_and_aliases): Count also callers of aliases.
3413 2013-09-06 Jan Hubicka <jh@suse.cz>
3416 * cgraph.h (symtab_semantically_equivalent_p): Declare.
3417 * tree-tailcall.c: Include ipa-utils.h.
3418 (find_tail_calls): Use it.
3419 * ipa-pure-const.c (check_call): Likewise.
3420 * ipa-utils.c (recursive_call_p): New function.
3421 * ipa-utils.h (recursive_call_p): Dclare.
3422 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
3423 (symtab_semantically_equivalent_p): New function.
3424 * Makefile.in (tree-tailcall.o): Update dependencies.
3426 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3428 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
3431 2013-09-06 Richard Biener <rguenther@suse.de>
3433 * lto-streamer.h (lto_global_var_decls): Remove.
3434 * Makefile.in (OBJS): Remove lto-symtab.o.
3435 (lto-symtab.o): Remove.
3436 (GTFILES): Remove lto-symtab.c
3437 * lto-symtab.c: Move to lto/
3439 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3441 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
3442 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
3443 (UNSPEC_FPINT_RINT): New constant definitions.
3444 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
3445 definition with 2 attributes.
3446 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
3447 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
3450 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3452 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
3453 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
3454 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
3455 attribute to "z196_alone".
3456 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
3459 2013-09-06 Richard Biener <rguenther@suse.de>
3461 * basic-block.h (class control_dependences): New.
3462 * tree-ssa-dce.c (control_dependence_map): Remove.
3464 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
3465 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
3466 find_pdom, find_control_dependence, find_all_control_dependences):
3468 (mark_control_dependent_edges_necessary,
3469 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
3470 tree_dce_done, perform_tree_ssa_dce): Adjust.
3471 * cfganal.c (set_control_dependence_map_bit,
3472 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
3473 find_all_control_dependences): Move from tree-ssa-dce.c and
3474 implement as methods of control_dependences class.
3475 (control_dependences::control_dependences): New.
3476 (control_dependences::~control_dependences): Likewise.
3477 (control_dependences::get_edges_dependent_on): Likewise.
3478 (control_dependences::get_edge): Likewise.
3480 2013-09-04 Jan Hubicka <jh@suse.cz>
3482 * tree.c (types_same_for_odr): Drop overactive check.
3483 * ipa-devirt.c (hash_type_name): Likewise.
3485 2013-09-04 Jan Hubicka <jh@suse.cz>
3487 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
3488 (analyze_functions): ... here.
3490 2013-09-04 Jan Hubicka <jh@suse.cz>
3493 * cgraphunit.c (analyze_functions): Clear AUX fields
3494 after processing; initialize assembler name has.
3496 2013-09-05 Jeff Law <law@redhat.com>
3498 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
3499 from thread_around_empty_block. Record threading path into PATH.
3500 Recurse if threading through the initial block is successful.
3501 (thread_across_edge): Corresponding changes to slightly simplify.
3503 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3505 * config/aarch64/aarch64.md
3506 (type): Remove frecpe, frecps, frecpx.
3507 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
3508 fix to be a TARGET_SIMD instruction.
3509 (aarch64_frecps): Remove.
3510 * config/aarch64/aarch64-simd.md
3511 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
3512 (aarch64_frecps<mode>): Handle all float/vector of float modes.
3514 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3515 Sofiane Naci <sofiane.naci@arm.com>
3517 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
3518 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
3519 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
3520 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
3521 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
3522 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
3523 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
3524 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
3525 into "clz, "rbit". Rename "shift" to "shift_imm".
3526 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
3527 changes. Update for attribute changes all occurrences of arlo_* and
3529 * config/arm/arm-fixed.md: Update for attribute changes
3530 all occurrences of arlo_* types.
3531 * config/arm/thumb2.md: Update for attribute changes all occurrences
3533 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
3534 (cortexa7_older_only): Likewise.
3535 (cortexa7_younger): Likewise.
3536 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3537 (1020alu_shift_op): Likewise.
3538 (1020alu_shift_reg_op): Likewise.
3539 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3540 (alu_shift_op): Likewise.
3541 (alu_shift_reg_op): Likewise.
3542 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3543 (11_alu_shift_op): Likewise.
3544 (11_alu_shift_reg_op): Likewise.
3545 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3546 (9_alu_shift_reg_op): Likewise.
3547 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3549 (cortex_a15_alu_shift): Likewise.
3550 (cortex_a15_alu_shift_reg): Likewise.
3551 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3553 (cortex_a5_alu_shift): Likewise.
3554 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3556 (cortex_a53_alu_shift): Likewise.
3557 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3559 (cortex_a7_alu_reg): Likewise.
3560 (cortex_a7_alu_shift): Likewise.
3561 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3563 (cortex_a8_alu_shift): Likewise.
3564 (cortex_a8_alu_shift_reg): Likewise.
3565 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3566 (cortex_a9_dp_shift): Likewise.
3567 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3569 * config/arm/cortex-r4.md
3570 (cortex_r4_alu): Update for attribute changes.
3571 (cortex_r4_mov): Likewise.
3572 (cortex_r4_alu_shift_reg): Likewise.
3573 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3574 (526_alu_shift_op): Likewise.
3575 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3576 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3577 (626te_alu_shift_op): Likewise.
3578 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3579 (726te_alu_shift_op): Likewise.
3580 (726te_alu_shift_reg_op): Likewise.
3581 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3582 (mp626_alu_shift_op): Likewise.
3583 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3584 (pj4_alu_conds): Likewise.
3585 (pj4_shift): Likewise.
3586 (pj4_shift_conds): Likewise.
3587 (pj4_alu_shift): Likewise.
3588 (pj4_alu_shift_conds): Likewise.
3589 * config/aarch64/aarch64.md: Update for attribute change
3590 all occurrences of arlo_* and shift* types.
3592 2013-09-05 Mike Stump <mikestump@comcast.net>
3594 * tree.h: Move documentation for tree_function_decl to tree-core.h
3595 with the declaration.
3597 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3600 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3601 looking for widest mode.
3603 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3605 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3607 2013-09-05 Richard Biener <rguenther@suse.de>
3609 PR tree-optimization/58137
3610 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3611 Do not create vectors of pointers.
3612 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3613 types for the components of the vector initializer.
3614 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3615 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3617 2013-09-05 Martin Jambor <mjambor@suse.cz>
3619 * ipa-prop.c (remove_described_reference): Accept missing references,
3620 return false if that hppens, otherwise return true.
3621 (cgraph_node_for_jfunc): New function.
3622 (try_decrement_rdesc_refcount): Likewise.
3623 (try_make_edge_direct_simple_call): Use them.
3624 (ipa_edge_removal_hook): Remove references from rdescs.
3625 (ipa_edge_duplication_hook): Clone rdescs and their references
3626 when the new edge has the same caller as the old one.
3627 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3628 reference before removing any edges.
3630 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3632 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3634 * thumb2.md (thumb2_storewb_parisi): New pattern.
3636 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3638 * config/aarch64/aarch64-option-extensions.def: Add
3639 AARCH64_OPT_EXTENSION of 'crc'.
3640 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3641 (AARCH64_ISA_CRC): Ditto.
3642 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3643 description of the CRC extension.
3645 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3647 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3648 * config/rs6000/linux.h: Ditto.
3649 * alpha/linux.h: Ditto.
3650 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3651 no_c99_libc_has_function.
3652 * config/c6x/uclinux-elf.h: Ditto.
3653 * config/lm32/uclinux-elf.h: Ditto.
3654 * config/m68k/uclinux.h: Ditto.
3655 * config/moxie/uclinux.h: Ditto.
3656 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3657 (crisv32-*-linux*, cris-*-linux*): Ditto.
3658 * config/bfin/bfin.c: Include "tm_p.h".
3660 2013-09-05 Richard Biener <rguenther@suse.de>
3662 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3663 check for a definition without a basic-block.
3665 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3666 Sofiane Naci <sofiane.naci@arm.com>
3668 * config/aarch64/aarch64.md
3669 (*movti_aarch64): Rename r_2_f and f_2_r.
3670 (*movsf_aarch64): Likewise.
3671 (*movdf_aarch64): Likewise.
3672 (*movtf_aarch64): Likewise.
3673 (aarch64_movdi_<mode>low): Likewise.
3674 (aarch64_movdi_<mode>high): Likewise.
3675 (aarch64_mov<mode>high_di): Likewise.
3676 (aarch64_mov<mode>low_di): Likewise.
3677 (aarch64_movtilow_tilow): Likewise.
3678 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3679 values to config/arm/types.md
3680 (attribute "conds"): Update for attribute change.
3681 (anddi3_insn): Likewise.
3682 (iordi3_insn): Likewise.
3683 (xordi3_insn): Likewise.
3684 (one_cmpldi2): Likewise.
3685 * config/arm/types.md (type): Add Neon types.
3686 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3687 use "type" attribute.
3688 (movmisalign<mode>_neon_store): Likewise.
3689 (movmisalign<mode>_neon_load): Likewise.
3690 (vec_set<mode>_internal): Likewise.
3691 (vec_setv2di_internal): Likewise.
3692 (vec_extract<mode>): Likewise.
3693 (vec_extractv2di): Likewise.
3694 (add<mode>3_neon): Likewise.
3695 (adddi3_neon): Likewise.
3696 (sub<mode>3_neon): Likewise.
3697 (subdi3_neon): Likewise.
3698 (mul<mode>3_neon): Likewise.
3699 (mul<mode>3add<mode>_neon): Likewise.
3700 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3701 (fma<VCVTF:mode>4)): Likewise.
3702 (fma<VCVTF:mode>4_intrinsic): Likewise.
3703 (fmsub<VCVTF:mode>4)): Likewise.
3704 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3705 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3706 (ior<mode>3): Likewise.
3707 (and<mode>3): Likewise.
3708 (anddi3_neon): Likewise.
3709 (orn<mode>3_neon): Likewise.
3710 (orndi3_neon): Likewise.
3711 (bic<mode>3_neon): Likewise.
3712 (bicdi3_neon): Likewise.
3713 (xor<mode>3): Likewise.
3714 (one_cmpl<mode>2): Likewise.
3715 (abs<mode>2): Likewise.
3716 (neg<mode>2): Likewise.
3717 (umin<mode>3_neon): Likewise.
3718 (umax<mode>3_neon): Likewise.
3719 (smin<mode>3_neon): Likewise.
3720 (smax<mode>3_neon): Likewise.
3721 (vashl<mode>3): Likewise.
3722 (vashr<mode>3_imm): Likewise.
3723 (vlshr<mode>3_imm): Likewise.
3724 (ashl<mode>3_signed): Likewise.
3725 (ashl<mode>3_unsigned): Likewise.
3726 (neon_load_count): Likewise.
3727 (ashldi3_neon_noclobber): Likewise.
3728 (signed_shift_di3_neon): Likewise.
3729 (unsigned_shift_di3_neon): Likewise.
3730 (ashrdi3_neon_imm_noclobber): Likewise.
3731 (lshrdi3_neon_imm_noclobber): Likewise.
3732 (widen_ssum<mode>3): Likewise.
3733 (widen_usum<mode>3): Likewise.
3734 (quad_halves_<code>v4si): Likewise.
3735 (quad_halves_<code>v4sf): Likewise.
3736 (quad_halves_<code>v8hi): Likewise.
3737 (quad_halves_<code>v16qi): Likewise.
3738 (reduc_splus_v2di): Likewise.
3739 (neon_vpadd_internal<mode>): Likewise.
3740 (neon_vpsmin<mode>): Likewise.
3741 (neon_vpsmax<mode>): Likewise.
3742 (neon_vpumin<mode>): Likewise.
3743 (neon_vpumax<mode>): Likewise.
3744 (ss_add<mode>_neon): Likewise.
3745 (us_add<mode>_neon): Likewise.
3746 (ss_sub<mode>_neon): Likewise.
3747 (us_sub<mode>_neon): Likewise.
3748 (neon_vadd<mode>_unspec): Likewise.
3749 (neon_vaddl<mode>): Likewise.
3750 (neon_vaddw<mode>): Likewise.
3751 (neon_vhadd<mode>): Likewise.
3752 (neon_vqadd<mode>): Likewise.
3753 (neon_vaddhn<mode>): Likewise.
3754 (neon_vmul<mode>): Likewise.
3755 (neon_vmla<mode>): Likewise.
3756 (neon_vmlal<mode>): Likewise.
3757 (neon_vmls<mode>): Likewise.
3758 (neon_vmlsl<mode>): Likewise.
3759 (neon_vqdmulh<mode>): Likewise.
3760 (neon_vqdmlal<mode>): Likewise.
3761 (neon_vqdmlsl<mode>): Likewise.
3762 (neon_vmull<mode>): Likewise.
3763 (neon_vqdmull<mode>): Likewise.
3764 (neon_vsub<mode>_unspec): Likewise.
3765 (neon_vsubl<mode>): Likewise.
3766 (neon_vsubw<mode>): Likewise.
3767 (neon_vqsub<mode>): Likewise.
3768 (neon_vhsub<mode>): Likewise.
3769 (neon_vsubhn<mode>): Likewise.
3770 (neon_vceq<mode>): Likewise.
3771 (neon_vcge<mode>): Likewise.
3772 (neon_vcgeu<mode>): Likewise.
3773 (neon_vcgt<mode>): Likewise.
3774 (neon_vcgtu<mode>): Likewise.
3775 (neon_vcle<mode>): Likewise.
3776 (neon_vclt<mode>): Likewise.
3777 (neon_vcage<mode>): Likewise.
3778 (neon_vcagt<mode>): Likewise.
3779 (neon_vtst<mode>): Likewise.
3780 (neon_vabd<mode>): Likewise.
3781 (neon_vabdl<mode>): Likewise.
3782 (neon_vaba<mode>): Likewise.
3783 (neon_vabal<mode>): Likewise.
3784 (neon_vmax<mode>): Likewise.
3785 (neon_vmin<mode>): Likewise.
3786 (neon_vpaddl<mode>): Likewise.
3787 (neon_vpadal<mode>): Likewise.
3788 (neon_vpmax<mode>): Likewise.
3789 (neon_vpmin<mode>): Likewise.
3790 (neon_vrecps<mode>): Likewise.
3791 (neon_vrsqrts<mode>): Likewise.
3792 (neon_vqabs<mode>): Likewise.
3793 (neon_vqneg<mode>): Likewise.
3794 (neon_vcls<mode>): Likewise.
3795 (clz<mode>2): Likewise.
3796 (popcount<mode>2): Likewise.
3797 (neon_vrecpe): Likewise.
3798 (neon_vrsqrte): Likewise.
3799 (neon_vget_lane<mode>_sext_internal): Likewise.
3800 (neon_vget_lane<mode>_zext_internal): Likewise.
3801 (neon_vdup_n<mode>): Likewise.
3802 (neon_vdup_nv2di): Likewise.
3803 (neon_vdpu_lane<mode>_internal): Likewise.
3804 (neon_vswp<mode>): Likewise.
3805 (float<mode><V_cvtto>2): Likewise.
3806 (floatuns<mode><V_cvtto>2): Likewise.
3807 (fix_trunc<mode><V_cvtto>)2): Likewise
3808 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3809 (neon_vcvt<mode>): Likewise.
3810 (neon_vcvtv4sfv4hf): Likewise.
3811 (neon_vcvtv4hfv4sf): Likewise.
3812 (neon_vcvt_n<mode>): Likewise.
3813 (neon_vmovn<mode>): Likewise.
3814 (neon_vqmovn<mode>): Likewise.
3815 (neon_vqmovun<mode>): Likewise.
3816 (neon_vmovl<mode>): Likewise.
3817 (neon_vmul_lane<mode>): Likewise.
3818 (neon_vmull_lane<mode>): Likewise.
3819 (neon_vqdmull_lane<mode>): Likewise.
3820 (neon_vqdmulh_lane<mode>): Likewise.
3821 (neon_vmla_lane<mode>): Likewise.
3822 (neon_vmlal_lane<mode>): Likewise.
3823 (neon_vqdmlal_lane<mode>): Likewise.
3824 (neon_vmls_lane<mode>): Likewise.
3825 (neon_vmlsl_lane<mode>): Likewise.
3826 (neon_vqdmlsl_lane<mode>): Likewise.
3827 (neon_vext<mode>): Likewise.
3828 (neon_vrev64<mode>): Likewise.
3829 (neon_vrev32<mode>): Likewise.
3830 (neon_vrev16<mode>): Likewise.
3831 (neon_vbsl<mode>_internal): Likewise.
3832 (neon_vshl<mode>): Likewise.
3833 (neon_vqshl<mode>): Likewise.
3834 (neon_vshr_n<mode>): Likewise.
3835 (neon_vshrn_n<mode>): Likewise.
3836 (neon_vqshrn_n<mode>): Likewise.
3837 (neon_vqshrun_n<mode>): Likewise.
3838 (neon_vshl_n<mode>): Likewise.
3839 (neon_vqshl_n<mode>): Likewise.
3840 (neon_vqshlu_n<mode>): Likewise.
3841 (neon_vshll_n<mode>): Likewise.
3842 (neon_vsra_n<mode>): Likewise.
3843 (neon_vsri_n<mode>): Likewise.
3844 (neon_vsli_n<mode>): Likewise.
3845 (neon_vtbl1v8qi): Likewise.
3846 (neon_vtbl2v8qi): Likewise.
3847 (neon_vtbl3v8qi): Likewise.
3848 (neon_vtbl4v8qi): Likewise.
3849 (neon_vtbx1v8qi): Likewise.
3850 (neon_vtbx2v8qi): Likewise.
3851 (neon_vtbx3v8qi): Likewise.
3852 (neon_vtbx4v8qi): Likewise.
3853 (neon_vtrn<mode>_internal): Likewise.
3854 (neon_vzip<mode>_internal): Likewise.
3855 (neon_vuzp<mode>_internal): Likewise.
3856 (neon_vld1<mode>): Likewise.
3857 (neon_vld1_lane<mode>): Likewise.
3858 (neon_vld1_dup<mode>): Likewise.
3859 (neon_vld1_dupv2di): Likewise.
3860 (neon_vst1<mode>): Likewise.
3861 (neon_vst1_lane<mode>): Likewise.
3862 (neon_vld2<mode>): Likewise.
3863 (neon_vld2_lane<mode>): Likewise.
3864 (neon_vld2_dup<mode>): Likewise.
3865 (neon_vst2<mode>): Likewise.
3866 (neon_vst2_lane<mode>): Likewise.
3867 (neon_vld3<mode>): Likewise.
3868 (neon_vld3qa<mode>): Likewise.
3869 (neon_vld3qb<mode>): Likewise.
3870 (neon_vld3_lane<mode>): Likewise.
3871 (neon_vld3_dup<mode>): Likewise.
3872 (neon_vst3<mode>): Likewise.
3873 (neon_vst3qa<mode>): Likewise.
3874 (neon_vst3qb<mode>): Likewise.
3875 (neon_vst3_lane<mode>): Likewise.
3876 (neon_vld4<mode>): Likewise.
3877 (neon_vld4qa<mode>): Likewise.
3878 (neon_vld4qb<mode>): Likewise.
3879 (neon_vld4_lane<mode>): Likewise.
3880 (neon_vld4_dup<mode>): Likewise.
3881 (neon_vst4<mode>): Likewise.
3882 (neon_vst4qa<mode>): Likewise.
3883 (neon_vst4qb<mode>): Likewise.
3884 (neon_vst4_lane<mode>): Likewise.
3885 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3886 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3887 (neon_vec_<US>mult_lo_<mode>): Likewise.
3888 (neon_vec_<US>mult_hi_<mode>): Likewise.
3889 (neon_vec_<US>shiftl_<mode>): Likewise.
3890 (neon_unpack<US>_<mode>): Likewise.
3891 (neon_vec_<US>mult_<mode>): Likewise.
3892 (vec_pack_trunc_<mode>): Likewise.
3893 (neon_vec_pack_trunk_<mode>): Likewise.
3894 (neon_vabd<mode>_2): Likewise.
3895 (neon_vabd<mode>_3): Likewise.
3896 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3897 (thumb2_movsi_vfp): Likewise.
3898 (movdi_vfp): Likewise.
3899 (movdi_vfp_cortexa8): Likewise.
3900 (movhf_vfp_neon): Likewise.
3901 (movhf_vfp): Likewiwse.
3902 (movsf_vfp): Likewiwse.
3903 (thumb2_movsf_vfp): Likewiwse.
3904 (movdf_vfp): Likewise.
3905 (thumb2_movdf_vfp): Likewise.
3906 (movsfcc_vfp): Likewise.
3907 (thumb2_movsfcc_vfp): Likewise.
3908 (movdfcc_vfp): Likewise.
3909 (thumb2_movdfcc_vfp): Likewise.
3910 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3911 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3912 (v10_v2c): Likewise.
3913 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3915 (cortex_a15_neon_int_2): Likewise.
3916 (cortex_a15_neon_int_3): Likewise.
3917 (cortex_a15_neon_int_4): Likewise.
3918 (cortex_a15_neon_int_5): Likewise.
3919 (cortex_a15_neon_vqneg_vqabs): Likewise.
3920 (cortex_a15_neon_vmov): Likewise.
3921 (cortex_a15_neon_vaba): Likewise.
3922 (cortex_a15_neon_vaba_qqq): Likewise.
3923 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3924 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3925 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3926 scalar_64_32_long_scalar): Likewise.
3927 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3928 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3929 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3930 lotype_qdd_64_32_long): Likewise.
3931 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3932 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3933 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3934 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3935 (cortex_a15_neon_shift_1): Likewise.
3936 (cortex_a15_neon_shift_2): Likewise.
3937 (cortex_a15_neon_shift_3): Likewise.
3938 (cortex_a15_neon_vshl_ddd): Likewise.
3939 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3940 (cortex_a15_neon_vsra_vrsra): Likewise.
3941 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3942 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3943 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3944 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3945 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3946 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3947 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3948 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3949 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3950 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3951 (cortex_a15_neon_bp_simple): Likewise.
3952 (cortex_a15_neon_bp_2cycle): Likewise.
3953 (cortex_a15_neon_bp_3cycle): Likewise.
3954 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3955 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3956 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3957 (cortex_a15_neon_vld2_4_regs): Likewise.
3958 (cortex_a15_neon_vld3_vld4): Likewise.
3959 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3960 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3961 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3962 (cortex_a15_neon_vst3_vst4): Likewise.
3963 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3964 (cortex_a15_neon_vld3_vld4_lane" 10
3965 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3966 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3967 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3968 (cortex_a15_neon_ldm_2): Likewise.0
3969 (cortex_a15_neon_stm_2): Likewise.
3970 (cortex_a15_neon_mcr): Likewise.
3971 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3972 (cortex_a15_neon_mrc): Likewise.
3973 (cortex_a15_neon_mrrc): Likewise.
3974 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3976 (cortex_a15_alu_shift): Likewise.
3977 (cortex_a15_alu_shift_reg): Likewise.
3978 (cortex_a15_mult32): Likewise.
3979 (cortex_a15_mult64): Likewise.
3980 (cortex_a15_block): Likewise.
3981 (cortex_a15_branch): Likewise.
3982 (cortex_a15_load1): Likewise.
3983 (cortex_a15_load3): Likewise.
3984 (cortex_a15_store1): Likewise.
3985 (cortex_a15_store3): Likewise.
3986 (cortex_a15_call): Likewise.
3987 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3988 (cortex_a5_f2r): Likewise.
3989 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3991 (cortex_a53_f2r): Likewise.
3992 * config/arm/cortex-a7.md
3993 (cortex_a7_branch): Update for attribute change.
3994 (cortex_a7_call): Likewise.
3995 (cortex_a7_alu_imm): Likewise.
3996 (cortex_a7_alu_reg): Likewise.
3997 (cortex_a7_alu_shift): Likewise.
3998 (cortex_a7_mul): Likewise.
3999 (cortex_a7_load1): Likewise.
4000 (cortex_a7_store1): Likewise.
4001 (cortex_a7_load2): Likewise.
4002 (cortex_a7_store2): Likewise.
4003 (cortex_a7_load3): Likewise.
4004 (cortex_a7_store3): Likewise.
4005 (cortex_a7_load4): Likewise.
4006 (cortex_a7_store4): Likewise.
4007 (cortex_a7_fpalu): Likewise.
4008 (cortex_a7_fconst): Likewise.
4009 (cortex_a7_fpmuls): Likewise.
4010 (cortex_a7_neon_mul): Likewise.
4011 (cortex_a7_fpmacs): Likewise.
4012 (cortex_a7_neon_mla: Likewise.
4013 (cortex_a7_fpmuld: Likewise.
4014 (cortex_a7_fpmacd: Likewise.
4015 (cortex_a7_fpfmad: Likewise.
4016 (cortex_a7_fdivs: Likewise.
4017 (cortex_a7_fdivd: Likewise.
4018 (cortex_a7_r2f: Likewise.
4019 (cortex_a7_f2r: Likewise.
4020 (cortex_a7_f_flags: Likewise.
4021 (cortex_a7_f_loads: Likewise.
4022 (cortex_a7_f_loadd: Likewise.
4023 (cortex_a7_f_stores: Likewise.
4024 (cortex_a7_f_stored: Likewise.
4025 (cortex_a7_neon): Likewise.
4026 * config/arm/cortex-a8-neon.md
4027 (cortex_a8_neon_mrc): Update for attribute change.
4028 (cortex_a8_neon_mrrc): Likewise.
4029 (cortex_a8_neon_int_1): Likewise.
4030 (cortex_a8_neon_int_2): Likewise.
4031 (cortex_a8_neon_int_3): Likewise.
4032 (cortex_a8_neon_int_4): Likewise.
4033 (cortex_a8_neon_int_5): Likewise.
4034 (cortex_a8_neon_vqneg_vqabs): Likewise.
4035 (cortex_a8_neon_vmov): Likewise.
4036 (cortex_a8_neon_vaba): Likewise.
4037 (cortex_a8_neon_vaba_qqq): Likewise.
4038 (cortex_a8_neon_vsma): Likewise.
4039 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4040 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4041 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4042 long_scalar): Likewise.
4043 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4044 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4045 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4046 long_scalar_qdd_64_32_long): Likewise.
4047 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4048 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4049 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4050 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4051 (cortex_a8_neon_shift_1): Likewise.
4052 (cortex_a8_neon_shift_2): Likewise.
4053 (cortex_a8_neon_shift_3): Likewise.
4054 (cortex_a8_neon_vshl_ddd): Likewise.
4055 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4056 (cortex_a8_neon_vsra_vrsra): Likewise.
4057 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4058 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4059 (cortex_a8_neon_fp_vsum): Likewise.
4060 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4061 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4062 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4063 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4064 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4065 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4066 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4067 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4068 (cortex_a8_neon_bp_simple): Likewise.
4069 (cortex_a8_neon_bp_2cycle): Likewise.
4070 (cortex_a8_neon_bp_3cycle): Likewise.
4071 (cortex_a8_neon_ldr): Likewise.
4072 (cortex_a8_neon_str): Likewise.
4073 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4074 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4075 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4076 (cortex_a8_neon_vld2_4_regs): Likewise.
4077 (cortex_a8_neon_vld3_vld4): Likewise.
4078 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4079 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4080 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4081 (cortex_a8_neon_vst3_vst4): Likewise.
4082 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4083 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4084 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4085 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4086 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4087 (cortex_a8_neon_mcr): Likewise.
4088 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4089 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4090 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4092 (ca9_neon_mrrc): Likewise.
4093 (cortex_a9_neon_int_1): Likewise.
4094 (cortex_a9_neon_int_2): Likewise.
4095 (cortex_a9_neon_int_3): Likewise.
4096 (cortex_a9_neon_int_4): Likewise.
4097 (cortex_a9_neon_int_5): Likewise.
4098 (cortex_a9_neon_vqneg_vqabs): Likewise.
4099 (cortex_a9_neon_vmov): Likewise.
4100 (cortex_a9_neon_vaba): Likewise.
4101 (cortex_a9_neon_vaba_qqq): Likewise.
4102 (cortex_a9_neon_vsma): Likewise.
4103 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4104 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4105 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4106 long_scalar): Likewise.
4107 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4108 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4109 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4110 long_scalar_qdd_64_32_long): Likewise.
4111 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4112 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4113 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4114 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4115 (cortex_a9_neon_shift_1): Likewise.
4116 (cortex_a9_neon_shift_2): Likewise.
4117 (cortex_a9_neon_shift_3): Likewise.
4118 (cortex_a9_neon_vshl_ddd): Likewise.
4119 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4120 (cortex_a9_neon_vsra_vrsra): Likewise.
4121 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4122 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4123 (cortex_a9_neon_fp_vsum): Likewise.
4124 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4125 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4126 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4127 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4128 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4129 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4130 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4131 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4132 (cortex_a9_neon_bp_simple): Likewise.
4133 (cortex_a9_neon_bp_2cycle): Likewise.
4134 (cortex_a9_neon_bp_3cycle): Likewise.
4135 (cortex_a9_neon_ldr): Likewise.
4136 (cortex_a9_neon_str): Likewise.
4137 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4138 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4139 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4140 (cortex_a9_neon_vld2_4_regs): Likewise.
4141 (cortex_a9_neon_vld3_vld4): Likewise.
4142 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4143 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4144 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4145 (cortex_a9_neon_vst3_vst4): Likewise.
4146 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4147 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4148 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4149 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4150 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4151 (cortex_a9_neon_mcr): Likewise.
4152 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4153 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4154 (cortex_a9_fps): Likewise.
4155 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
4157 (cortex_m4_fmuls): Likewise.
4158 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
4160 (cortex_r4_mrc): Likewise.
4161 * config/arm/iterators.md: Update comment referring to neon_type.
4162 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
4163 (iwmmxt_movsi_insn): Likewise.
4164 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
4166 (pj4_core_to_vfp): Likewise.
4167 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
4169 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
4170 (vfp_fstore): Likewise.
4171 * doc/md.texi: Change references to neon_type to refer to type.
4173 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4175 * tree.h (DECL_BUILT_IN): Fix typo in comment.
4177 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
4179 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
4182 2013-09-04 Easwaran Raman <eraman@google.com>
4185 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
4186 (build_and_add_sum): Use it.
4187 (appears_later_in_bb): Simplify code.
4189 2013-09-04 Teresa Johnson <tejohnson@google.com>
4191 * dumpfile.c (dump_finish): Don't close stderr/stdout.
4193 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
4195 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
4197 2013-09-04 Jan Hubicka <jh@suse.cz>
4199 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
4200 * ipa-devirt.c: Include diganostic.h
4201 (odr_type_d): Add types and types_set.
4202 (hash_type_name): Work for types with vtables during LTO.
4203 (odr_hasher::remove): Fix comment; destroy types_set.
4204 (add_type_duplicate): New function,
4205 (get_odr_type): Use it.
4206 (dump_type_inheritance_graph): Dump type duplicates.
4207 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
4209 * tree.c (types_same_for_odr): Give exact answers on types with
4212 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4214 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
4215 explaining their differences.
4217 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
4219 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
4221 2013-09-03 Jeff Law <law@redhat.com>
4223 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
4224 when not threading through a joiner block. Pass joiner/no joiner
4225 state to register_jump_thread.
4226 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
4227 state from argument rather than implying on path length.
4228 Dump the entire jump thread path into debugging dump.
4229 * tree-flow.h (register_jump_thread): Update prototype.
4231 2013-08-29 Xinliang David Li <davidxl@google.com>
4233 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4234 Remove a trivial gcc_assert.
4236 2013-08-29 Xinliang David Li <davidxl@google.com>
4238 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4239 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
4240 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4241 Delay base decl alignment adjustment.
4242 * tree-vectorizer.c (vect_destroy_datarefs): New function.
4243 * tree-vectorizer.h: New data structure.
4244 (set_dr_misalignment): New function.
4245 (dr_misalignment): Ditto.
4246 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
4247 (vectorizable_load): Ditto.
4248 (ensure_base_align): New function.
4249 (vectorize_loops): Add dbg_cnt support.
4250 (execute_vect_slp): Ditto.
4251 * dbgcnt.def: New debug counter.
4252 * Makefile: New dependency.
4254 2013-09-03 Meador Inge <meadori@codesourcery.com>
4258 2013-08-30 Meador Inge <meadori@codesourcery.com>
4260 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4262 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4264 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4265 function descriptor.
4267 2013-09-03 Richard Biener <rguenther@suse.de>
4269 * tree-affine.c (add_elt_to_tree): Fix association issue,
4270 avoid useless converts and make sure to always return a
4271 properly typed result.
4273 2013-09-03 Richard Biener <rguenther@suse.de>
4276 * fold-const.c (negate_expr_p): Fix division case.
4277 (negate_expr): Likewise.
4279 2013-09-03 Richard Biener <rguenther@suse.de>
4282 * tree-streamer-out.c: Include tm.h.
4283 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
4285 2013-09-03 Jan Hubicka <jh@suse.cz>
4287 * tree-profile.c (tree_profiling): Cleanup CFG when done.
4289 2013-09-03 Alan Modra <amodra@gmail.com>
4291 * config.gcc (powerpc*-*-linux*): Add support for little-endian
4292 multilibs to big-endian target and vice versa.
4293 * config/rs6000/t-linux64: Use := assignment on all vars.
4294 (MULTILIB_EXTRA_OPTS): Remove fPIC.
4295 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
4296 * config/rs6000/t-linux64le: New file.
4297 * config/rs6000/t-linux64bele: New file.
4298 * config/rs6000/t-linux64lebe: New file.
4300 2013-09-02 Jan Hubicka <jh@suse.cz>
4302 * ipa-inline-transform.c (inline_transform): Do not
4303 optimize_inline_calls when not optimizing.
4305 2013-09-02 Jan Hubicka <jh@suse.cz>
4307 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
4308 duplicated nodes for assembler names.
4309 * symtab.c (symtab_unregister_node): Do not attempt to unlink
4310 hard registers from assembler name hash.
4312 2013-09-02 Jan Hubicka <jh@suse.cz>
4314 * ipa-split.c (execute_split_functions): Split externally visible
4315 functions called once.
4317 2013-09-02 Martin Jambor <mjambor@suse.cz>
4320 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
4321 linked list. When finding the correct duplicate, also consider also
4322 the caller in additon to its inlined_to node.
4324 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
4326 * config/aarch64/aarch64-simd-builtins.def
4327 (dup_lane_scalar): Remove.
4328 * config/aarch64/aarch64-simd.md
4329 (aarch64_simd_dup): Add 'w->w' alternative.
4330 (aarch64_dup_lane<mode>): Allow for VALL.
4331 (aarch64_dup_lane_scalar<mode>): Remove.
4332 (aarch64_dup_lane_<vswap_width_name><mode>): New.
4333 (aarch64_get_lane_signed<mode>): Add w->w altenative.
4334 (aarch64_get_lane_unsigned<mode>): Likewise.
4335 (aarch64_get_lane<mode>): Likewise.
4336 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
4337 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
4338 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
4339 (VCON): Change container of V2SF.
4340 (vswap_width_name): Likewise.
4341 * config/aarch64/arm_neon.h
4342 (__aarch64_vdup_lane_any): New.
4343 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4344 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
4345 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4347 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
4350 * expr.c (emit_move_complex): Do not move complex FP values as parts if
4351 the source or the destination is a single hard register.
4353 2013-09-02 Richard Biener <rguenther@suse.de>
4356 * tree-scalar-evolution.c (instantiate_scev_name): Allow
4359 2013-09-02 Richard Biener <rguenther@suse.de>
4361 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
4362 arithmetic to sizetype.
4364 2013-09-02 Bin Cheng <bin.cheng@arm.com>
4366 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
4367 Find auto-increment use both before and after candidate.
4369 2013-09-02 Marek Polacek <polacek@redhat.com>
4371 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
4373 2013-09-01 Jan Hubicka <jh@suse.cz>
4375 * Makefile.in: Add ipa-profile.o
4376 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
4377 * cgraph.c (struct cgraph_propagate_frequency_data,
4378 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
4379 ipa-profile.c; replace cgraph_ by ipa_ prefix.
4380 * cgraph.h (cgraph_propagate_frequency): Remove.
4381 * ipa-inline-analysis.c: Include ipa-utils.h;
4382 drop duplicated cfgloop.h.
4383 (inline_update_callee_summaries): Update.
4384 * ipa-profile.c: New file.
4385 * ipa-utils.h (ipa_propagate_frequency): Declare.
4386 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
4387 data-streamer.h, value-prof.h.
4388 (symtab_remove_unreachable_nodes): Update profile.
4389 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
4390 account_time_size, cmp_counts, dump_histogram,
4391 ipa_profile_generate_summary, ipa_profile_write_summary,
4392 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
4393 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
4394 Move to ipa-profile.c.
4396 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
4398 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
4400 2013-09-01 Jan Hubicka <jh@suse.cz>
4402 * common.opt (fdevirtualize-speculatively): New function.
4403 * invoke.texi (fdevirtualize-speculatively): Document.
4404 * ipa-devirt.c: Include ipa-inline.h
4405 (likely_target_p): New function.
4406 (ipa_devirt): New function.
4407 (gate_ipa_devirt): New function.
4408 (pass_data_ipa_devirt): New static var.
4409 (pass_ipa_devirt): Likewise.
4410 (make_pass_ipa_devirt): New function.
4411 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
4412 (common_handle_option): Disable devirtualization when
4413 value range profiling is available.
4414 * passes.def (pass_ipa_devirt): Add.
4415 * timever.def (TV_IPA_DEVIRT): New timevar.
4416 * tree-pass.h (make_pass_ipa_devirt):
4418 2013-09-01 Iain Sandoe <iain@codesourcery.com>
4420 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
4421 include sanitize(undefined).
4423 2013-08-31 Diego Novillo <dnovillo@google.com>
4425 * Makefile.in (TREE_CORE_H): Define.
4427 (GTFILES): Add tree-core.h.
4428 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
4430 * tree-core.h: New file.
4431 Move all data structures, enum, typedefs, global
4432 declarations and constants from ...
4435 2013-08-31 Jan Hubicka <jh@suse.cz>
4437 * bulitins.c (expand_builtin): Do not early exit for gcov
4438 instrumented functions.
4440 2013-08-31 Marek Polacek <polacek@redhat.com>
4442 * ubsan.c: Include tm_p.h.
4444 2013-08-31 Jan Hubicka <jh@suse.cz>
4446 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
4449 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
4450 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
4451 * tree-cfg.c (verify_gimple_label): ... here.
4452 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
4453 (ipa_merge_profiles): New function.
4454 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
4455 (lto_input_function_body): Likewise.
4456 * ipa-utils.h (ipa_merge_profiles): Declare.
4457 * lto-streamer.h (lto_input_function_body): Update prototype.
4458 (emit_label_in_global_context_p): Remove.
4459 * lto-symtab.c: Include ipa-utils.h
4460 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4462 2013-08-31 Jan Hubicka <jh@suse.cz>
4464 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
4466 2013-08-31 Jan Hubicka <jh@suse.cz>
4468 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
4470 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
4472 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
4473 "cmp" RTX before signed_comparison_operator check to account
4476 2013-08-30 Jan Hubicka <jh@suse.cz>
4478 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
4479 (detect_type_change_1): Rename to ...
4480 (detect_type_change): ... this one; early return on non-polymorphic
4482 (detect_type_change_ssa): Add comp_type parameter; update
4483 use of detect_type_change.
4484 (compute_complex_assign_jump_func): Add param_type parameter;
4485 update use of detect_type_change_ssa.
4486 (compute_complex_ancestor_jump_func): Likewise.
4487 (ipa_get_callee_param_type): New function.
4488 (ipa_compute_jump_functions_for_edge): Compute parameter type;
4489 update calls to the jump function computation functions.
4491 2013-08-30 Teresa Johnson <tejohnson@google.com>
4492 Steven Bosscher <steven@gcc.gnu.org>
4494 * cfgrtl.c (fixup_new_cold_bb): New routine.
4495 (commit_edge_insertions): Invoke fixup_partitions.
4496 (find_partition_fixes): New routine.
4497 (fixup_partitions): Ditto.
4498 (verify_hot_cold_block_grouping): Update comments.
4499 (rtl_verify_edges): Invoke find_partition_fixes.
4500 (rtl_verify_bb_pointers): Update comments.
4501 (rtl_verify_bb_layout): Ditto.
4502 * basic-block.h (probably_never_executed_edge_p): Declare.
4503 (fixup_partitions): Ditto.
4504 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
4505 * bb-reorder.c (sanitize_hot_paths): New function.
4506 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
4508 * predict.c (probably_never_executed_edge_p): New routine.
4509 * cfg.c (check_bb_profile): Add partition insanity warnings.
4511 2013-08-30 Meador Inge <meadori@codesourcery.com>
4513 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4515 2013-08-30 Marek Polacek <polacek@redhat.com>
4517 * Makefile.in (ubsan.o): Add.
4518 (c-family/c-ubsan.o): Add.
4519 (builtins.o): Add ubsan.h dependency.
4520 * ubsan.h: New file.
4521 * ubsan.c: New file.
4522 * common.opt: Add -fsanitize=undefined option.
4523 (flag_sanitize): Add variable.
4524 (fsanitize=): Add option. Add Driver.
4525 (fsanitize=thread): Remove option.
4526 (fsanitize=address): Likewise.
4527 (static-libubsan): New option.
4528 * doc/invoke.texi: Document the new flag and -static-libubsan.
4529 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
4530 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
4531 * builtin-attrs.def (ATTR_COLD): Define.
4532 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4533 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
4534 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
4535 * flag-types.h (sanitize_code): New enum.
4536 * opts.c (common_handle_option): Parse command line arguments
4537 of -fsanitize=. Add -fsanitize=unreachable option.
4538 * varasm.c (get_variable_section): Adjust.
4539 (assemble_noswitch_variable): Likewise.
4540 (assemble_variable): Likewise.
4541 (output_constant_def_contents): Likewise.
4542 (categorize_decl_for_section): Likewise.
4543 (place_block_symbol): Likewise.
4544 (output_object_block): Likewise.
4545 * builtins.def: Likewise.
4546 * toplev.c (compile_file): Likewise.
4547 (process_options): Likewise.
4548 * cppbuiltin.c: Likewise.
4549 * tsan.c (tsan_pass): Likewise.
4550 (tsan_gate): Likewise.
4551 (tsan_gate_O0): Likewise.
4552 * cfgexpand.c (partition_stack_vars): Likewise.
4553 (expand_stack_vars): Likewise.
4554 (defer_stack_allocation): Likewise.
4555 (expand_used_vars): Likewise.
4556 * cfgcleanup.c (old_insns_match_p): Likewise.
4557 * asan.c (asan_finish_file): Likewise.
4558 (asan_instrument): Likewise.
4559 (gate_asan): Likewise.
4560 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4561 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4562 (asan_global_struct): Use pointer_sized_int_node instead
4563 calling build_nonstandard_integer_type.
4564 (initialize_sanitizer_builtins): Likewise.
4565 (asan_finish_file): Likewise.
4566 * gcc.c: Document %{%:function(args):X}.
4567 (static_spec_functions): Add sanitize.
4568 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4569 store funcval != NULL there.
4570 (do_spec_1): Adjust handle_spec_function caller.
4571 (handle_braces): Allow %:function(args) as condition.
4572 (sanitize_spec_function): New function.
4573 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4574 (LIBUBSAN_SPEC): Likewise.
4575 (LIBUBSAN_EARLY_SPEC): Likewise.
4576 (SANITIZER_SPEC): Handle libubsan.
4577 (SANITIZER_EARLY_SPEC): Likewise.
4578 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4579 instead of fsanitize=address.
4580 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4581 instead of fsanitize=address*.
4582 * builtins.c: Include ubsan.h.
4583 (fold_builtin_0): Instrument __builtin_unreachable.
4584 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4585 instead of flag_asan.
4586 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4587 (pointer_sized_int_node): Define.
4588 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4590 2013-08-30 Mike Stump <mikestump@comcast.net>
4592 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4595 2013-08-29 Jan Hubicka <jh@suse.cz>
4597 * cgraph.c (cgraph_function_body_availability): Handle weakref
4599 * passes.def: Remove pass_fixup_cfg.
4600 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4601 track when we need to remove functions.
4602 (gate_ipa_inline): Execute inlining always; add comment why.
4603 (pass_data_ipa_inline): Remove TODO_remove_functions.
4604 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4605 do not produce summaries.
4606 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4607 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4608 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4609 constant pool and vtable.
4611 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4613 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4614 New arm_neon.h's internal macros to specify 64-bit constants.
4615 Avoid using stdint.h's macros.
4617 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4619 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4620 before applying REG_P.
4622 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4624 PR tree-optimization/58277
4625 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4626 after seeing too many stmts with vdef in between dombb and current
4627 bb, invalidate everything.
4629 2013-08-30 Richard Biener <rguenther@suse.de>
4631 * fold-const.c (fold_single_bit_test): Fix overflow test.
4633 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4635 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4636 and which can live in a register, always retrieve the value on entry.
4637 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4638 passed by invisible reference specially.
4639 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4640 (vt_add_function_parameter): Correctly deal with a parameter passed by
4641 invisible reference.
4643 2013-08-30 Jan Hubicka <jh@suse.cz>
4645 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4647 2013-08-30 Richard Biener <rguenther@suse.de>
4649 PR tree-optimization/58228
4650 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4651 allow invariant loads in nested loop vectorization.
4653 2013-08-30 Richard Biener <rguenther@suse.de>
4655 PR tree-optimization/58223
4656 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4657 (has_anti_or_output_dependence): ... this and adjust to also
4658 look for output dependences.
4659 (mark_nodes_having_upstream_mem_writes): Adjust.
4660 (rdg_flag_uses): Likewise.
4662 2013-08-30 Richard Biener <rguenther@suse.de>
4664 PR tree-optimization/58010
4665 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4666 assert that we have a loop-closed PHI.
4668 2013-08-29 Jan Hubicka <jh@suse.cz>
4670 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4671 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4672 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4673 (lto_free_function_in_decl_state_for_node): New function.
4675 2013-08-29 Xinliang David Li <davidxl@google.com>
4677 * loop-unroll.c (report_unroll_peel): Minor message change.
4678 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4679 Emit alignment peeling message with default -fopt-info.
4680 (vect_loop_versioning): Emit loop version info message.
4681 * tree-vectorizer.c (vectorize_loops): Minor message change.
4682 (execute_vect_slp): Ditto.
4684 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4686 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4687 of the clone from the DECL_NAME of the original function.
4689 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4691 * passes.c (register_pass): Add overload.
4692 * tree-pass.h (register_pass): Forward declare it. Add comment.
4694 2013-08-29 Jan Hubicka <jh@suse.cz>
4696 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4697 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4698 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4699 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4700 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4701 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4702 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4703 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4705 2013-08-29 Teresa Johnson <tejohnson@google.com>
4707 * dumpfile.c (dump_loc): Output column number.
4708 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4709 * doc/invoke.texi: Document optall -fopt-info flag.
4710 * profile.c (read_profile_edge_counts): Use new dump framework.
4711 (compute_branch_probabilities): Ditto.
4712 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4713 when pass not in any opt group.
4714 * pass_manager.h (pass_manager::get_pass_profile): New method.
4715 * value-prof.c (check_counter): Use new dump framework.
4716 (check_ic_target): Ditto.
4717 * coverage.c (get_coverage_counts): Ditto.
4718 (coverage_init): Setup new dump framework.
4720 2013-08-29 Richard Biener <rguenther@suse.de>
4722 PR tree-optimization/58246
4723 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4724 handle the dominance check inside a basic-block.
4726 2013-08-29 Richard Biener <rguenther@suse.de>
4729 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4730 of default defs that appear in abnormal PHI nodes.
4732 2013-08-29 Richard Biener <rguenther@suse.de>
4734 PR tree-optimization/57685
4735 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4736 single-use operands to avoid exponential complexity.
4738 2013-08-28 Dehao Chen <dehao@google.com>
4740 * ipa-inline.c (edge_badness): Fix integer underflow.
4742 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4744 * gtm-builtins.def (_ITM_free): Declare leaf.
4746 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4749 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4750 (*tls_local_dynamic_base_64_largepic): Likewise.
4751 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4752 Remove predicate from call operand.
4753 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4754 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4756 2013-08-28 Jeff Law <law@redhat.com>
4758 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4759 checks for the number of predecessors and successors allowed.
4760 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4761 which require copying a joiner block if there is a request which
4762 is a subpath that requires no joiner block copying.
4764 2013-08-28 Jan Hubicka <jh@suse.cz>
4766 * lto-streamer-out.c (DFS_write_tree_body): Drop
4767 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4768 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4769 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4770 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4771 Do not read DECL_ERROR_ISSUED.
4772 (unpack_ts_decl_with_vis_value_fields): Do not read
4774 (lto_input_ts_binfo_tree_pointers): Do not read
4775 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4776 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4777 write DECL_ERROR_ISSUED..
4778 (pack_ts_decl_with_vis_value_fields): Do not write
4780 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4781 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4782 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4783 * tree.h (tree_decl_common): Update comment.
4784 (DECL_ERROR_ISSUED): Remove.
4786 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4789 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4791 2013-08-28 Jan Hubicka <jh@suse.cz>
4793 * builtins.def (free): Declare leaf.
4795 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4798 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4799 * configure: Regenerate.
4801 2013-08-27 Martin Jambor <mjambor@suse.cz>
4803 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4804 (ipa_ancestor_jf_data): Likewise.
4805 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4806 (ipa_get_jf_pass_through_type_preserved): New function.
4807 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4808 (ipa_get_jf_ancestor_type_preserved): New function.
4809 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4810 (ipa_get_jf_ancestor_result): Likewise.
4811 (propagate_vals_accross_pass_through): Use
4812 ipa_get_jf_pass_through_result to do all the value mappings.
4813 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4814 type_preserved flag.
4815 (ipa_set_jf_cst_copy): New function.
4816 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4817 (ipa_set_jf_arith_pass_through): Likewise.
4818 (ipa_set_ancestor_jf): Likewise.
4819 (compute_complex_assign_jump_func): Set type_preserved instead of
4821 (ipa_compute_jump_functions_for_edge): Likewise.
4822 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4823 (update_jump_functions_after_inlining): Update type_preserved.
4824 Explicitely create jump functions when combining one with pass_through.
4825 (ipa_write_jump_function): Stream the type_preserved flags.
4826 (ipa_read_jump_function): Likewise.
4828 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4829 Aldy Hernandez <aldyh@redhat.com>
4831 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4832 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4833 * function.h (struct function): Add has_force_vect_loops and
4835 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4836 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4838 * gimple.def: Update CLAUSES comments.
4839 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4840 (gimple_build_omp_for): Add argument to prototype.
4841 (gimple_omp_for_kind): New.
4842 (gimple_omp_for_set_kind): New.
4843 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4844 GOVD_DATA_SHARE_CLASS.
4845 (enum omp_region_type): Add ORT_SIMD.
4846 (gimple_add_tmp_var): Handle ORT_SIMD.
4847 (gimplify_var_or_parm_decl): Same.
4848 (is_gimple_stmt): Same.
4849 (omp_firstprivatize_variable): Same.
4850 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4851 (omp_notice_variable): Handle ORT_SIMD.
4852 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4853 (omp_check_private): Handle ORT_SIMD.
4854 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4856 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4857 Handle OMP_CLAUSE_LASTPRIVATE.
4858 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4860 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4861 (gimplify_expr): Handle OMP_SIMD.
4862 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4863 (expand_GOMP_SIMD_VF): New.
4864 (expand_GOMP_SIMD_LAST_LANE): New.
4865 * internal-fn.def (GOMP_SIMD_LANE): New.
4866 (GOMP_SIMD_VF): New.
4867 (GOMP_SIMD_LAST_LANE): New.
4868 * omp-low.c: Include target.h.
4869 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4871 (check_omp_nesting_restrictions): Same.
4873 (lower_rec_simd_input_clauses): New.
4874 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4876 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4877 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4878 (expand_omp_build_assign): New.
4879 (expand_omp_for_init_counts): New.
4880 (expand_omp_for_init_vars): New.
4881 (extract_omp_for_update_vars): New.
4882 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4883 and rewrite accordingly.
4884 (expand_omp_simd): New.
4885 (expand_omp_for): Use expand_omp_simd.
4886 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4887 (lower_omp_for): Do not lower the body.
4888 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4890 * tree-flow.h (find_omp_clause): Remove prototype.
4891 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4892 forcing vectorization of the loop, or if flag_tree_vectorize.
4893 (gate_tree_if_conversion): Similarly.
4894 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4895 gimple_build_omp_for.
4896 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4897 * tree-parloops (create_parallel_loop): Pass kind argument to
4898 gimple_build_omp_for.
4899 * tree-pretty-print.c (dump_omp_clause): Add cases for
4900 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4901 OMP_CLAUSE__SIMDUID_.
4902 (dump_generic_node): Handle OMP_SIMD.
4903 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4904 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4905 unroll OMP_SIMD loops here.
4906 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4907 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4909 (vect_analyze_data_refs): Handle simd loops.
4910 * tree-vect-loop.c (vectorizable_live_operation): Handle
4912 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4913 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4914 (vectorizable_load): Same.
4915 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4916 (struct simduid_to_vf): New.
4917 (simduid_to_vf::hash): New.
4918 (simduid_to-vf::equal): New.
4919 (struct simd_array_to_simduid): New.
4920 (simd_array_to_simduid::hash): New.
4921 (simd_array_to_simduid::equal): New.
4922 (adjust_simduid_builtins): New.
4923 (struct note_simd_array_uses_struct): New.
4924 (note_simd_array_uses_cb): New.
4925 (note_simd_array_uses): New.
4926 (vectorize_loops): Handle simd hints and adjust simd builtins
4928 * tree-vectorizer.h (struct _stmt_vec_info): Add
4929 simd_lane_access_p field.
4930 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4931 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4932 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4933 (omp_clause_code_name): Same.
4934 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4935 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4936 * tree.def (OMP_SIMD): New entry.
4937 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4938 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4939 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4940 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4941 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4942 (OMP_CLAUSE_LINEAR_STEP): New.
4943 (OMP_CLAUSE_SAFELEN_EXPR): New.
4944 (OMP_CLAUSE__SIMDUID__DECL): New.
4945 (find_omp_clause): New prototype.
4947 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4949 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4950 Haswell processor detection.
4952 2013-08-27 Christian Widmer <shadow@umbrox.de>
4955 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4956 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4957 AVX2 capable processors.
4959 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4961 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4962 of vget_low_* with implementations in terms of other intrinsics.
4964 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4967 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4968 values to -1, 0 and 1.
4970 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4972 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4973 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4974 (<optab><mode>3_insn): Remove.
4975 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4976 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4977 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4978 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4979 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4980 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4981 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4982 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4983 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4984 (ror<mode>3_insn): Likewise.
4985 * config/aarch64/predicates.md (aarch64_simd_register): New.
4987 2013-08-27 Richard Biener <rguenther@suse.de>
4989 PR tree-optimization/57521
4990 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4991 one edge is non-critical.
4992 (find_phi_replacement_condition): Make sure to use a non-critical
4993 edge. Cleanup and remove old bug workarounds.
4994 (bb_postdominates_preds): Remove.
4995 (if_convertible_loop_p_1): Do not compute post-dominators.
4996 (combine_blocks): Do not free post-dominators.
4997 (main_tree_if_conversion): Likewise.
4998 (pass_data_if_conversion): Add TODO_verify_ssa.
5000 2013-08-27 DJ Delorie <dj@redhat.com>
5002 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
5004 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
5006 * function.c (assign_parm_find_data_types): Set passed_mode and
5007 nominal_mode to the TYPE_MODE of nominal_type for the built
5008 pointer type in case of the struct-pass-by-reference.
5010 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5012 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
5013 (UINT16_TYPE): Change default to "unsigned int".
5015 * config/avr/avr.opt (mfract-convert-truncate): New option.
5016 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5017 is set, round negative fractional integers according to n1169
5018 when converting to integer types.
5020 2013-08-26 Jan Hubicka <jh@suse.cz>
5022 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5023 methods can not be called indirectly when their address is not taken.
5025 2013-08-26 Jan Hubicka <jh@suse.cz>
5027 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5030 2013-08-26 Jan Hubicka <jh@suse.cz>
5032 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5035 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5037 * reload.c (find_valid_class): Allow classes that do not include
5038 FIRST_PSEUDO_REGISTER - 1.
5040 2013-08-26 Jan Hubicka <jh@suse.cz>
5042 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5043 fix edge count/frequency when speculation failed; fix type check
5044 for the direct call.
5046 2013-08-26 Jan Hubicka <jh@suse.cz>
5048 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5050 2013-08-26 Jan Hubicka <jh@suse.cz>
5052 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5053 to be changed by edge redirection.
5055 2013-08-26 Jan Hubicka <jh@suse.cz>
5057 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5058 formating; add sanity check.
5059 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5060 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5061 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5062 (resolve_noninline_speculation): Update callee keys, too.
5064 2013-08-26 Jan Hubicka <jh@suse.cz>
5066 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5067 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5069 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5071 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5074 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5076 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5077 (x86_64_elf_select_section): Ditto.
5078 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5080 (ix86_pass_by_reference): Ditto.
5081 (output_set_got): Ditto.
5082 (ix86_unary_operator_ok): Ditto.
5083 (ix86_expand_builtin): Ditto.
5085 2013-08-23 Jan Hubicka <jh@suse.cz>
5087 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5089 2013-08-23 Jan Hubicka <jh@suse.cz>
5091 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5092 (tree_decl_with_vis): Add FINAL field.
5094 2013-08-23 Jeff Law <law@redhat.com>
5096 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5097 the debugging dump when the expression is fully redundant.
5099 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5101 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5102 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5103 * pretty-print.c (pp_formatted_text_data): Likewise.
5104 (pp_write_text_to_stream): Likewise.
5105 (pp_write_text_as_dot_label_to_stream): Likewise.
5106 (pp_append_r): Likewise.
5107 (pp_format): Likewise.
5108 (pp_flush): Likewise.
5109 (pp_clear_output_area): Likewise.
5110 (pp_append_text): Likewise.
5111 (pp_formatted_text): Likewise.
5112 (pp_remaining_character_count_for_line): Likewise.
5113 (pp_newline): Likewise.
5114 (pp_character): Likewise.
5115 (output_buffer::~output_buffer): Define.
5116 (pretty_printer::~pretty_printer): Destruct output buffer.
5117 * pretty-print.h (output_buffer::~output_buffer): Declare.
5118 (pretty_printer::~pretty_printer): Declare virtual.
5120 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5123 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
5124 HOST_WIDE_INT_M1U): New macros.
5125 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
5126 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
5127 unsigned -1 for lshift.
5128 * cse.c (cse_insn): Likewise.
5129 * double-int.c (rshift_double, lshift_double): Likewise.
5130 * builtins.c (fold_builtin_bitop): Likewise.
5131 * combine.c (force_to_mode): Likewise.
5132 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
5133 * simplify-rtx.c (simplify_const_unary_operation,
5134 simplify_const_binary_operation): Likewise.
5135 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
5136 check_va_list_escapes): Likewise.
5137 * rtlanal.c (nonzero_bits1): Likewise.
5138 * expmed.c (expand_smod_pow2): Likewise.
5139 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
5141 2013-08-23 Jan Hubicka <jh@suse.cz>
5143 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
5144 as having address taken.
5146 2013-08-23 Jan Hubicka <jh@suse.cz>
5148 * ipa-utils.h (method_class_type): Declare.
5149 * ipa-devirt.c (method_class_type): Export.
5151 * cgraphunit.c (analyze_functions): Do basic devirtualization;
5152 do not walk base classes of anonymous types.
5154 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
5156 PR rtl-optimization/58220
5158 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
5159 handle SEQUENCE insns properly.
5161 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5163 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
5165 (pp_newline_and_indent): Likewise.
5166 (pp_separate_with): Likewise.
5167 * pretty-print.c (pp_newline_and_flush): Define.
5168 (pp_newline_and_indent): Likewise.
5169 (pp_separate_with): Likewise.
5171 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5174 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
5175 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
5177 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
5179 * config/i386/predicates.md (ext_sse_reg_operand): New.
5180 * config/i386/i386.md (*movti_internal): Use
5181 predicate to determine if EVEX is needed.
5182 (*movsi_internal): Ditto.
5183 (*movdf_internal): Ditto.
5184 (*movsf_internal): Ditto.
5185 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
5187 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5189 PR tree-optimization/58209
5190 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
5191 (find_tail_calls): Give up for pointer result types if m is non-NULL.
5192 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
5193 emit POINTER_PLUS_EXPR.
5194 (create_tailcall_accumulator): For pointer result type accumulate in
5197 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
5199 * configure.ac: Add backslashes missing from the last change.
5200 * configure: Regenerate.
5202 2013-08-22 Jan Hubicka <jh@susue.cz>
5204 * ipa.c (function_and_variable_visibility): First remember function
5205 was global and then make it local.
5207 2013-08-22 Julian Brown <julian@codesourcery.com>
5209 * configure.ac: Add aarch64 to list of arches which use "nop" in
5211 * configure: Regenerate.
5213 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5215 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
5216 gnu_libc_has_function.
5217 * config/s390/tpf.h: Likewise.
5219 2013-08-22 Jan Hubicka <jh@susue.cz>
5221 * timevar.c (validate_phases): Add cast.
5223 2013-08-22 Jan Hubicka <jh@susue.cz>
5225 * timevar.c (validate_phases): Use size_t for memory.
5226 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
5228 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
5230 * pretty-print.h (output_buffer::output_buffer): Declare.
5231 (pretty_printer::pretty_printer): Likewise.
5232 (pp_construct): Remove.
5233 * pretty-print.c (output_buffer::output_buffer): Define.
5234 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
5235 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
5236 (print_gimple_expr): Likewise.
5237 (print_gimple_seq): Likewise.
5238 (gimple_dump_bb): Likewise.
5239 * sched-vis.c (dump_value_slim): Likewise.
5240 (dump_insn_slim): Likewise.
5241 (dump_rtl_slim): Likewise.
5242 (str_pattern_slim): Likewise.
5243 * tree-mudflap.c (mf_varname_tree): Likewise.
5244 * graph.c (print_graph_cfg): Likewise.
5245 (start_graph_dump): Likewise.
5246 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
5248 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
5249 pretty printer initialization.
5250 * coretypes.h (diagnostic_context): Remove superflous type alias
5252 (pretty_printer): Likewise. Declare directly as a class.
5253 (pretty_print_info): Remove declaration as class.
5254 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5255 and pp_clear_output_area.
5256 (asan_add_global): Likewise.
5258 2013-08-22 Jan Hubicka <jh@suse.cz>
5260 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5261 * ipa-utils.h (update_type_inheritance_graph): Declare.
5262 (possible_polymorphic_call_target_p): Declare.
5263 (possible_polymorphic_call_target_p): New.
5264 * ipa-devirt.c: Update toplevel comments.
5265 (cached_polymorphic_call_targets): Move up.
5266 (odr_type_d): Move ID down.
5267 (polymorphic_type_binfo_p): Update comment.
5268 (odr_hasher::remove): Likewise;
5269 (get_odr_type): Set anonymous_namespace.
5270 (dump_odr_type): Dump it.
5271 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
5272 (maybe_record_node): Record node in cached_polymorphic_call_targets.
5273 (record_binfo): Add comment.
5274 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
5276 (devirt_node_removal_hook): Do not iCE when cache is freed.
5277 (possible_polymorphic_call_target_p): New predicate.
5278 (update_type_inheritance_graph): New function.
5280 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
5281 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5282 Sergey Lega <sergey.s.lega@intel.com>
5283 Anna Tikhonova <anna.tikhonova@intel.com>
5284 Ilya Tocar <ilya.tocar@intel.com>
5285 Andrey Turetskiy <andrey.turetskiy@intel.com>
5286 Ilya Verbin <ilya.verbin@intel.com>
5287 Kirill Yukhin <kirill.yukhin@intel.com>
5288 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5290 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
5291 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
5292 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
5293 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
5294 (OPTION_MASK_ISA_AVX2_UNSET): Update.
5295 (OPTION_MASK_ISA_AVX512F_UNSET): New.
5296 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
5297 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
5298 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
5299 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
5300 OPT_mavx512pf, OPT_mavx512er cases.
5301 * config/i386/constraints.md (v): New constraint.
5302 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
5303 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
5304 (bit_AVX512CD): New.
5305 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5306 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
5307 * config/i386/i386-c.c (ix86_target_macros_internal):
5308 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
5310 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
5311 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
5312 * config/i386/i386.c (regclass_map, dbx_register_map)
5313 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
5314 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
5315 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
5317 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
5318 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
5319 -mavx512pf options. Fix formatting.
5320 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
5321 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
5322 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
5323 -mavx512cd, -mavx512pf options.
5324 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
5325 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
5326 (ix86_preferred_output_reload_class): Replace SSE_REGS with
5328 (ix86_hard_regno_mode_ok): Support 512-bit registers.
5329 (ix86_set_reg_reg_cost): Ditto.
5330 (x86_order_regs_for_local_alloc): Ditto.
5331 (MAX_VECT_LEN): Extend to 64-byte.
5332 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
5333 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
5334 (TARGET_AVX512ER, TARGET_AVX512CD): New.
5335 (BIGGEST_ALIGNMENT): Extend to 512-bits.
5336 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
5337 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
5338 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
5339 (SSE_REG_MODE_P): Support new modes.
5340 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
5341 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
5342 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
5343 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
5344 (REG_CLASS_CONTENTS): Add new registers.
5345 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
5346 (EXT_REX_SSE_REGNO_P): New.
5347 (HI_REGISTER_NAMES): Add new registers.
5348 * config/i386/i386.md: Define constants for new registers.
5349 (mode): Add new 512-bit modes.
5350 (prefix): Support evex prefix.
5351 (isa): Support avx512f, noavx512f, fma_avx512f.
5352 (ssemodesuffix): Add new 512-bit modes.
5354 (*movxi_internal_avx512f): Ditto.
5355 (*movdi_internal): Replace constraint "x" with the new constraint "v".
5357 (*movsi_internal): Likewise.
5358 (*movdf_internal): Likewise.
5359 (*movsf_internal): Likewise.
5360 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
5361 (<code><mode>3): Likewise.
5362 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
5364 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
5365 with the new constraint "v".
5366 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
5368 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
5369 with the new constraint "v".
5370 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
5371 (<sse2>_storedqu<avxsizesuffix>): Likewise.
5372 (*<plusminus_insn><mode>3): Likewise.
5373 (<sse>_vm<plusminus_insn><mode>3): Likewise.
5374 (*mul<mode>3): Likewise.
5375 (<sse>_vmmul<mode>3): Likewise.
5376 (<sse>_div<mode>3): Likewise.
5377 (<sse>_vmdiv<mode>3): Likewise.
5378 (<sse>_sqrt<mode>2): Likewise.
5379 (<sse>_vmsqrt<mode>2): Likewise.
5380 (*<code><mode>3_finite): Likewise.
5381 (*<code><mode>3) <smaxmin>: Likewise.
5382 (<sse>_vm<code><mode>3): Likewise.
5383 (*<code><mode>3) <any_logic>: Likewise.
5384 (*fma_fmadd_<mode>): Likewise.
5385 (*fma_fmsub_<mode>): Likewise.
5386 (*fma_fnmadd_<mode>): Likewise.
5387 (*fma_fnmsub_<mode>): Likewise.
5388 (*fma_fmaddsub_<mode>): Likewise.
5389 (*fma_fmsubadd_<mode>): Likewise.
5390 (*fmai_fmadd_<mode>): Likewise.
5391 (*fmai_fmsub_<mode>): Likewise.
5392 (*fmai_fnmadd_<mode>): Likewise.
5393 (*fmai_fnmsub_<mode>): Likewise.
5394 (sse_cvtsi2ss): Likewise.
5395 (sse_cvtsi2ssq): Likewise.
5396 (sse_cvtss2si): Likewise.
5397 (sse_cvtss2si_2): Likewise.
5398 (sse_cvtss2siq): Likewise.
5399 (sse_cvtss2siq_2): Likewise.
5400 (sse_cvttss2si): Likewise.
5401 (sse_cvtss2siq_2): Likewise.
5402 (float<sseintvecmodelower><mode>2): Likewise.
5403 (sse2_cvtsd2si_2): Likewise.
5404 (sse2_cvtsd2siq_2): Likewise.
5405 (*<plusminus_insn><mode>3): Likewise.
5406 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
5407 (*<sse4_1_avx2>_mul<mode>3): Likewise.
5408 (ashr<mode>3): Likewise.
5409 (<shift_insn><mode>3): Likewise.
5410 (avx2_<code><mode>3): Likewise.
5411 (*avx2_<code><mode>3): Likewise.
5412 (*andnot<mode>3): Likewise.
5413 (*<code><mode>3) <any_logic>: Likewise.
5414 (abs<mode>2): Likewise.
5415 (avx2_permvar<mode>): Likewise.
5416 (avx2_perm<mode>_1): Likewise.
5417 (*avx_vpermilp<mode>): Likewise.
5418 (avx_vpermilvar<mode>3): Likewise.
5419 (avx2_ashrv<mode>): Likewise.
5420 (avx2_<shift_insn>v<mode>): Likewise.
5421 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
5423 * doc/rtl.texi: Document XImode.
5425 2013-08-21 Jeff Law <law@redhat.com>
5427 * tree-flow.h (register_jump_thread): Pass vector of edges
5428 instead of each important edge.
5429 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
5430 thread path into a vector and pass that to register_jump_thread.
5431 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
5432 passed in edge vector to the current 3-edge form.
5435 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5437 * dce.c (fini_dce): Call df_analyze again just in case
5438 delete_unmarked_insns removed anything.
5440 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
5442 * reload.h (struct reg_equivs): Rename to ..
5443 (struct reg_equivs_s): .. this.
5445 2013-08-20 Martin Liska <marxin.liska@gmail.com>
5447 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
5449 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5451 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
5453 2013-08-21 Jeff Law <law@redhat.com>
5455 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
5456 simplify assignments too. If the RHS collapses to a singleton
5457 range, then return the value for the range.
5459 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
5461 * config/i386/sse.md (V16): Rename to...
5463 (mov<mode>): Update iterator name.
5464 (*mov<mode>_internal): Ditto.
5465 (push<mode>1): Ditto.
5466 (movmisalign<mode>): Ditto.
5468 2013-08-20 Jan Hubicka <jh@suse.cz>
5471 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
5472 entry for direct edges.
5473 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
5475 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5477 Revert my last two changes, r201865 and r201864:
5480 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5482 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5483 instances can own GC refs.
5485 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5486 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5487 (gcc::context::gt_pch_nx): Likewise.
5488 (gcc::context::gt_pch_nx): Likewise.
5489 * ggc.h (gt_ggc_mx <T>): New.
5490 (gt_pch_nx_with_op <T>): New.
5491 (gt_pch_nx <T>): New.
5492 * passes.c (opt_pass::gt_ggc_mx): New.
5493 (opt_pass::gt_pch_nx): New.
5494 (opt_pass::gt_pch_nx_with_op): New.
5495 (pass_manager::gt_ggc_mx): New.
5496 (pass_manager::gt_pch_nx): New.
5497 (pass_manager::gt_pch_nx_with_op): New.
5498 (pass_manager::operator new): Use
5499 ggc_internal_cleared_alloc_stat rather than xcalloc.
5500 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5501 (pass_manager::gt_ggc_mx): New.
5502 (pass_manager::gt_pch_nx): New.
5503 (pass_manager::gt_pch_nx_with_op): New.
5504 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5505 (opt_pass::operator new): New.
5506 (opt_pass::gt_ggc_mx): New.
5507 (opt_pass::gt_pch_nx): New.
5508 (opt_pass::gt_pch_nx_with_op): New.
5511 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5513 * Makefile.in (GTFILES): Add context.h.
5514 * context.c (gcc::context::operator new): New.
5515 (gcc::context::gt_ggc_mx): New.
5516 (gcc::context::gt_pch_nx): New.
5517 (gcc::context::gt_pch_nx): New.
5518 * context.h (gcc::context): Add GTY((user)) marking.
5519 (gcc::context::operator new): New.
5520 (gcc::context::gt_ggc_mx): New.
5521 (gcc::context::gt_pch_nx): New.
5522 (gcc::context::gt_pch_nx): New.
5523 (g): Add GTY marking.
5524 (gt_ggc_mx (gcc::context *)): New.
5525 (gt_pch_nx (gcc::context *)): New.
5526 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5527 void *cookie)): New.
5528 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5530 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5532 * dce.c (fini_dce): Call df_analyze again just in case
5533 delete_unmarked_insns removed anything.
5535 2013-08-20 Teresa Johnson <tejohnson@google.com>
5537 PR rtl-optimizations/57451
5538 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5539 from crossing split section boundaries.
5541 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5543 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5544 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5545 with MULTLIB_DEFAULTS.
5547 2013-08-20 Nick Clifton <nickc@redhat.com>
5549 * target.def (narrow_volatile_bitfield): Note that the default
5550 value is false, not !TARGET_STRICT_ALIGN.
5551 * doc/tm.texi: Regenerate.
5553 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5555 Fix LIB_SPEC for systems without libpthread.
5557 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5558 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5560 * config/i386/linux-common.h: Likewise.
5561 * config/mips/linux-common.h: Likewise.
5563 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5565 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5568 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5570 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5571 instances can own GC refs.
5573 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5574 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5575 (gcc::context::gt_pch_nx): Likewise.
5576 (gcc::context::gt_pch_nx): Likewise.
5577 * ggc.h (gt_ggc_mx <T>): New.
5578 (gt_pch_nx_with_op <T>): New.
5579 (gt_pch_nx <T>): New.
5580 * passes.c (opt_pass::gt_ggc_mx): New.
5581 (opt_pass::gt_pch_nx): New.
5582 (opt_pass::gt_pch_nx_with_op): New.
5583 (pass_manager::gt_ggc_mx): New.
5584 (pass_manager::gt_pch_nx): New.
5585 (pass_manager::gt_pch_nx_with_op): New.
5586 (pass_manager::operator new): Use
5587 ggc_internal_cleared_alloc_stat rather than xcalloc.
5588 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5589 (pass_manager::gt_ggc_mx): New.
5590 (pass_manager::gt_pch_nx): New.
5591 (pass_manager::gt_pch_nx_with_op): New.
5592 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5593 (opt_pass::operator new): New.
5594 (opt_pass::gt_ggc_mx): New.
5595 (opt_pass::gt_pch_nx): New.
5596 (opt_pass::gt_pch_nx_with_op): New.
5598 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5600 * Makefile.in (GTFILES): Add context.h.
5601 * context.c (gcc::context::operator new): New.
5602 (gcc::context::gt_ggc_mx): New.
5603 (gcc::context::gt_pch_nx): New.
5604 (gcc::context::gt_pch_nx): New.
5605 * context.h (gcc::context): Add GTY((user)) marking.
5606 (gcc::context::operator new): New.
5607 (gcc::context::gt_ggc_mx): New.
5608 (gcc::context::gt_pch_nx): New.
5609 (gcc::context::gt_pch_nx): New.
5610 (g): Add GTY marking.
5611 (gt_ggc_mx (gcc::context *)): New.
5612 (gt_pch_nx (gcc::context *)): New.
5613 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5614 void *cookie)): New.
5615 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5617 2013-08-20 Alan Modra <amodra@gmail.com>
5620 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5621 (rs6000_emit_epilogue): Likewise.
5623 2013-08-19 Dehao Chen <dehao@google.com>
5625 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5627 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5628 Jakub Jelinek <jakub@redhat.com>
5630 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5631 (BUILT_IN_FABSD64): Likewise.
5632 (BUILT_IN_FABSD128): Likewise.
5633 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5634 (fold_builtin_1): Likewise.
5635 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5636 destination and source operands.
5637 (*abstd2_fpr): Likewise.
5638 (*nabstd2_fpr): Likewise.
5640 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5642 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5645 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5647 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5650 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5652 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5653 * builtins.c (default_libc_has_function): New.
5654 (gnu_libc_has_function): Ditto.
5655 (no_c99_libc_has_function): Ditto.
5656 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5657 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5658 (fold_builtin_sincos): Likewise.
5659 (fold_builtin_cexp): Likewise.
5660 * builtins.def (DEF_C94_BUILTIN): Likewise.
5661 (DEF_C99_BUILTIN): Likewise.
5662 (DEF_C99_C90RES_BUILTIN): Likewise.
5663 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5664 definitions to using this define.
5665 * config/darwin-protos.h (darwin_libc_has_function): New.
5666 * config/darwin.c (darwin_libc_has_function): Ditto.
5667 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5668 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5669 * config/darwin.h: Ditto.
5670 * config/elfos.h: Ditto.
5671 * config/freebsd.h: Ditto.
5672 * config/i386/cygming.h: Ditto.
5673 * config/i386/djgpp.h: Ditto.
5674 * config/i386/i386-interix.h: Ditto.
5675 * config/microblaze/microblaze.h: Ditto.
5676 * config/mmix/mmix.h: Ditto.
5677 * config/gnu-user.h: Ditto.
5678 * config/ia64/hpux.h: Ditto.
5679 * config/pa/pa-hpux.h: Ditto.
5680 * config/pdp11/pdp11.h: Ditto.
5681 * config/picochip/picochip.h: Ditto.
5682 * config/linux.h: Ditto.
5683 * config/netbsd.h: Ditto.
5684 * config/openbsd.h: Ditto.
5685 * config/rs6000/aix43.h: Ditto.
5686 * config/rs6000/aix51.h: Ditto.
5687 * config/rs6000/aix52.h: Ditto.
5688 * config/rs6000/aix53.h: Ditto.
5689 * config/rs6000/aix61.h: Ditto.
5690 * config/rs6000/darwin.h: Ditto.
5691 * config/rs6000/linux.h: Ditto.
5692 * config/rs6000/linux64.h: Ditto.
5693 * config/s390/tpf.h: Ditto.
5694 * config/sol2-10.h: Ditto.
5695 * config/sol2.h: Ditto.
5696 * config/vms/vms.h: Ditto.
5697 * config/vxworks.h: Ditto.
5698 * config/linux-android.c (linux_android_libc_has_function):
5699 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5700 * config/linux-protos.h (linux_android_libc_has_function):
5702 * config/i386/i386.c (ix86_libc_has_function): New.
5703 * config/i386/i386-protos.h
5704 (ix86_libc_has_function): New declaration.
5705 * config/i386/i386.md
5706 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5707 ("isinf<mode>2): Likewise.
5708 * convert.c (convert_to_integer): Using new target hook
5709 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5710 TARGET_C99_FUNCTIONS.
5711 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5712 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5713 * coretypes.h (function_class): New enum for different
5714 classes of functions.
5715 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5716 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5717 (TARGET_HAS_SINCOS): Likewise.
5718 (TARGET_LIBC_HAS_FUNCTION): New.
5719 * doc/tm.texi: Regenerated.
5720 * targhooks.h (default_libc_has_function): New declaration.
5721 (no_c99_libc_has_function): Ditto.
5722 (gnu_libc_has_function): Ditto.
5723 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5724 and TARGET_HAS_SINCOS.
5726 2013-08-18 Jan Hubicka <jh@suse.cz>
5728 * Makeifle-in (ipa-devirt.o): New.
5729 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5730 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5731 (analyze_functions): Look into possible targets of polymorphic call.
5732 * dumpfile.c (dump_files): Add type-inheritance dump.
5733 * dumpfile.h (TDI_inheritance): New.
5734 * ipa-devirt.c: New file.
5735 * ipa-utils.h (odr_type_d): Forward declare.
5736 (odr_type): New type.
5737 (build_type_inheritance_graph): Declare.
5738 (possible_polymorphic_call_targets): Declare and introduce inline
5739 variant when only edge is pased.
5740 (dump_possible_polymorphic_call_targets): Likewise.
5741 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5742 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5743 (types_same_for_odr): ... here.
5744 * tree.h (type_in_anonymous_namespace_p): Declare.
5746 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5748 PR tree-optimization/58006
5749 * tree-parloops.c (take_address_of): Don't ICE if get_name
5751 (eliminate_local_variables_stmt): Remove clobber stmts.
5753 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5755 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5756 error message is issued for an alias to undefined symbol.
5758 2013-08-18 Jan Hubicka <jh@suse.cz>
5760 * cgraph.c (cgraph_create_indirect_edge): Discover
5761 polymorphic calls and record basic info into indirect_info.
5762 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5763 devirtualization, ignore objc function calls.
5764 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5765 call with no parm index info.
5766 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5767 * tree.c (virtual_method_call_p): New function.
5768 * tree.h (virtual_method_call_p): Declare.
5770 2013-08-16 Jan Hubicka <jh@suse.cz>
5773 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5775 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5777 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5780 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5782 * gengtype.c (type_for_name): Add special-case support for
5783 locating types within the "gcc::" namespace.
5784 (open_base_files): Emit a "using namespace gcc" directive.
5786 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5789 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5790 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5792 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5793 array instead of each individual operand as a separate argument.
5794 (emit_fusion_gpr_load): Likewise.
5795 (expand_fusion_gpr_load): Add new function declaration.
5797 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5798 signature to have the operands passed as an array, instead of as
5799 separate arguments. Allow ZERO_EXTEND to be in the memory
5800 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5801 depend on the register live/dead flags when peepholes are run.
5802 (expand_fusion_gpr_load): New function to be called from the
5803 peephole2 pass, to change the register that addis sets to be the
5805 (emit_fusion_gpr_load): Change the calling signature to have the
5806 operands passed as an array, instead of as separate arguments.
5807 Allow ZERO_EXTEND to be in the memory address, and also
5808 SIGN_EXTEND if -mpower8-fusion-sign.
5810 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5812 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5813 adjust the register addis loads up in the peephole2 pass. Do not
5814 depend on the register live/dead state when the peephole pass is done.
5816 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5818 * gengtype.c (create_user_defined_type): Ensure that the kind
5819 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5820 declaration is seen before the GTY((user)) marking.
5822 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5825 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5827 2013-08-16 Jan Hubicka <jh@suse.cz>
5829 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5830 arugment expected_type.
5831 (gimple_fold_call): Use it.
5832 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5833 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5834 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5835 (try_make_edge_direct_virtual_call): Likewise.
5836 * tree.c (obj_type_ref_class): New.
5837 * tree.h (obj_type_ref_class): Use it.
5839 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5841 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5842 (rtl_slim_pp): Likewise.
5843 (init_rtl_slim_pretty_print): Likewise.
5844 (dump_value_slim): Don't call it. Use local pretty printer.
5845 (dump_insn_slim): Likewise.
5846 (dump_rtl_slim): Likewise.
5847 (str_pattern_slim): Likewise.
5848 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5851 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5853 PR tree-optimization/58164
5854 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5855 walk gimple_goto_dest of GIMPLE_GOTO.
5857 PR tree-optimization/58165
5858 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5859 bi_call must be the last stmt in a bb, don't split_block, instead
5860 use fallthru edge from it and give up if there is none.
5861 Release conds vector when returning early.
5863 2013-08-14 Xinliang David Li <davidxl@google.com>
5865 * config/i386/i386.c (ix86_option_override_internal):
5866 Remove unused variable and field.
5868 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5871 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5872 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5873 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5874 and Linux, correct BLKmode alignment when 128-bit alignment is
5875 required and compatibility flag is not set.
5876 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5877 for zero-size arguments when compatibility flag is not set.
5879 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5881 PR tree-optimization/58145
5882 * tree-sra.c (build_ref_for_offset): If prev_base has
5883 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5885 2013-08-14 Xinliang David Li <davidxl@google.com>
5887 * config/i386/i386.c (ix86_option_override_internal):
5888 Fix uninitialized variable error.
5890 2013-08-14 Xinliang David Li <davidxl@google.com>
5892 * config/i386/i386.opt: Define two new options.
5893 * config/i386/x86-tune.def: Add arch selector field in macros.
5894 * config/i386/i386.h: Adjust macro definition.
5895 * config/i386/i386.c (ix86_option_override_internal):
5897 (parse_mtune_ctrl_str): New function.
5898 (set_ix86_tune_features): New function.
5899 (ix86_function_specific_restore): Call the new helper function.
5901 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5903 PR rtl-optimization/57662
5904 * sel-sched.c (code_motion_process_successors): When the current insn
5905 is removed after the recursive traversal, break from the loop.
5906 Add comments and debug printouts.
5908 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5909 Alexandre Oliva <aoliva@redhat.com>
5912 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5913 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5914 there also UNSPEC_PLTOFF.
5916 2013-08-14 Marek Polacek <polacek@redhat.com>
5918 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5919 NUM_CONDITIONS bit positions.
5921 2013-08-13 Cary Coutant <ccoutant@google.com>
5923 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5924 (attr_checksum): Hash vector contents instead of pointer.
5925 (attr_checksum_ordered): Likewise.
5927 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5929 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5930 when Pmode != word_mode. Add length_address attribute.
5931 (sse3_monitor_<mode>): Merge from sse3_monitor and
5932 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5933 Pmode != word_mode. Update insn length attribute.
5934 * config/i386/i386.c (ix86_option_override_internal): Update
5935 ix86_gen_monitor selection for merged sse3_monitor insn.
5937 2013-08-13 Julian Brown <julian@codesourcery.com>
5939 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5940 perform invalid legitimization on greater-than-word-size modes for
5943 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5945 * ira.c (setup_class_translate_array): Use aclass instead of cl
5946 for classes not fully covered by allocno classes.
5948 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5950 PR tree-optimization/57661
5951 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5952 * tree-inline.c (tree_function_versioning): Initialize it.
5953 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5954 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5955 that is not being copied.
5958 * asan.c (instrument_strlen_call): Fix typo in comment.
5959 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5961 2013-08-13 Steve Ellcey <sellcey@mips.com>
5963 * config/mips/mips.md (prefetch): Use lw instead of ld on
5964 loongson in 32bit mode.
5966 2013-08-13 Nick Clifton <nickc@redhat.com>
5968 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5970 2013-08-13 Jan Hubicka <jh@suse.cz>
5972 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5973 introduced edge; fix typo in sanity check.
5974 (cgraph_resolve_speculation): Export; improve diagnostic.
5975 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5976 speculation at type mismatch.
5977 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5978 (cgraph_resolve_speculation): Declare.
5979 (symtab_can_be_discarded): New function.
5980 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5981 * ipa-inline-transform.c (speculation_removed): New global var.
5982 (clone_inlined_nodes): See if speculation can be removed.
5983 (inline_call): If speculations was removed, we growths may not match.
5984 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5985 (speculation_useful_p): New function.
5986 (resolve_noninline_speculation): New function.
5987 (inline_small_functions): Resolve useless speculations.
5988 * ipa-inline.h (speculation_useful_p): Declare
5989 * ipa.c (can_replace_by_local_alias): Simplify.
5990 (ipa_profile): Produce speculative calls in non-lto, too;
5991 add simple cost model; produce local aliases.
5993 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5995 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5998 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
6000 * config/i386/i386.c (ix86_function_versions): Use error + inform.
6002 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
6004 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
6005 iterator instead of X87MODEF.
6007 2013-08-12 Perez Read <netfirewall@gmail.com>
6010 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
6011 operand 0 for intel asm alternative.
6012 (*movabs<mode>_2): Ditto for operand 1.
6014 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6016 * config/aarch64/arm_none.h
6017 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6019 2013-08-12 Nick Clifton <nickc@redhat.com>
6021 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6023 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6025 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6026 expand for QI/HImode operand to produce more effictive code for
6027 unsigned char(short) --> float(double) conversion.
6029 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6031 * doc/invoke.texi: Mention that -ftls-model does not force the final
6034 2013-08-12 Marek Polacek <polacek@redhat.com>
6035 Marc Glisse <marc.glisse@inria.fr>
6037 PR tree-optimization/57980
6038 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6039 when creating -1 constant.
6041 2013-08-10 Jan Hubicka <jh@suse.cz>
6043 Workaround binutils PR14342.
6044 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6045 (gimple_init_edge_profiler): Likewise.
6046 (gimple_gen_ic_func_profiler): Likewise.
6048 2013-08-09 Jan Hubicka <jh@suse.cz>
6050 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6052 2013-08-09 Xinliang David Li <davidxl@google.com>
6054 * config/i386/stringop.def: New file.
6055 * config/i386/stringop.opt: New file.
6056 * config/i386/i386-opts.h: Include stringopt.def.
6057 * config/i386/i386.opt: Include stringopt.opt.
6058 * config/i386/i386.c (ix86_option_override_internal):
6059 Override default size based stringop inline strategies with options.
6060 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6063 2013-08-09 Jan Hubicka <jh@suse.cz>
6065 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6067 2013-08-09 Jan Hubicka <jh@suse.cz>
6069 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6071 (dump_cgraph_node): Dump profile-id.
6072 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6073 and common_target_probability.
6074 * lto-cgraph.c (lto_output_edge): Stream common targets.
6075 (lto_output_node): Stream profile ids.
6076 (input_node): Stream profile ids.
6077 (input_edge): Stream common targets.
6078 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6079 * ipa.c: Include value-prof.h
6080 (ipa_profile_generate_summary): Turn indirect call statement histograms
6081 into common targets.
6082 (ipa_profile): Turn common targets into speculative edges.
6084 2013-08-09 Jan Hubicka <jh@suse.cz>
6086 * cgraph.h (cgraph_node): Add profile_id.
6087 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6088 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6089 (del_node_map): Update.
6090 (find_func_by_funcdef_no): Replace by ...
6091 (find_func_by_profile_id): ... this one.
6092 (gimple_ic_transform): Do not remove useful histograms when
6093 speculation is not done; dump info when indirect call removal
6095 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6096 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6097 (__gcov_indirect_call_profiler_v2): .. this one.
6098 * profile.h (init_node_map): Update.
6099 * coverage.c (coverage_compute_profile_id): New function.
6100 * coverage.h (coverage_compute_profile_id): Declare.
6101 * tree-profile.c (init_ic_make_global_vars): Make
6102 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6103 (gimple_init_edge_profiler): Update prototype of
6104 __gcov_indirect_call_profiler.
6105 (gimple_gen_ic_func_profiler): Simplify.
6106 (tree_profiling): Use init_node_map
6108 2013-08-09 Jan Hubicka <jh@suse.cz>
6110 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6111 non-speculative refs.
6112 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6113 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6114 (cgraph_set_call_stmt): Likewise.
6115 (cgraph_create_edge_1): Fix release checking compilatoin;
6117 (cgraph_free_edge): Free indirect info.
6118 (cgraph_turn_edge_to_speculative): New function.
6119 (cgraph_speculative_call_info): New function.
6120 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6121 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6122 (dump_cgraph_node): Dump speculation.
6123 (verify_edge_count_and_frequency): Accept speculative edges.
6124 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
6125 (verify_cgraph_node): Handle speculation.
6126 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
6127 (cgraph_set_call_stmt): Update prototype.
6128 (cgraph_make_edge_direct): Update prototype.
6129 (cgraph_speculative_call_info): Declare.
6130 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
6131 to change; update call of ipa_find_references.
6132 * ipa-ref.c (ipa_record_reference): Fix return value; clear
6133 lto_stmt_uid and speculative flags.
6134 (ipa_dump_references): Dump speculation.
6135 (ipa_clone_references): Clone speculative flag.
6136 (ipa_clone_referring): Likewise.
6137 (ipa_clone_ref): New function.
6138 (ipa_find_reference): Look into lto_stmt_uids
6139 (ipa_clear_stmts_in_references): Do not clear speculative calls.
6140 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
6141 (ipa_find_reference): Update declaration.
6142 (ipa_clone_ref): Declare.
6143 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
6144 stream speculative flag.
6145 (lto_output_ref): Stream statements uids and speculation.
6146 (input_ref): Likewise.
6147 (input_edge): Stream speuclation.
6148 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
6149 (cgraph_set_call_stmt_including_clones): Handle speculation.
6150 * ipa-inline.c (heap_edge_removal_hook): New function.
6151 (inline_small_functions): Register it.
6152 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
6153 also initialize refs.
6154 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
6156 (try_make_edge_direct_simple_call): Likewise.
6157 (try_make_edge_direct_simple_call): Likewise.
6158 (update_indirect_edges_after_inlining): Likewise.
6159 (remove_described_reference): Look proper lto_stmt_uid.
6160 (propagate_controlled_uses): Likewise.
6161 (propagate_controlled_uses): Liekwise.
6162 * tree-inline.c (copy_bb): Copy speculative edges.
6163 (redirect_all_calls): New function.
6164 (copy_cfg_body): Do redirection after loop info is updated.
6165 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
6167 2013-08-09 Jan Hubicka <jh@suse.cz>
6169 * lto-streamer-out.c (output_function): Renumber PHIs.
6170 * lto-streamer-in.c (input_function): Likewise.
6172 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
6174 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
6175 (get_lane_unsigned): Likewise.
6176 (dup_lane_scalar): Likewise.
6177 (get_lane): enable for VALL.
6178 * config/aarch64/aarch64-simd.md
6179 (aarch64_dup_lane_scalar<mode>): Remove.
6180 (aarch64_get_lane_signed<mode>): Likewise.
6181 (aarch64_get_lane_unsigned<mode>): Likewise.
6182 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
6183 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
6184 (aarch64_get_lane<mode>): Enable for all vector modes.
6185 (aarch64_get_lanedi): Remove misleading constraints.
6186 * config/aarch64/arm_neon.h
6187 (__aarch64_vget_lane_any): Define.
6188 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
6189 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
6190 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
6191 * config/aarch64/iterators.md (VDQQH): New.
6195 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
6197 * configure.ac: Add GAS check for LEON instructions on SPARC.
6198 * configure: Regenerate.
6199 * config.in: Likewise.
6200 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
6202 * config/sparc/sparc.opt (LEON, LEON3): New masks.
6203 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
6205 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
6206 (AS_LEON_FLAG): New macro.
6207 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
6208 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
6209 Deal with LEON and LEON3 for the memory model.
6210 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
6211 (atomic_compare_and_swap<mode>_1): Likewise.
6212 (*atomic_compare_and_swap<mode>_1): Likewise.
6214 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6216 * config/arm/neon.md (vcond): Fix floating-point vector
6217 comparisons against 0.
6219 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
6221 * lra-constraints.c (emit_spill_move): Remove assert.
6222 (process_alt_operands): Add more debugging
6223 output. Increase reject for spilling into memory. Decrease
6224 reject for reloading scratch.
6225 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
6227 2013-08-08 Steve Ellcey <sellcey@mips.com>
6229 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
6230 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
6231 micromips incompatible. Add nan2008.
6232 (MULTILIB_DIRNAMES): Add nan2008.
6233 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6234 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
6235 and micromips incompatible. Add nan2008.
6236 (MULTILIB_DIRNAMES): Add nan2008.
6237 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6239 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
6241 PR rtl-optimization/58079
6242 * combine.c (combine_simplify_rtx): Avoid using SUBST if
6243 simplify_comparison has widened a comparison with an integer.
6245 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6247 * config/arm/neon.md (movmisalign<mode>): Disable when we
6248 don't allow unaligned accesses.
6249 (*movmisalign<mode>_neon_store): Likewise.
6250 (*movmisalign<mode>_neon_load): Likewise.
6251 (*movmisalign<mode>_neon_store): Likewise.
6252 (*movmisalign<mode>_neon_load): Likewise.
6254 2013-08-08 Jan Hubicka <jh@suse.cz>
6256 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6257 (make_pass_rebuild_cgraph_edges): Also clear references.
6258 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6259 * ipa-inline-transform.c (inline_transform): Remove all references
6261 * cgraphunit.c (expand_function): Remove all references after
6263 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6264 (ipa_find_reference): Rewrite to iterator.
6265 (remove_stmt_references): Likewise.
6266 (ipa_clear_stmts_in_references): New function.
6267 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6268 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
6270 * ipa-split.c (split_function): Remove references in split function.
6272 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6275 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
6276 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
6278 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6281 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
6282 suggested mode for the assignment isn't compatible with the
6285 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
6288 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
6290 2013-08-07 Xinliang David Li <davidxl@google.com>
6292 * config/i386/i386.opt: New option -mtune-ctrl=.
6293 * config/i386/x86-tune.def: New file.
6294 * config/i386/i386.h: include x86-tune.def.
6295 * config/i386/i386.c (ix86_option_override_internal):
6296 Parsing -mtune-ctrl= option and set tune features.
6298 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
6301 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
6302 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
6305 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6307 * rtl.h (update_alignments): Declare.
6308 * final.c (grow_label_align): New function extracted from...
6309 (shorten_branches): ...here. Call it.
6310 (update_alignments): New function.
6311 * reorg.c (sibling_labels): New variable.
6312 (get_label_before): Add SIBLING parameter. If it is non-zero, push
6313 the new label along with it onto the sibling_labels vector.
6314 (fill_simple_delay_slots): Adjust call to get_label_before.
6315 (fill_slots_from_thread): Likewise.
6316 (relax_delay_slots): Likewise.
6317 (make_return_insns): Likewise.
6318 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
6320 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6322 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
6323 document its semantics.
6324 (diagnostic_report_diagnostic): Adjust accordingly.
6326 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6328 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
6329 (sparc_option_override): ...and port to new C++ pass API.
6330 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
6332 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
6334 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
6336 2013-08-06 Caroline Tice <cmtice@google.com>
6338 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
6339 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
6340 * tree-pass.h: Add pass_vtable_verify.
6341 * varasm.c (assemble_variable): Add code to properly set the comdat
6342 section and name for the .vtable_map_vars section.
6343 (assemble_vtyv_preinit_initializer): New function.
6344 (default_sectin_type_flags): Make sure .vtable_map_vars section has
6346 * output.h: Add function decl for assemble_vtv_preinit_initializer.
6347 * vtable-verify.c: New file.
6348 * vtable-verify.h: New file.
6349 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
6350 initialiation levels.
6351 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
6352 * passes.def: Insert pass_vtable_verify.
6353 * aclocal.m4: Reorder includes.
6354 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
6355 -fvtv-counts options.
6356 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
6357 as appropriate, if -fvtable-verify=... is used.
6358 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
6359 -fvtable-verify=... is used.
6360 * Makefile.in (OBJS): Add vtable-verify.o to list.
6361 (vtable-verify.o): Add new build rule.
6362 (GTFILES): Add vtable-verify.c to list.
6363 * common.opt (fvtable-verify=): New flag.
6364 (vtv_priority): Values for fvtable-verify= flag.
6365 (fvtv-counts): New flag.
6366 (fvtv-debug): New flag.
6367 * tree.h (save_vtable_map_decl): New extern function decl.
6369 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6371 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
6372 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
6373 (pass_data_rl78_devirt): ...new pass_data instance and...
6374 (make_pass_rl78_devirt): ...new function.
6375 (rl78_asm_file_start): Port pass registration to new C++ API.
6377 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6379 * coretypes.h (rtl_opt_pass): Add.
6380 (gcc::context): Add.
6381 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
6382 (epiphany_init): Port to new C++ pass API.
6383 (epiphany_optimize_mode_switching): Likewise.
6384 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
6385 (pass_manager::get_pass_mode_switching): New.
6386 (pass_manager::get_pass_peephole2): New.
6387 * mode-switching.c (pass_mode_switching): Add clone method.
6388 * recog.c (pass_peephole2): Add clone method.
6389 (pass_split_all_insns): Add clone method.
6391 2013-08-06 David Malcolm <dmalcolm@redhat.com>
6393 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
6394 (mips_option_override): ...here, porting to new C++ API for passes.
6396 2013-08-06 Jan Hubicka <jh@suse.cz>
6398 * cgraph.c (cgraph_get_body): New function based on lto.c
6400 * cgraph.h (cgraph_get_body): Declare.
6401 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
6403 * cgraphunit.c (expand_function): Get body prior expanding.
6404 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
6405 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
6407 * passes.c (do_per_function_toporder): Get body.
6408 * tree-inline.c (expand_call_inline): Get body prior inlining it.
6409 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
6411 2013-08-06 Martin Jambor <mjambor@suse.cz>
6414 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
6415 is not re-finalized. Rename second parameter to no_collect.
6417 2013-08-06 Martin Jambor <mjambor@suse.cz>
6420 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
6421 MEM_REF has proper alignment information.
6423 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
6426 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
6428 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
6429 rtx (*) (rtx, ...) with insn_gen_fn.
6430 * genoutput.c (output_insn_data): Cast gen_? function pointers to
6431 insn_gen_fn::stored_funcptr. Add initializer braces.
6433 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6435 Rewrite how instances of passes are cloned to remove assumptions
6436 about their sizes (thus allowing pass subclasses to have
6437 additional data fields, albeit non-GC-managed ones at this point).
6439 * passes.c (make_pass_instance): Now that passes have clone
6440 methods, rewrite this function to eliminate XNEW and memcpy
6441 calls that used hardcoded sizes. Since this function no longer
6442 creates pass instances, rename it to...
6443 (add_pass_instance): ...this. Document the old way that passes were
6444 numbered and flagged, and rework this function to continue using it.
6445 (next_pass_1): Add an initial_pass argument for use by
6447 (position_pass): When adding multiple instances of a pass, use
6448 the pass's clone method, rather than relying on the XNEW/memcpy
6449 within the former make_pass_instance (now add_pass_instance).
6450 (pass_manager::pass_manager): When invoking next_pass_1, also supply
6451 the initial instance of the current pass within the pass manager.
6453 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6455 This is the automated part of the conversion of passes from C
6456 structs to C++ classes.
6458 Patch autogenerated by refactor_passes.py from
6459 https://github.com/davidmalcolm/gcc-refactoring-scripts
6460 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
6462 * asan.c (pass_asan): Convert from a global struct to a subclass of
6463 gimple_opt_pass along with...
6464 (pass_data_asan): ...new pass_data instance and...
6465 (make_pass_asan): ...new function.
6466 (pass_asan_O0): Convert from a global struct to a subclass of
6467 gimple_opt_pass along with...
6468 (pass_data_asan_O0): ...new pass_data instance and...
6469 (make_pass_asan_O0): ...new function.
6470 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
6471 subclass of rtl_opt_pass along with...
6472 (pass_data_inc_dec): ...new pass_data instance and...
6473 (make_pass_inc_dec): ...new function.
6474 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
6475 a subclass of rtl_opt_pass along with...
6476 (pass_data_reorder_blocks): ...new pass_data instance and...
6477 (make_pass_reorder_blocks): ...new function.
6478 (pass_duplicate_computed_gotos): Convert from a global struct to a
6479 subclass of rtl_opt_pass along with...
6480 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
6481 (make_pass_duplicate_computed_gotos): ...new function.
6482 (pass_partition_blocks): Convert from a global struct to a subclass of
6483 rtl_opt_pass along with...
6484 (pass_data_partition_blocks): ...new pass_data instance and...
6485 (make_pass_partition_blocks): ...new function.
6486 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
6487 struct to a subclass of rtl_opt_pass along with...
6488 (pass_data_branch_target_load_optimize1): ...new pass_data instance
6490 (make_pass_branch_target_load_optimize1): ...new function.
6491 (pass_branch_target_load_optimize2): Convert from a global struct to a
6492 subclass of rtl_opt_pass along with...
6493 (pass_data_branch_target_load_optimize2): ...new pass_data instance
6495 (make_pass_branch_target_load_optimize2): ...new function.
6496 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
6497 of rtl_opt_pass along with...
6498 (pass_data_jump): ...new pass_data instance and...
6499 (make_pass_jump): ...new function.
6500 (pass_jump2): Convert from a global struct to a subclass of
6501 rtl_opt_pass along with...
6502 (pass_data_jump2): ...new pass_data instance and...
6503 (make_pass_jump2): ...new function.
6504 * cfgexpand.c (pass_expand): Convert from a global struct to a
6505 subclass of rtl_opt_pass along with...
6506 (pass_data_expand): ...new pass_data instance and...
6507 (make_pass_expand): ...new function.
6508 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
6509 of rtl_opt_pass along with...
6510 (pass_data_free_cfg): ...new pass_data instance and...
6511 (make_pass_free_cfg): ...new function.
6512 (pass_into_cfg_layout_mode): Convert from a global struct to a
6513 subclass of rtl_opt_pass along with...
6514 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
6515 (make_pass_into_cfg_layout_mode): ...new function.
6516 (pass_outof_cfg_layout_mode): Convert from a global struct to a
6517 subclass of rtl_opt_pass along with...
6518 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
6519 (make_pass_outof_cfg_layout_mode): ...new function.
6520 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
6521 struct to a subclass of gimple_opt_pass along with...
6522 (pass_data_build_cgraph_edges): ...new pass_data instance and...
6523 (make_pass_build_cgraph_edges): ...new function.
6524 (pass_rebuild_cgraph_edges): Convert from a global struct to a
6525 subclass of gimple_opt_pass along with...
6526 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
6527 (make_pass_rebuild_cgraph_edges): ...new function.
6528 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
6529 subclass of gimple_opt_pass along with...
6530 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
6532 (make_pass_remove_cgraph_callee_edges): ...new function.
6533 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
6534 struct to a subclass of rtl_opt_pass along with...
6535 (pass_data_stack_adjustments): ...new pass_data instance and...
6536 (make_pass_stack_adjustments): ...new function.
6537 * combine.c (pass_combine): Convert from a global struct to a subclass
6538 of rtl_opt_pass along with...
6539 (pass_data_combine): ...new pass_data instance and...
6540 (make_pass_combine): ...new function.
6541 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6542 global struct to a subclass of rtl_opt_pass along with...
6543 (pass_data_compare_elim_after_reload): ...new pass_data instance
6545 (make_pass_compare_elim_after_reload): ...new function.
6546 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6547 of rtl_opt_pass along with...
6548 (pass_data_rtl_cprop): ...new pass_data instance and...
6549 (make_pass_rtl_cprop): ...new function.
6550 * cse.c (pass_cse): Convert from a global struct to a subclass of
6551 rtl_opt_pass along with...
6552 (pass_data_cse): ...new pass_data instance and...
6553 (make_pass_cse): ...new function.
6554 (pass_cse2): Convert from a global struct to a subclass of
6555 rtl_opt_pass along with...
6556 (pass_data_cse2): ...new pass_data instance and...
6557 (make_pass_cse2): ...new function.
6558 (pass_cse_after_global_opts): Convert from a global struct to a
6559 subclass of rtl_opt_pass along with...
6560 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6561 (make_pass_cse_after_global_opts): ...new function.
6562 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6563 of rtl_opt_pass along with...
6564 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6565 (make_pass_ud_rtl_dce): ...new function.
6566 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6567 rtl_opt_pass along with...
6568 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6569 (make_pass_fast_rtl_dce): ...new function.
6570 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6571 a subclass of rtl_opt_pass along with...
6572 (pass_data_df_initialize_opt): ...new pass_data instance and...
6573 (make_pass_df_initialize_opt): ...new function.
6574 (pass_df_initialize_no_opt): Convert from a global struct to a
6575 subclass of rtl_opt_pass along with...
6576 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6577 (make_pass_df_initialize_no_opt): ...new function.
6578 (pass_df_finish): Convert from a global struct to a subclass of
6579 rtl_opt_pass along with...
6580 (pass_data_df_finish): ...new pass_data instance and...
6581 (make_pass_df_finish): ...new function.
6582 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6583 rtl_opt_pass along with...
6584 (pass_data_rtl_dse1): ...new pass_data instance and...
6585 (make_pass_rtl_dse1): ...new function.
6586 (pass_rtl_dse2): Convert from a global struct to a subclass of
6587 rtl_opt_pass along with...
6588 (pass_data_rtl_dse2): ...new pass_data instance and...
6589 (make_pass_rtl_dse2): ...new function.
6590 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6591 subclass of rtl_opt_pass along with...
6592 (pass_data_dwarf2_frame): ...new pass_data instance and...
6593 (make_pass_dwarf2_frame): ...new function.
6594 * except.c (pass_set_nothrow_function_flags): Convert from a global
6595 struct to a subclass of rtl_opt_pass along with...
6596 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6598 (make_pass_set_nothrow_function_flags): ...new function.
6599 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6600 subclass of rtl_opt_pass along with...
6601 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6603 (make_pass_convert_to_eh_region_ranges): ...new function.
6604 * final.c (pass_compute_alignments): Convert from a global struct to a
6605 subclass of rtl_opt_pass along with...
6606 (pass_data_compute_alignments): ...new pass_data instance and...
6607 (make_pass_compute_alignments): ...new function.
6608 (pass_final): Convert from a global struct to a subclass of
6609 rtl_opt_pass along with...
6610 (pass_data_final): ...new pass_data instance and...
6611 (make_pass_final): ...new function.
6612 (pass_shorten_branches): Convert from a global struct to a subclass of
6613 rtl_opt_pass along with...
6614 (pass_data_shorten_branches): ...new pass_data instance and...
6615 (make_pass_shorten_branches): ...new function.
6616 (pass_clean_state): Convert from a global struct to a subclass of
6617 rtl_opt_pass along with...
6618 (pass_data_clean_state): ...new pass_data instance and...
6619 (make_pass_clean_state): ...new function.
6620 * function.c (pass_instantiate_virtual_regs): Convert from a global
6621 struct to a subclass of rtl_opt_pass along with...
6622 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6623 (make_pass_instantiate_virtual_regs): ...new function.
6624 (pass_leaf_regs): Convert from a global struct to a subclass of
6625 rtl_opt_pass along with...
6626 (pass_data_leaf_regs): ...new pass_data instance and...
6627 (make_pass_leaf_regs): ...new function.
6628 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6629 subclass of rtl_opt_pass along with...
6630 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6632 (make_pass_thread_prologue_and_epilogue): ...new function.
6633 (pass_match_asm_constraints): Convert from a global struct to a
6634 subclass of rtl_opt_pass along with...
6635 (pass_data_match_asm_constraints): ...new pass_data instance and...
6636 (make_pass_match_asm_constraints): ...new function.
6637 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6638 subclass of rtl_opt_pass along with...
6639 (pass_data_rtl_fwprop): ...new pass_data instance and...
6640 (make_pass_rtl_fwprop): ...new function.
6641 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6642 rtl_opt_pass along with...
6643 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6644 (make_pass_rtl_fwprop_addr): ...new function.
6645 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6646 rtl_opt_pass along with...
6647 (pass_data_rtl_pre): ...new pass_data instance and...
6648 (make_pass_rtl_pre): ...new function.
6649 (pass_rtl_hoist): Convert from a global struct to a subclass of
6650 rtl_opt_pass along with...
6651 (pass_data_rtl_hoist): ...new pass_data instance and...
6652 (make_pass_rtl_hoist): ...new function.
6653 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6654 subclass of gimple_opt_pass along with...
6655 (pass_data_lower_cf): ...new pass_data instance and...
6656 (make_pass_lower_cf): ...new function.
6657 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6658 from a global struct to a subclass of gimple_opt_pass along with...
6659 (pass_data_strength_reduction): ...new pass_data instance and...
6660 (make_pass_strength_reduction): ...new function.
6661 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6662 of rtl_opt_pass along with...
6663 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6664 (make_pass_rtl_ifcvt): ...new function.
6665 (pass_if_after_combine): Convert from a global struct to a subclass of
6666 rtl_opt_pass along with...
6667 (pass_data_if_after_combine): ...new pass_data instance and...
6668 (make_pass_if_after_combine): ...new function.
6669 (pass_if_after_reload): Convert from a global struct to a subclass of
6670 rtl_opt_pass along with...
6671 (pass_data_if_after_reload): ...new pass_data instance and...
6672 (make_pass_if_after_reload): ...new function.
6673 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6674 a subclass of rtl_opt_pass along with...
6675 (pass_data_initialize_regs): ...new pass_data instance and...
6676 (make_pass_initialize_regs): ...new function.
6677 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6678 of ipa_opt_pass_d along with...
6679 (pass_data_ipa_cp): ...new pass_data instance and...
6680 (make_pass_ipa_cp): ...new function.
6681 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6682 global struct to a subclass of gimple_opt_pass along with...
6683 (pass_data_inline_parameters): ...new pass_data instance and...
6684 (make_pass_inline_parameters): ...new function.
6685 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6686 subclass of gimple_opt_pass along with...
6687 (pass_data_early_inline): ...new pass_data instance and...
6688 (make_pass_early_inline): ...new function.
6689 (pass_ipa_inline): Convert from a global struct to a subclass of
6690 ipa_opt_pass_d along with...
6691 (pass_data_ipa_inline): ...new pass_data instance and...
6692 (make_pass_ipa_inline): ...new function.
6693 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6694 struct to a subclass of gimple_opt_pass along with...
6695 (pass_data_local_pure_const): ...new pass_data instance and...
6696 (make_pass_local_pure_const): ...new function.
6697 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6698 ipa_opt_pass_d along with...
6699 (pass_data_ipa_pure_const): ...new pass_data instance and...
6700 (make_pass_ipa_pure_const): ...new function.
6701 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6702 to a subclass of ipa_opt_pass_d along with...
6703 (pass_data_ipa_reference): ...new pass_data instance and...
6704 (make_pass_ipa_reference): ...new function.
6705 * ipa-split.c (pass_split_functions): Convert from a global struct to
6706 a subclass of gimple_opt_pass along with...
6707 (pass_data_split_functions): ...new pass_data instance and...
6708 (make_pass_split_functions): ...new function.
6709 (pass_feedback_split_functions): Convert from a global struct to a
6710 subclass of gimple_opt_pass along with...
6711 (pass_data_feedback_split_functions): ...new pass_data instance and...
6712 (make_pass_feedback_split_functions): ...new function.
6713 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6714 global struct to a subclass of simple_ipa_opt_pass along with...
6715 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6717 (make_pass_ipa_function_and_variable_visibility): ...new function.
6718 (pass_ipa_free_inline_summary): Convert from a global struct to a
6719 subclass of simple_ipa_opt_pass along with...
6720 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6721 (make_pass_ipa_free_inline_summary): ...new function.
6722 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6723 subclass of ipa_opt_pass_d along with...
6724 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6726 (make_pass_ipa_whole_program_visibility): ...new function.
6727 (pass_ipa_profile): Convert from a global struct to a subclass of
6728 ipa_opt_pass_d along with...
6729 (pass_data_ipa_profile): ...new pass_data instance and...
6730 (make_pass_ipa_profile): ...new function.
6731 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6732 ipa_opt_pass_d along with...
6733 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6734 (make_pass_ipa_cdtor_merge): ...new function.
6735 * ira.c (pass_ira): Convert from a global struct to a subclass of
6736 rtl_opt_pass along with...
6737 (pass_data_ira): ...new pass_data instance and...
6738 (make_pass_ira): ...new function.
6739 (pass_reload): Convert from a global struct to a subclass of
6740 rtl_opt_pass along with...
6741 (pass_data_reload): ...new pass_data instance and...
6742 (make_pass_reload): ...new function.
6743 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6744 subclass of rtl_opt_pass along with...
6745 (pass_data_cleanup_barriers): ...new pass_data instance and...
6746 (make_pass_cleanup_barriers): ...new function.
6747 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6748 of rtl_opt_pass along with...
6749 (pass_data_loop2): ...new pass_data instance and...
6750 (make_pass_loop2): ...new function.
6751 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6752 rtl_opt_pass along with...
6753 (pass_data_rtl_loop_init): ...new pass_data instance and...
6754 (make_pass_rtl_loop_init): ...new function.
6755 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6756 rtl_opt_pass along with...
6757 (pass_data_rtl_loop_done): ...new pass_data instance and...
6758 (make_pass_rtl_loop_done): ...new function.
6759 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6760 subclass of rtl_opt_pass along with...
6761 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6762 (make_pass_rtl_move_loop_invariants): ...new function.
6763 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6764 rtl_opt_pass along with...
6765 (pass_data_rtl_unswitch): ...new pass_data instance and...
6766 (make_pass_rtl_unswitch): ...new function.
6767 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6768 subclass of rtl_opt_pass along with...
6769 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6771 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6772 (pass_rtl_doloop): Convert from a global struct to a subclass of
6773 rtl_opt_pass along with...
6774 (pass_data_rtl_doloop): ...new pass_data instance and...
6775 (make_pass_rtl_doloop): ...new function.
6776 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6777 a subclass of rtl_opt_pass along with...
6778 (pass_data_lower_subreg): ...new pass_data instance and...
6779 (make_pass_lower_subreg): ...new function.
6780 (pass_lower_subreg2): Convert from a global struct to a subclass of
6781 rtl_opt_pass along with...
6782 (pass_data_lower_subreg2): ...new pass_data instance and...
6783 (make_pass_lower_subreg2): ...new function.
6784 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6785 struct to a subclass of ipa_opt_pass_d along with...
6786 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6787 (make_pass_ipa_lto_gimple_out): ...new function.
6788 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6789 of ipa_opt_pass_d along with...
6790 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6791 (make_pass_ipa_lto_finish_out): ...new function.
6792 * mode-switching.c (pass_mode_switching): Convert from a global struct
6793 to a subclass of rtl_opt_pass along with...
6794 (pass_data_mode_switching): ...new pass_data instance and...
6795 (make_pass_mode_switching): ...new function.
6796 * modulo-sched.c (pass_sms): Convert from a global struct to a
6797 subclass of rtl_opt_pass along with...
6798 (pass_data_sms): ...new pass_data instance and...
6799 (make_pass_sms): ...new function.
6800 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6801 subclass of gimple_opt_pass along with...
6802 (pass_data_expand_omp): ...new pass_data instance and...
6803 (make_pass_expand_omp): ...new function.
6804 (pass_lower_omp): Convert from a global struct to a subclass of
6805 gimple_opt_pass along with...
6806 (pass_data_lower_omp): ...new pass_data instance and...
6807 (make_pass_lower_omp): ...new function.
6808 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6809 of gimple_opt_pass along with...
6810 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6811 (make_pass_diagnose_omp_blocks): ...new function.
6812 * passes.c (pass_early_local_passes): Convert from a global struct to
6813 a subclass of simple_ipa_opt_pass along with...
6814 (pass_data_early_local_passes): ...new pass_data instance and...
6815 (make_pass_early_local_passes): ...new function.
6816 (pass_all_early_optimizations): Convert from a global struct to a
6817 subclass of gimple_opt_pass along with...
6818 (pass_data_all_early_optimizations): ...new pass_data instance and...
6819 (make_pass_all_early_optimizations): ...new function.
6820 (pass_all_optimizations): Convert from a global struct to a subclass
6821 of gimple_opt_pass along with...
6822 (pass_data_all_optimizations): ...new pass_data instance and...
6823 (make_pass_all_optimizations): ...new function.
6824 (pass_all_optimizations_g): Convert from a global struct to a subclass
6825 of gimple_opt_pass along with...
6826 (pass_data_all_optimizations_g): ...new pass_data instance and...
6827 (make_pass_all_optimizations_g): ...new function.
6828 (pass_rest_of_compilation): Convert from a global struct to a subclass
6829 of rtl_opt_pass along with...
6830 (pass_data_rest_of_compilation): ...new pass_data instance and...
6831 (make_pass_rest_of_compilation): ...new function.
6832 (pass_postreload): Convert from a global struct to a subclass of
6833 rtl_opt_pass along with...
6834 (pass_data_postreload): ...new pass_data instance and...
6835 (make_pass_postreload): ...new function.
6836 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6837 subclass of rtl_opt_pass along with...
6838 (pass_data_gcse2): ...new pass_data instance and...
6839 (make_pass_gcse2): ...new function.
6840 * postreload.c (pass_postreload_cse): Convert from a global struct to
6841 a subclass of rtl_opt_pass along with...
6842 (pass_data_postreload_cse): ...new pass_data instance and...
6843 (make_pass_postreload_cse): ...new function.
6844 * predict.c (pass_profile): Convert from a global struct to a subclass
6845 of gimple_opt_pass along with...
6846 (pass_data_profile): ...new pass_data instance and...
6847 (make_pass_profile): ...new function.
6848 (pass_strip_predict_hints): Convert from a global struct to a subclass
6849 of gimple_opt_pass along with...
6850 (pass_data_strip_predict_hints): ...new pass_data instance and...
6851 (make_pass_strip_predict_hints): ...new function.
6852 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6853 of rtl_opt_pass along with...
6854 (pass_data_peephole2): ...new pass_data instance and...
6855 (make_pass_peephole2): ...new function.
6856 (pass_split_all_insns): Convert from a global struct to a subclass of
6857 rtl_opt_pass along with...
6858 (pass_data_split_all_insns): ...new pass_data instance and...
6859 (make_pass_split_all_insns): ...new function.
6860 (pass_split_after_reload): Convert from a global struct to a subclass
6861 of rtl_opt_pass along with...
6862 (pass_data_split_after_reload): ...new pass_data instance and...
6863 (make_pass_split_after_reload): ...new function.
6864 (pass_split_before_regstack): Convert from a global struct to a
6865 subclass of rtl_opt_pass along with...
6866 (pass_data_split_before_regstack): ...new pass_data instance and...
6867 (make_pass_split_before_regstack): ...new function.
6868 (pass_split_before_sched2): Convert from a global struct to a subclass
6869 of rtl_opt_pass along with...
6870 (pass_data_split_before_sched2): ...new pass_data instance and...
6871 (make_pass_split_before_sched2): ...new function.
6872 (pass_split_for_shorten_branches): Convert from a global struct to a
6873 subclass of rtl_opt_pass along with...
6874 (pass_data_split_for_shorten_branches): ...new pass_data instance
6876 (make_pass_split_for_shorten_branches): ...new function.
6877 * ree.c (pass_ree): Convert from a global struct to a subclass of
6878 rtl_opt_pass along with...
6879 (pass_data_ree): ...new pass_data instance and...
6880 (make_pass_ree): ...new function.
6881 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6882 subclass of rtl_opt_pass along with...
6883 (pass_data_stack_regs): ...new pass_data instance and...
6884 (make_pass_stack_regs): ...new function.
6885 (pass_stack_regs_run): Convert from a global struct to a subclass of
6886 rtl_opt_pass along with...
6887 (pass_data_stack_regs_run): ...new pass_data instance and...
6888 (make_pass_stack_regs_run): ...new function.
6889 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6890 subclass of rtl_opt_pass along with...
6891 (pass_data_cprop_hardreg): ...new pass_data instance and...
6892 (make_pass_cprop_hardreg): ...new function.
6893 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6894 subclass of rtl_opt_pass along with...
6895 (pass_data_reginfo_init): ...new pass_data instance and...
6896 (make_pass_reginfo_init): ...new function.
6897 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6898 of rtl_opt_pass along with...
6899 (pass_data_regmove): ...new pass_data instance and...
6900 (make_pass_regmove): ...new function.
6901 * regrename.c (pass_regrename): Convert from a global struct to a
6902 subclass of rtl_opt_pass along with...
6903 (pass_data_regrename): ...new pass_data instance and...
6904 (make_pass_regrename): ...new function.
6905 * reorg.c (pass_delay_slots): Convert from a global struct to a
6906 subclass of rtl_opt_pass along with...
6907 (pass_data_delay_slots): ...new pass_data instance and...
6908 (make_pass_delay_slots): ...new function.
6909 (pass_machine_reorg): Convert from a global struct to a subclass of
6910 rtl_opt_pass along with...
6911 (pass_data_machine_reorg): ...new pass_data instance and...
6912 (make_pass_machine_reorg): ...new function.
6913 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6914 of rtl_opt_pass along with...
6915 (pass_data_sched): ...new pass_data instance and...
6916 (make_pass_sched): ...new function.
6917 (pass_sched2): Convert from a global struct to a subclass of
6918 rtl_opt_pass along with...
6919 (pass_data_sched2): ...new pass_data instance and...
6920 (make_pass_sched2): ...new function.
6921 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6922 to a subclass of rtl_opt_pass along with...
6923 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6924 (make_pass_stack_ptr_mod): ...new function.
6925 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6926 to a subclass of rtl_opt_pass along with...
6927 (pass_data_rtl_store_motion): ...new pass_data instance and...
6928 (make_pass_rtl_store_motion): ...new function.
6929 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6930 of gimple_opt_pass along with...
6931 (pass_data_tracer): ...new pass_data instance and...
6932 (make_pass_tracer): ...new function.
6933 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6934 to a subclass of gimple_opt_pass along with...
6935 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6936 (make_pass_diagnose_tm_blocks): ...new function.
6937 (pass_lower_tm): Convert from a global struct to a subclass of
6938 gimple_opt_pass along with...
6939 (pass_data_lower_tm): ...new pass_data instance and...
6940 (make_pass_lower_tm): ...new function.
6941 (pass_tm_init): Convert from a global struct to a subclass of
6942 gimple_opt_pass along with...
6943 (pass_data_tm_init): ...new pass_data instance and...
6944 (make_pass_tm_init): ...new function.
6945 (pass_tm_mark): Convert from a global struct to a subclass of
6946 gimple_opt_pass along with...
6947 (pass_data_tm_mark): ...new pass_data instance and...
6948 (make_pass_tm_mark): ...new function.
6949 (pass_tm_edges): Convert from a global struct to a subclass of
6950 gimple_opt_pass along with...
6951 (pass_data_tm_edges): ...new pass_data instance and...
6952 (make_pass_tm_edges): ...new function.
6953 (pass_tm_memopt): Convert from a global struct to a subclass of
6954 gimple_opt_pass along with...
6955 (pass_data_tm_memopt): ...new pass_data instance and...
6956 (make_pass_tm_memopt): ...new function.
6957 (pass_ipa_tm): Convert from a global struct to a subclass of
6958 simple_ipa_opt_pass along with...
6959 (pass_data_ipa_tm): ...new pass_data instance and...
6960 (make_pass_ipa_tm): ...new function.
6961 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6962 subclass of gimple_opt_pass along with...
6963 (pass_data_call_cdce): ...new pass_data instance and...
6964 (make_pass_call_cdce): ...new function.
6965 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6966 subclass of gimple_opt_pass along with...
6967 (pass_data_build_cfg): ...new pass_data instance and...
6968 (make_pass_build_cfg): ...new function.
6969 (pass_split_crit_edges): Convert from a global struct to a subclass of
6970 gimple_opt_pass along with...
6971 (pass_data_split_crit_edges): ...new pass_data instance and...
6972 (make_pass_split_crit_edges): ...new function.
6973 (pass_warn_function_return): Convert from a global struct to a
6974 subclass of gimple_opt_pass along with...
6975 (pass_data_warn_function_return): ...new pass_data instance and...
6976 (make_pass_warn_function_return): ...new function.
6977 (pass_warn_function_noreturn): Convert from a global struct to a
6978 subclass of gimple_opt_pass along with...
6979 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6980 (make_pass_warn_function_noreturn): ...new function.
6981 (pass_warn_unused_result): Convert from a global struct to a subclass
6982 of gimple_opt_pass along with...
6983 (pass_data_warn_unused_result): ...new pass_data instance and...
6984 (make_pass_warn_unused_result): ...new function.
6985 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6986 a subclass of gimple_opt_pass along with...
6987 (pass_data_merge_phi): ...new pass_data instance and...
6988 (make_pass_merge_phi): ...new function.
6989 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6990 a subclass of gimple_opt_pass along with...
6991 (pass_data_lower_complex): ...new pass_data instance and...
6992 (make_pass_lower_complex): ...new function.
6993 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6994 gimple_opt_pass along with...
6995 (pass_data_lower_complex_O0): ...new pass_data instance and...
6996 (make_pass_lower_complex_O0): ...new function.
6997 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6998 subclass of gimple_opt_pass along with...
6999 (pass_data_lower_eh): ...new pass_data instance and...
7000 (make_pass_lower_eh): ...new function.
7001 (pass_refactor_eh): Convert from a global struct to a subclass of
7002 gimple_opt_pass along with...
7003 (pass_data_refactor_eh): ...new pass_data instance and...
7004 (make_pass_refactor_eh): ...new function.
7005 (pass_lower_resx): Convert from a global struct to a subclass of
7006 gimple_opt_pass along with...
7007 (pass_data_lower_resx): ...new pass_data instance and...
7008 (make_pass_lower_resx): ...new function.
7009 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
7010 of gimple_opt_pass along with...
7011 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
7012 (make_pass_lower_eh_dispatch): ...new function.
7013 (pass_cleanup_eh): Convert from a global struct to a subclass of
7014 gimple_opt_pass along with...
7015 (pass_data_cleanup_eh): ...new pass_data instance and...
7016 (make_pass_cleanup_eh): ...new function.
7017 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7018 to a subclass of simple_ipa_opt_pass along with...
7019 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7020 (make_pass_ipa_lower_emutls): ...new function.
7021 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7022 a subclass of gimple_opt_pass along with...
7023 (pass_data_if_conversion): ...new pass_data instance and...
7024 (make_pass_if_conversion): ...new function.
7025 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7026 subclass of gimple_opt_pass along with...
7027 (pass_data_build_ssa): ...new pass_data instance and...
7028 (make_pass_build_ssa): ...new function.
7029 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7030 global struct to a subclass of gimple_opt_pass along with...
7031 (pass_data_loop_distribution): ...new pass_data instance and...
7032 (make_pass_loop_distribution): ...new function.
7033 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7034 subclass of gimple_opt_pass along with...
7035 (pass_data_mudflap_1): ...new pass_data instance and...
7036 (make_pass_mudflap_1): ...new function.
7037 (pass_mudflap_2): Convert from a global struct to a subclass of
7038 gimple_opt_pass along with...
7039 (pass_data_mudflap_2): ...new pass_data instance and...
7040 (make_pass_mudflap_2): ...new function.
7041 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7042 subclass of gimple_opt_pass along with...
7043 (pass_data_mudflap_1): ...new pass_data instance and...
7044 (make_pass_mudflap_1): ...new function.
7045 (pass_mudflap_2): Convert from a global struct to a subclass of
7046 gimple_opt_pass along with...
7047 (pass_data_mudflap_2): ...new pass_data instance and...
7048 (make_pass_mudflap_2): ...new function.
7049 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7050 gimple_opt_pass along with...
7051 (pass_data_nrv): ...new pass_data instance and...
7052 (make_pass_nrv): ...new function.
7053 (pass_return_slot): Convert from a global struct to a subclass of
7054 gimple_opt_pass along with...
7055 (pass_data_return_slot): ...new pass_data instance and...
7056 (make_pass_return_slot): ...new function.
7057 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7058 to a subclass of gimple_opt_pass along with...
7059 (pass_data_object_sizes): ...new pass_data instance and...
7060 (make_pass_object_sizes): ...new function.
7061 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7062 global struct to a subclass of gimple_opt_pass along with...
7063 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7065 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7066 (pass_fixup_cfg): Convert from a global struct to a subclass of
7067 gimple_opt_pass along with...
7068 (pass_data_fixup_cfg): ...new pass_data instance and...
7069 (make_pass_fixup_cfg): ...new function.
7070 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7071 (make_pass_mudflap_1): ...new function.
7072 (pass_mudflap_2): Replace declaration with that of...
7073 (make_pass_mudflap_2): ...new function.
7074 (pass_asan): Replace declaration with that of...
7075 (make_pass_asan): ...new function.
7076 (pass_asan_O0): Replace declaration with that of...
7077 (make_pass_asan_O0): ...new function.
7078 (pass_tsan): Replace declaration with that of...
7079 (make_pass_tsan): ...new function.
7080 (pass_tsan_O0): Replace declaration with that of...
7081 (make_pass_tsan_O0): ...new function.
7082 (pass_lower_cf): Replace declaration with that of...
7083 (make_pass_lower_cf): ...new function.
7084 (pass_refactor_eh): Replace declaration with that of...
7085 (make_pass_refactor_eh): ...new function.
7086 (pass_lower_eh): Replace declaration with that of...
7087 (make_pass_lower_eh): ...new function.
7088 (pass_lower_eh_dispatch): Replace declaration with that of...
7089 (make_pass_lower_eh_dispatch): ...new function.
7090 (pass_lower_resx): Replace declaration with that of...
7091 (make_pass_lower_resx): ...new function.
7092 (pass_build_cfg): Replace declaration with that of...
7093 (make_pass_build_cfg): ...new function.
7094 (pass_early_tree_profile): Replace declaration with that of...
7095 (make_pass_early_tree_profile): ...new function.
7096 (pass_cleanup_eh): Replace declaration with that of...
7097 (make_pass_cleanup_eh): ...new function.
7098 (pass_sra): Replace declaration with that of...
7099 (make_pass_sra): ...new function.
7100 (pass_sra_early): Replace declaration with that of...
7101 (make_pass_sra_early): ...new function.
7102 (pass_early_ipa_sra): Replace declaration with that of...
7103 (make_pass_early_ipa_sra): ...new function.
7104 (pass_tail_recursion): Replace declaration with that of...
7105 (make_pass_tail_recursion): ...new function.
7106 (pass_tail_calls): Replace declaration with that of...
7107 (make_pass_tail_calls): ...new function.
7108 (pass_tree_loop): Replace declaration with that of...
7109 (make_pass_tree_loop): ...new function.
7110 (pass_tree_loop_init): Replace declaration with that of...
7111 (make_pass_tree_loop_init): ...new function.
7112 (pass_lim): Replace declaration with that of...
7113 (make_pass_lim): ...new function.
7114 (pass_tree_unswitch): Replace declaration with that of...
7115 (make_pass_tree_unswitch): ...new function.
7116 (pass_predcom): Replace declaration with that of...
7117 (make_pass_predcom): ...new function.
7118 (pass_iv_canon): Replace declaration with that of...
7119 (make_pass_iv_canon): ...new function.
7120 (pass_scev_cprop): Replace declaration with that of...
7121 (make_pass_scev_cprop): ...new function.
7122 (pass_empty_loop): Replace declaration with that of...
7123 (make_pass_empty_loop): ...new function.
7124 (pass_record_bounds): Replace declaration with that of...
7125 (make_pass_record_bounds): ...new function.
7126 (pass_graphite): Replace declaration with that of...
7127 (make_pass_graphite): ...new function.
7128 (pass_graphite_transforms): Replace declaration with that of...
7129 (make_pass_graphite_transforms): ...new function.
7130 (pass_if_conversion): Replace declaration with that of...
7131 (make_pass_if_conversion): ...new function.
7132 (pass_loop_distribution): Replace declaration with that of...
7133 (make_pass_loop_distribution): ...new function.
7134 (pass_vectorize): Replace declaration with that of...
7135 (make_pass_vectorize): ...new function.
7136 (pass_slp_vectorize): Replace declaration with that of...
7137 (make_pass_slp_vectorize): ...new function.
7138 (pass_complete_unroll): Replace declaration with that of...
7139 (make_pass_complete_unroll): ...new function.
7140 (pass_complete_unrolli): Replace declaration with that of...
7141 (make_pass_complete_unrolli): ...new function.
7142 (pass_parallelize_loops): Replace declaration with that of...
7143 (make_pass_parallelize_loops): ...new function.
7144 (pass_loop_prefetch): Replace declaration with that of...
7145 (make_pass_loop_prefetch): ...new function.
7146 (pass_iv_optimize): Replace declaration with that of...
7147 (make_pass_iv_optimize): ...new function.
7148 (pass_tree_loop_done): Replace declaration with that of...
7149 (make_pass_tree_loop_done): ...new function.
7150 (pass_ch): Replace declaration with that of...
7151 (make_pass_ch): ...new function.
7152 (pass_ccp): Replace declaration with that of...
7153 (make_pass_ccp): ...new function.
7154 (pass_phi_only_cprop): Replace declaration with that of...
7155 (make_pass_phi_only_cprop): ...new function.
7156 (pass_build_ssa): Replace declaration with that of...
7157 (make_pass_build_ssa): ...new function.
7158 (pass_build_alias): Replace declaration with that of...
7159 (make_pass_build_alias): ...new function.
7160 (pass_build_ealias): Replace declaration with that of...
7161 (make_pass_build_ealias): ...new function.
7162 (pass_dominator): Replace declaration with that of...
7163 (make_pass_dominator): ...new function.
7164 (pass_dce): Replace declaration with that of...
7165 (make_pass_dce): ...new function.
7166 (pass_dce_loop): Replace declaration with that of...
7167 (make_pass_dce_loop): ...new function.
7168 (pass_cd_dce): Replace declaration with that of...
7169 (make_pass_cd_dce): ...new function.
7170 (pass_call_cdce): Replace declaration with that of...
7171 (make_pass_call_cdce): ...new function.
7172 (pass_merge_phi): Replace declaration with that of...
7173 (make_pass_merge_phi): ...new function.
7174 (pass_split_crit_edges): Replace declaration with that of...
7175 (make_pass_split_crit_edges): ...new function.
7176 (pass_pre): Replace declaration with that of...
7177 (make_pass_pre): ...new function.
7178 (pass_profile): Replace declaration with that of...
7179 (make_pass_profile): ...new function.
7180 (pass_strip_predict_hints): Replace declaration with that of...
7181 (make_pass_strip_predict_hints): ...new function.
7182 (pass_lower_complex_O0): Replace declaration with that of...
7183 (make_pass_lower_complex_O0): ...new function.
7184 (pass_lower_complex): Replace declaration with that of...
7185 (make_pass_lower_complex): ...new function.
7186 (pass_lower_vector): Replace declaration with that of...
7187 (make_pass_lower_vector): ...new function.
7188 (pass_lower_vector_ssa): Replace declaration with that of...
7189 (make_pass_lower_vector_ssa): ...new function.
7190 (pass_lower_omp): Replace declaration with that of...
7191 (make_pass_lower_omp): ...new function.
7192 (pass_diagnose_omp_blocks): Replace declaration with that of...
7193 (make_pass_diagnose_omp_blocks): ...new function.
7194 (pass_expand_omp): Replace declaration with that of...
7195 (make_pass_expand_omp): ...new function.
7196 (pass_expand_omp_ssa): Replace declaration with that of...
7197 (make_pass_expand_omp_ssa): ...new function.
7198 (pass_object_sizes): Replace declaration with that of...
7199 (make_pass_object_sizes): ...new function.
7200 (pass_strlen): Replace declaration with that of...
7201 (make_pass_strlen): ...new function.
7202 (pass_fold_builtins): Replace declaration with that of...
7203 (make_pass_fold_builtins): ...new function.
7204 (pass_stdarg): Replace declaration with that of...
7205 (make_pass_stdarg): ...new function.
7206 (pass_early_warn_uninitialized): Replace declaration with that of...
7207 (make_pass_early_warn_uninitialized): ...new function.
7208 (pass_late_warn_uninitialized): Replace declaration with that of...
7209 (make_pass_late_warn_uninitialized): ...new function.
7210 (pass_cse_reciprocals): Replace declaration with that of...
7211 (make_pass_cse_reciprocals): ...new function.
7212 (pass_cse_sincos): Replace declaration with that of...
7213 (make_pass_cse_sincos): ...new function.
7214 (pass_optimize_bswap): Replace declaration with that of...
7215 (make_pass_optimize_bswap): ...new function.
7216 (pass_optimize_widening_mul): Replace declaration with that of...
7217 (make_pass_optimize_widening_mul): ...new function.
7218 (pass_warn_function_return): Replace declaration with that of...
7219 (make_pass_warn_function_return): ...new function.
7220 (pass_warn_function_noreturn): Replace declaration with that of...
7221 (make_pass_warn_function_noreturn): ...new function.
7222 (pass_cselim): Replace declaration with that of...
7223 (make_pass_cselim): ...new function.
7224 (pass_phiopt): Replace declaration with that of...
7225 (make_pass_phiopt): ...new function.
7226 (pass_forwprop): Replace declaration with that of...
7227 (make_pass_forwprop): ...new function.
7228 (pass_phiprop): Replace declaration with that of...
7229 (make_pass_phiprop): ...new function.
7230 (pass_tree_ifcombine): Replace declaration with that of...
7231 (make_pass_tree_ifcombine): ...new function.
7232 (pass_dse): Replace declaration with that of...
7233 (make_pass_dse): ...new function.
7234 (pass_nrv): Replace declaration with that of...
7235 (make_pass_nrv): ...new function.
7236 (pass_rename_ssa_copies): Replace declaration with that of...
7237 (make_pass_rename_ssa_copies): ...new function.
7238 (pass_sink_code): Replace declaration with that of...
7239 (make_pass_sink_code): ...new function.
7240 (pass_fre): Replace declaration with that of...
7241 (make_pass_fre): ...new function.
7242 (pass_check_data_deps): Replace declaration with that of...
7243 (make_pass_check_data_deps): ...new function.
7244 (pass_copy_prop): Replace declaration with that of...
7245 (make_pass_copy_prop): ...new function.
7246 (pass_vrp): Replace declaration with that of...
7247 (make_pass_vrp): ...new function.
7248 (pass_uncprop): Replace declaration with that of...
7249 (make_pass_uncprop): ...new function.
7250 (pass_return_slot): Replace declaration with that of...
7251 (make_pass_return_slot): ...new function.
7252 (pass_reassoc): Replace declaration with that of...
7253 (make_pass_reassoc): ...new function.
7254 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7255 (make_pass_rebuild_cgraph_edges): ...new function.
7256 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7257 (make_pass_remove_cgraph_callee_edges): ...new function.
7258 (pass_build_cgraph_edges): Replace declaration with that of...
7259 (make_pass_build_cgraph_edges): ...new function.
7260 (pass_local_pure_const): Replace declaration with that of...
7261 (make_pass_local_pure_const): ...new function.
7262 (pass_tracer): Replace declaration with that of...
7263 (make_pass_tracer): ...new function.
7264 (pass_warn_unused_result): Replace declaration with that of...
7265 (make_pass_warn_unused_result): ...new function.
7266 (pass_diagnose_tm_blocks): Replace declaration with that of...
7267 (make_pass_diagnose_tm_blocks): ...new function.
7268 (pass_lower_tm): Replace declaration with that of...
7269 (make_pass_lower_tm): ...new function.
7270 (pass_tm_init): Replace declaration with that of...
7271 (make_pass_tm_init): ...new function.
7272 (pass_tm_mark): Replace declaration with that of...
7273 (make_pass_tm_mark): ...new function.
7274 (pass_tm_memopt): Replace declaration with that of...
7275 (make_pass_tm_memopt): ...new function.
7276 (pass_tm_edges): Replace declaration with that of...
7277 (make_pass_tm_edges): ...new function.
7278 (pass_split_functions): Replace declaration with that of...
7279 (make_pass_split_functions): ...new function.
7280 (pass_feedback_split_functions): Replace declaration with that of...
7281 (make_pass_feedback_split_functions): ...new function.
7282 (pass_strength_reduction): Replace declaration with that of...
7283 (make_pass_strength_reduction): ...new function.
7284 (pass_ipa_lower_emutls): Replace declaration with that of...
7285 (make_pass_ipa_lower_emutls): ...new function.
7286 (pass_ipa_function_and_variable_visibility): Replace declaration with
7288 (make_pass_ipa_function_and_variable_visibility): ...new function.
7289 (pass_ipa_tree_profile): Replace declaration with that of...
7290 (make_pass_ipa_tree_profile): ...new function.
7291 (pass_early_local_passes): Replace declaration with that of...
7292 (make_pass_early_local_passes): ...new function.
7293 (pass_ipa_whole_program_visibility): Replace declaration with that
7295 (make_pass_ipa_whole_program_visibility): ...new function.
7296 (pass_ipa_lto_gimple_out): Replace declaration with that of...
7297 (make_pass_ipa_lto_gimple_out): ...new function.
7298 (pass_ipa_increase_alignment): Replace declaration with that of...
7299 (make_pass_ipa_increase_alignment): ...new function.
7300 (pass_ipa_inline): Replace declaration with that of...
7301 (make_pass_ipa_inline): ...new function.
7302 (pass_ipa_free_lang_data): Replace declaration with that of...
7303 (make_pass_ipa_free_lang_data): ...new function.
7304 (pass_ipa_free_inline_summary): Replace declaration with that of...
7305 (make_pass_ipa_free_inline_summary): ...new function.
7306 (pass_ipa_cp): Replace declaration with that of...
7307 (make_pass_ipa_cp): ...new function.
7308 (pass_ipa_reference): Replace declaration with that of...
7309 (make_pass_ipa_reference): ...new function.
7310 (pass_ipa_pure_const): Replace declaration with that of...
7311 (make_pass_ipa_pure_const): ...new function.
7312 (pass_ipa_pta): Replace declaration with that of...
7313 (make_pass_ipa_pta): ...new function.
7314 (pass_ipa_lto_finish_out): Replace declaration with that of...
7315 (make_pass_ipa_lto_finish_out): ...new function.
7316 (pass_ipa_tm): Replace declaration with that of...
7317 (make_pass_ipa_tm): ...new function.
7318 (pass_ipa_profile): Replace declaration with that of...
7319 (make_pass_ipa_profile): ...new function.
7320 (pass_ipa_cdtor_merge): Replace declaration with that of...
7321 (make_pass_ipa_cdtor_merge): ...new function.
7322 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
7324 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7325 (pass_init_datastructures): Replace declaration with that of...
7326 (make_pass_init_datastructures): ...new function.
7327 (pass_fixup_cfg): Replace declaration with that of...
7328 (make_pass_fixup_cfg): ...new function.
7329 (pass_expand): Replace declaration with that of...
7330 (make_pass_expand): ...new function.
7331 (pass_instantiate_virtual_regs): Replace declaration with that of...
7332 (make_pass_instantiate_virtual_regs): ...new function.
7333 (pass_rtl_fwprop): Replace declaration with that of...
7334 (make_pass_rtl_fwprop): ...new function.
7335 (pass_rtl_fwprop_addr): Replace declaration with that of...
7336 (make_pass_rtl_fwprop_addr): ...new function.
7337 (pass_jump): Replace declaration with that of...
7338 (make_pass_jump): ...new function.
7339 (pass_jump2): Replace declaration with that of...
7340 (make_pass_jump2): ...new function.
7341 (pass_lower_subreg): Replace declaration with that of...
7342 (make_pass_lower_subreg): ...new function.
7343 (pass_cse): Replace declaration with that of...
7344 (make_pass_cse): ...new function.
7345 (pass_fast_rtl_dce): Replace declaration with that of...
7346 (make_pass_fast_rtl_dce): ...new function.
7347 (pass_ud_rtl_dce): Replace declaration with that of...
7348 (make_pass_ud_rtl_dce): ...new function.
7349 (pass_rtl_dce): Replace declaration with that of...
7350 (make_pass_rtl_dce): ...new function.
7351 (pass_rtl_dse1): Replace declaration with that of...
7352 (make_pass_rtl_dse1): ...new function.
7353 (pass_rtl_dse2): Replace declaration with that of...
7354 (make_pass_rtl_dse2): ...new function.
7355 (pass_rtl_dse3): Replace declaration with that of...
7356 (make_pass_rtl_dse3): ...new function.
7357 (pass_rtl_cprop): Replace declaration with that of...
7358 (make_pass_rtl_cprop): ...new function.
7359 (pass_rtl_pre): Replace declaration with that of...
7360 (make_pass_rtl_pre): ...new function.
7361 (pass_rtl_hoist): Replace declaration with that of...
7362 (make_pass_rtl_hoist): ...new function.
7363 (pass_rtl_store_motion): Replace declaration with that of...
7364 (make_pass_rtl_store_motion): ...new function.
7365 (pass_cse_after_global_opts): Replace declaration with that of...
7366 (make_pass_cse_after_global_opts): ...new function.
7367 (pass_rtl_ifcvt): Replace declaration with that of...
7368 (make_pass_rtl_ifcvt): ...new function.
7369 (pass_into_cfg_layout_mode): Replace declaration with that of...
7370 (make_pass_into_cfg_layout_mode): ...new function.
7371 (pass_outof_cfg_layout_mode): Replace declaration with that of...
7372 (make_pass_outof_cfg_layout_mode): ...new function.
7373 (pass_loop2): Replace declaration with that of...
7374 (make_pass_loop2): ...new function.
7375 (pass_rtl_loop_init): Replace declaration with that of...
7376 (make_pass_rtl_loop_init): ...new function.
7377 (pass_rtl_move_loop_invariants): Replace declaration with that of...
7378 (make_pass_rtl_move_loop_invariants): ...new function.
7379 (pass_rtl_unswitch): Replace declaration with that of...
7380 (make_pass_rtl_unswitch): ...new function.
7381 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
7382 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7383 (pass_rtl_doloop): Replace declaration with that of...
7384 (make_pass_rtl_doloop): ...new function.
7385 (pass_rtl_loop_done): Replace declaration with that of...
7386 (make_pass_rtl_loop_done): ...new function.
7387 (pass_web): Replace declaration with that of...
7388 (make_pass_web): ...new function.
7389 (pass_cse2): Replace declaration with that of...
7390 (make_pass_cse2): ...new function.
7391 (pass_df_initialize_opt): Replace declaration with that of...
7392 (make_pass_df_initialize_opt): ...new function.
7393 (pass_df_initialize_no_opt): Replace declaration with that of...
7394 (make_pass_df_initialize_no_opt): ...new function.
7395 (pass_reginfo_init): Replace declaration with that of...
7396 (make_pass_reginfo_init): ...new function.
7397 (pass_inc_dec): Replace declaration with that of...
7398 (make_pass_inc_dec): ...new function.
7399 (pass_stack_ptr_mod): Replace declaration with that of...
7400 (make_pass_stack_ptr_mod): ...new function.
7401 (pass_initialize_regs): Replace declaration with that of...
7402 (make_pass_initialize_regs): ...new function.
7403 (pass_combine): Replace declaration with that of...
7404 (make_pass_combine): ...new function.
7405 (pass_if_after_combine): Replace declaration with that of...
7406 (make_pass_if_after_combine): ...new function.
7407 (pass_ree): Replace declaration with that of...
7408 (make_pass_ree): ...new function.
7409 (pass_partition_blocks): Replace declaration with that of...
7410 (make_pass_partition_blocks): ...new function.
7411 (pass_match_asm_constraints): Replace declaration with that of...
7412 (make_pass_match_asm_constraints): ...new function.
7413 (pass_regmove): Replace declaration with that of...
7414 (make_pass_regmove): ...new function.
7415 (pass_split_all_insns): Replace declaration with that of...
7416 (make_pass_split_all_insns): ...new function.
7417 (pass_fast_rtl_byte_dce): Replace declaration with that of...
7418 (make_pass_fast_rtl_byte_dce): ...new function.
7419 (pass_lower_subreg2): Replace declaration with that of...
7420 (make_pass_lower_subreg2): ...new function.
7421 (pass_mode_switching): Replace declaration with that of...
7422 (make_pass_mode_switching): ...new function.
7423 (pass_sms): Replace declaration with that of...
7424 (make_pass_sms): ...new function.
7425 (pass_sched): Replace declaration with that of...
7426 (make_pass_sched): ...new function.
7427 (pass_ira): Replace declaration with that of...
7428 (make_pass_ira): ...new function.
7429 (pass_reload): Replace declaration with that of...
7430 (make_pass_reload): ...new function.
7431 (pass_clean_state): Replace declaration with that of...
7432 (make_pass_clean_state): ...new function.
7433 (pass_branch_prob): Replace declaration with that of...
7434 (make_pass_branch_prob): ...new function.
7435 (pass_value_profile_transformations): Replace declaration with that
7437 (make_pass_value_profile_transformations): ...new function.
7438 (pass_postreload_cse): Replace declaration with that of...
7439 (make_pass_postreload_cse): ...new function.
7440 (pass_gcse2): Replace declaration with that of...
7441 (make_pass_gcse2): ...new function.
7442 (pass_split_after_reload): Replace declaration with that of...
7443 (make_pass_split_after_reload): ...new function.
7444 (pass_branch_target_load_optimize1): Replace declaration with that
7446 (make_pass_branch_target_load_optimize1): ...new function.
7447 (pass_thread_prologue_and_epilogue): Replace declaration with that
7449 (make_pass_thread_prologue_and_epilogue): ...new function.
7450 (pass_stack_adjustments): Replace declaration with that of...
7451 (make_pass_stack_adjustments): ...new function.
7452 (pass_peephole2): Replace declaration with that of...
7453 (make_pass_peephole2): ...new function.
7454 (pass_if_after_reload): Replace declaration with that of...
7455 (make_pass_if_after_reload): ...new function.
7456 (pass_regrename): Replace declaration with that of...
7457 (make_pass_regrename): ...new function.
7458 (pass_cprop_hardreg): Replace declaration with that of...
7459 (make_pass_cprop_hardreg): ...new function.
7460 (pass_reorder_blocks): Replace declaration with that of...
7461 (make_pass_reorder_blocks): ...new function.
7462 (pass_branch_target_load_optimize2): Replace declaration with that
7464 (make_pass_branch_target_load_optimize2): ...new function.
7465 (pass_leaf_regs): Replace declaration with that of...
7466 (make_pass_leaf_regs): ...new function.
7467 (pass_split_before_sched2): Replace declaration with that of...
7468 (make_pass_split_before_sched2): ...new function.
7469 (pass_compare_elim_after_reload): Replace declaration with that of...
7470 (make_pass_compare_elim_after_reload): ...new function.
7471 (pass_sched2): Replace declaration with that of...
7472 (make_pass_sched2): ...new function.
7473 (pass_stack_regs): Replace declaration with that of...
7474 (make_pass_stack_regs): ...new function.
7475 (pass_stack_regs_run): Replace declaration with that of...
7476 (make_pass_stack_regs_run): ...new function.
7477 (pass_df_finish): Replace declaration with that of...
7478 (make_pass_df_finish): ...new function.
7479 (pass_compute_alignments): Replace declaration with that of...
7480 (make_pass_compute_alignments): ...new function.
7481 (pass_duplicate_computed_gotos): Replace declaration with that of...
7482 (make_pass_duplicate_computed_gotos): ...new function.
7483 (pass_variable_tracking): Replace declaration with that of...
7484 (make_pass_variable_tracking): ...new function.
7485 (pass_free_cfg): Replace declaration with that of...
7486 (make_pass_free_cfg): ...new function.
7487 (pass_machine_reorg): Replace declaration with that of...
7488 (make_pass_machine_reorg): ...new function.
7489 (pass_cleanup_barriers): Replace declaration with that of...
7490 (make_pass_cleanup_barriers): ...new function.
7491 (pass_delay_slots): Replace declaration with that of...
7492 (make_pass_delay_slots): ...new function.
7493 (pass_split_for_shorten_branches): Replace declaration with that of...
7494 (make_pass_split_for_shorten_branches): ...new function.
7495 (pass_split_before_regstack): Replace declaration with that of...
7496 (make_pass_split_before_regstack): ...new function.
7497 (pass_convert_to_eh_region_ranges): Replace declaration with that
7499 (make_pass_convert_to_eh_region_ranges): ...new function.
7500 (pass_shorten_branches): Replace declaration with that of...
7501 (make_pass_shorten_branches): ...new function.
7502 (pass_set_nothrow_function_flags): Replace declaration with that of...
7503 (make_pass_set_nothrow_function_flags): ...new function.
7504 (pass_dwarf2_frame): Replace declaration with that of...
7505 (make_pass_dwarf2_frame): ...new function.
7506 (pass_final): Replace declaration with that of...
7507 (make_pass_final): ...new function.
7508 (pass_rtl_seqabstr): Replace declaration with that of...
7509 (make_pass_rtl_seqabstr): ...new function.
7510 (pass_release_ssa_names): Replace declaration with that of...
7511 (make_pass_release_ssa_names): ...new function.
7512 (pass_early_inline): Replace declaration with that of...
7513 (make_pass_early_inline): ...new function.
7514 (pass_inline_parameters): Replace declaration with that of...
7515 (make_pass_inline_parameters): ...new function.
7516 (pass_update_address_taken): Replace declaration with that of...
7517 (make_pass_update_address_taken): ...new function.
7518 (pass_convert_switch): Replace declaration with that of...
7519 (make_pass_convert_switch): ...new function.
7520 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
7521 to a subclass of simple_ipa_opt_pass along with...
7522 (pass_data_ipa_tree_profile): ...new pass_data instance and...
7523 (make_pass_ipa_tree_profile): ...new function.
7524 * tree-sra.c (pass_sra_early): Convert from a global struct to a
7525 subclass of gimple_opt_pass along with...
7526 (pass_data_sra_early): ...new pass_data instance and...
7527 (make_pass_sra_early): ...new function.
7528 (pass_sra): Convert from a global struct to a subclass of
7529 gimple_opt_pass along with...
7530 (pass_data_sra): ...new pass_data instance and...
7531 (make_pass_sra): ...new function.
7532 (pass_early_ipa_sra): Convert from a global struct to a subclass of
7533 gimple_opt_pass along with...
7534 (pass_data_early_ipa_sra): ...new pass_data instance and...
7535 (make_pass_early_ipa_sra): ...new function.
7536 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
7537 subclass of gimple_opt_pass along with...
7538 (pass_data_ccp): ...new pass_data instance and...
7539 (make_pass_ccp): ...new function.
7540 (pass_fold_builtins): Convert from a global struct to a subclass of
7541 gimple_opt_pass along with...
7542 (pass_data_fold_builtins): ...new pass_data instance and...
7543 (make_pass_fold_builtins): ...new function.
7544 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7545 subclass of gimple_opt_pass along with...
7546 (pass_data_copy_prop): ...new pass_data instance and...
7547 (make_pass_copy_prop): ...new function.
7548 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7549 global struct to a subclass of gimple_opt_pass along with...
7550 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7551 (make_pass_rename_ssa_copies): ...new function.
7552 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7553 subclass of gimple_opt_pass along with...
7554 (pass_data_dce): ...new pass_data instance and...
7555 (make_pass_dce): ...new function.
7556 (pass_dce_loop): Convert from a global struct to a subclass of
7557 gimple_opt_pass along with...
7558 (pass_data_dce_loop): ...new pass_data instance and...
7559 (make_pass_dce_loop): ...new function.
7560 (pass_cd_dce): Convert from a global struct to a subclass of
7561 gimple_opt_pass along with...
7562 (pass_data_cd_dce): ...new pass_data instance and...
7563 (make_pass_cd_dce): ...new function.
7564 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7565 subclass of gimple_opt_pass along with...
7566 (pass_data_dominator): ...new pass_data instance and...
7567 (make_pass_dominator): ...new function.
7568 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7569 gimple_opt_pass along with...
7570 (pass_data_phi_only_cprop): ...new pass_data instance and...
7571 (make_pass_phi_only_cprop): ...new function.
7572 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7573 subclass of gimple_opt_pass along with...
7574 (pass_data_dse): ...new pass_data instance and...
7575 (make_pass_dse): ...new function.
7576 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7577 a subclass of gimple_opt_pass along with...
7578 (pass_data_forwprop): ...new pass_data instance and...
7579 (make_pass_forwprop): ...new function.
7580 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7581 struct to a subclass of gimple_opt_pass along with...
7582 (pass_data_tree_ifcombine): ...new pass_data instance and...
7583 (make_pass_tree_ifcombine): ...new function.
7584 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7585 subclass of gimple_opt_pass along with...
7586 (pass_data_ch): ...new pass_data instance and...
7587 (make_pass_ch): ...new function.
7588 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7589 subclass of gimple_opt_pass along with...
7590 (pass_data_tree_loop): ...new pass_data instance and...
7591 (make_pass_tree_loop): ...new function.
7592 (pass_tree_loop_init): Convert from a global struct to a subclass of
7593 gimple_opt_pass along with...
7594 (pass_data_tree_loop_init): ...new pass_data instance and...
7595 (make_pass_tree_loop_init): ...new function.
7596 (pass_lim): Convert from a global struct to a subclass of
7597 gimple_opt_pass along with...
7598 (pass_data_lim): ...new pass_data instance and...
7599 (make_pass_lim): ...new function.
7600 (pass_tree_unswitch): Convert from a global struct to a subclass of
7601 gimple_opt_pass along with...
7602 (pass_data_tree_unswitch): ...new pass_data instance and...
7603 (make_pass_tree_unswitch): ...new function.
7604 (pass_predcom): Convert from a global struct to a subclass of
7605 gimple_opt_pass along with...
7606 (pass_data_predcom): ...new pass_data instance and...
7607 (make_pass_predcom): ...new function.
7608 (pass_vectorize): Convert from a global struct to a subclass of
7609 gimple_opt_pass along with...
7610 (pass_data_vectorize): ...new pass_data instance and...
7611 (make_pass_vectorize): ...new function.
7612 (pass_graphite): Convert from a global struct to a subclass of
7613 gimple_opt_pass along with...
7614 (pass_data_graphite): ...new pass_data instance and...
7615 (make_pass_graphite): ...new function.
7616 (pass_graphite_transforms): Convert from a global struct to a subclass
7617 of gimple_opt_pass along with...
7618 (pass_data_graphite_transforms): ...new pass_data instance and...
7619 (make_pass_graphite_transforms): ...new function.
7620 (pass_check_data_deps): Convert from a global struct to a subclass of
7621 gimple_opt_pass along with...
7622 (pass_data_check_data_deps): ...new pass_data instance and...
7623 (make_pass_check_data_deps): ...new function.
7624 (pass_iv_canon): Convert from a global struct to a subclass of
7625 gimple_opt_pass along with...
7626 (pass_data_iv_canon): ...new pass_data instance and...
7627 (make_pass_iv_canon): ...new function.
7628 (pass_scev_cprop): Convert from a global struct to a subclass of
7629 gimple_opt_pass along with...
7630 (pass_data_scev_cprop): ...new pass_data instance and...
7631 (make_pass_scev_cprop): ...new function.
7632 (pass_record_bounds): Convert from a global struct to a subclass of
7633 gimple_opt_pass along with...
7634 (pass_data_record_bounds): ...new pass_data instance and...
7635 (make_pass_record_bounds): ...new function.
7636 (pass_complete_unroll): Convert from a global struct to a subclass of
7637 gimple_opt_pass along with...
7638 (pass_data_complete_unroll): ...new pass_data instance and...
7639 (make_pass_complete_unroll): ...new function.
7640 (pass_complete_unrolli): Convert from a global struct to a subclass of
7641 gimple_opt_pass along with...
7642 (pass_data_complete_unrolli): ...new pass_data instance and...
7643 (make_pass_complete_unrolli): ...new function.
7644 (pass_parallelize_loops): Convert from a global struct to a subclass
7645 of gimple_opt_pass along with...
7646 (pass_data_parallelize_loops): ...new pass_data instance and...
7647 (make_pass_parallelize_loops): ...new function.
7648 (pass_loop_prefetch): Convert from a global struct to a subclass of
7649 gimple_opt_pass along with...
7650 (pass_data_loop_prefetch): ...new pass_data instance and...
7651 (make_pass_loop_prefetch): ...new function.
7652 (pass_iv_optimize): Convert from a global struct to a subclass of
7653 gimple_opt_pass along with...
7654 (pass_data_iv_optimize): ...new pass_data instance and...
7655 (make_pass_iv_optimize): ...new function.
7656 (pass_tree_loop_done): Convert from a global struct to a subclass of
7657 gimple_opt_pass along with...
7658 (pass_data_tree_loop_done): ...new pass_data instance and...
7659 (make_pass_tree_loop_done): ...new function.
7660 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7661 struct to a subclass of gimple_opt_pass along with...
7662 (pass_data_cse_reciprocals): ...new pass_data instance and...
7663 (make_pass_cse_reciprocals): ...new function.
7664 (pass_cse_sincos): Convert from a global struct to a subclass of
7665 gimple_opt_pass along with...
7666 (pass_data_cse_sincos): ...new pass_data instance and...
7667 (make_pass_cse_sincos): ...new function.
7668 (pass_optimize_bswap): Convert from a global struct to a subclass of
7669 gimple_opt_pass along with...
7670 (pass_data_optimize_bswap): ...new pass_data instance and...
7671 (make_pass_optimize_bswap): ...new function.
7672 (pass_optimize_widening_mul): Convert from a global struct to a
7673 subclass of gimple_opt_pass along with...
7674 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7675 (make_pass_optimize_widening_mul): ...new function.
7676 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7677 subclass of gimple_opt_pass along with...
7678 (pass_data_phiopt): ...new pass_data instance and...
7679 (make_pass_phiopt): ...new function.
7680 (pass_cselim): Convert from a global struct to a subclass of
7681 gimple_opt_pass along with...
7682 (pass_data_cselim): ...new pass_data instance and...
7683 (make_pass_cselim): ...new function.
7684 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7685 subclass of gimple_opt_pass along with...
7686 (pass_data_phiprop): ...new pass_data instance and...
7687 (make_pass_phiprop): ...new function.
7688 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7689 subclass of gimple_opt_pass along with...
7690 (pass_data_pre): ...new pass_data instance and...
7691 (make_pass_pre): ...new function.
7692 (pass_fre): Convert from a global struct to a subclass of
7693 gimple_opt_pass along with...
7694 (pass_data_fre): ...new pass_data instance and...
7695 (make_pass_fre): ...new function.
7696 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7697 subclass of gimple_opt_pass along with...
7698 (pass_data_reassoc): ...new pass_data instance and...
7699 (make_pass_reassoc): ...new function.
7700 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7701 subclass of gimple_opt_pass along with...
7702 (pass_data_sink_code): ...new pass_data instance and...
7703 (make_pass_sink_code): ...new function.
7704 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7705 subclass of gimple_opt_pass along with...
7706 (pass_data_strlen): ...new pass_data instance and...
7707 (make_pass_strlen): ...new function.
7708 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7709 struct to a subclass of gimple_opt_pass along with...
7710 (pass_data_build_alias): ...new pass_data instance and...
7711 (make_pass_build_alias): ...new function.
7712 (pass_build_ealias): Convert from a global struct to a subclass of
7713 gimple_opt_pass along with...
7714 (pass_data_build_ealias): ...new pass_data instance and...
7715 (make_pass_build_ealias): ...new function.
7716 (pass_ipa_pta): Convert from a global struct to a subclass of
7717 simple_ipa_opt_pass along with...
7718 (pass_data_ipa_pta): ...new pass_data instance and...
7719 (make_pass_ipa_pta): ...new function.
7720 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7721 subclass of gimple_opt_pass along with...
7722 (pass_data_uncprop): ...new pass_data instance and...
7723 (make_pass_uncprop): ...new function.
7724 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7725 global struct to a subclass of gimple_opt_pass along with...
7726 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7727 (make_pass_late_warn_uninitialized): ...new function.
7728 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7729 to a subclass of gimple_opt_pass along with...
7730 (pass_data_init_datastructures): ...new pass_data instance and...
7731 (make_pass_init_datastructures): ...new function.
7732 (pass_early_warn_uninitialized): Convert from a global struct to a
7733 subclass of gimple_opt_pass along with...
7734 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7735 (make_pass_early_warn_uninitialized): ...new function.
7736 (pass_update_address_taken): Convert from a global struct to a
7737 subclass of gimple_opt_pass along with...
7738 (pass_data_update_address_taken): ...new pass_data instance and...
7739 (make_pass_update_address_taken): ...new function.
7740 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7741 struct to a subclass of gimple_opt_pass along with...
7742 (pass_data_release_ssa_names): ...new pass_data instance and...
7743 (make_pass_release_ssa_names): ...new function.
7744 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7745 subclass of gimple_opt_pass along with...
7746 (pass_data_stdarg): ...new pass_data instance and...
7747 (make_pass_stdarg): ...new function.
7748 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7749 global struct to a subclass of gimple_opt_pass along with...
7750 (pass_data_convert_switch): ...new pass_data instance and...
7751 (make_pass_convert_switch): ...new function.
7752 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7753 to a subclass of gimple_opt_pass along with...
7754 (pass_data_tail_recursion): ...new pass_data instance and...
7755 (make_pass_tail_recursion): ...new function.
7756 (pass_tail_calls): Convert from a global struct to a subclass of
7757 gimple_opt_pass along with...
7758 (pass_data_tail_calls): ...new pass_data instance and...
7759 (make_pass_tail_calls): ...new function.
7760 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7761 struct to a subclass of gimple_opt_pass along with...
7762 (pass_data_lower_vector): ...new pass_data instance and...
7763 (make_pass_lower_vector): ...new function.
7764 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7765 gimple_opt_pass along with...
7766 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7767 (make_pass_lower_vector_ssa): ...new function.
7768 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7769 to a subclass of gimple_opt_pass along with...
7770 (pass_data_slp_vectorize): ...new pass_data instance and...
7771 (make_pass_slp_vectorize): ...new function.
7772 (pass_ipa_increase_alignment): Convert from a global struct to a
7773 subclass of simple_ipa_opt_pass along with...
7774 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7775 (make_pass_ipa_increase_alignment): ...new function.
7776 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7777 gimple_opt_pass along with...
7778 (pass_data_vrp): ...new pass_data instance and...
7779 (make_pass_vrp): ...new function.
7780 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7781 subclass of simple_ipa_opt_pass along with...
7782 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7783 (make_pass_ipa_free_lang_data): ...new function.
7784 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7785 gimple_opt_pass along with...
7786 (pass_data_tsan): ...new pass_data instance and...
7787 (make_pass_tsan): ...new function.
7788 (pass_tsan_O0): Convert from a global struct to a subclass of
7789 gimple_opt_pass along with...
7790 (pass_data_tsan_O0): ...new pass_data instance and...
7791 (make_pass_tsan_O0): ...new function.
7792 * var-tracking.c (pass_variable_tracking): Convert from a global
7793 struct to a subclass of rtl_opt_pass along with...
7794 (pass_data_variable_tracking): ...new pass_data instance and...
7795 (make_pass_variable_tracking): ...new function.
7796 * web.c (pass_web): Convert from a global struct to a subclass of
7797 rtl_opt_pass along with...
7798 (pass_data_web): ...new pass_data instance and...
7799 (make_pass_web): ...new function.
7800 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7801 declaration with that of...
7802 (make_pass_mode_switch_use): ...new function.
7803 (pass_resolve_sw_modes): Replace declaration with that of...
7804 (make_pass_resolve_sw_modes): ...new function.
7805 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7806 from a global struct to a subclass of rtl_opt_pass along with...
7807 (pass_data_mode_switch_use): ...new pass_data instance and...
7808 (make_pass_mode_switch_use): ...new function.
7809 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7810 from a global struct to a subclass of rtl_opt_pass along with...
7811 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7812 (make_pass_resolve_sw_modes): ...new function.
7813 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7814 struct to a subclass of rtl_opt_pass along with...
7815 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7816 (make_pass_insert_vzeroupper): ...new function.
7817 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7818 global struct to a subclass of rtl_opt_pass along with...
7819 (pass_data_work_around_errata): ...new pass_data instance and...
7820 (make_pass_work_around_errata): ...new function.
7821 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7822 struct to a subclass of rtl_opt_pass along with...
7823 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7824 (make_pass_mips_machine_reorg2): ...new function.
7826 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7828 * passes.c (pass_manager::operator new): New.
7830 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7832 Handwritten part of conversion of passes to C++ classes.
7834 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7835 (toplev.o): Add dep on PASS_MANAGER_H.
7836 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7837 of early local pases to reflect this moving from a global to a
7838 member of gcc::pass_manager.
7839 (cgraph_add_new_function): Likewise.
7840 * lto-cgraph.c (lto_output_node): Update for conversion of
7841 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7842 * passes.c (opt_pass::clone): New.
7843 (opt_pass::gate): New.
7844 (opt_pass::execute): New.
7845 (opt_pass::opt_pass): New.
7846 (pass_manager::execute_early_local_passes): New.
7847 (pass_manager::execute_pass_mode_switching): new.
7848 (finish_optimization_passes): Convert to...
7849 (pass_manager::finish_optimization_passes): ...this.
7850 (finish_optimization_passes): Update for conversion of passes to
7852 (register_dump_files_1): Use has_gate since we cannot portably
7853 check a vtable entry against NULL.
7854 (dump_one_pass): Likewise.
7855 (ipa_write_summaries_2): Likewise.
7856 (ipa_write_optimization_summaries_1): Likewise.
7857 (ipa_read_summaries_1): Likewise.
7858 (ipa_read_optimization_summaries_1): Likewise.
7859 (execute_ipa_stmt_fixups): Likewise.
7860 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7861 pass-creation functions rather than wiring up globals, and
7862 storing the results in fields of pass_manager generated using
7864 (pass_manager::dump_profile_report): Update for conversion of
7865 passes to C++ classes.
7866 (pass_manager::execute_ipa_summary_passes): Likewise.
7867 (execute_one_ipa_transform_pass): Likewise.
7868 (execute_one_pass): Use has_gate and has_execute since we cannot
7869 portably check a vtable entry against NULL.
7870 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7871 (pass_manager): Use pass-instances.def to add fields for the
7872 various pass instances.
7873 * toplev.c (finalize): Update for move of
7874 finish_optimization_passes to a method of gcc::pass_manager.
7875 * toplev.h (finish_optimization_passes): Move to method of class
7877 * tree-pass.h (struct pass_data): New.
7878 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7879 (opt_pass::gate): Convert to virtual function.
7880 (opt_pass::~opt_pass): New.
7881 (opt_pass::clone): New.
7882 (opt_pass::execute): Convert to virtual function.
7883 (opt_pass::opt_pass): New.
7884 (opt_pass::ctxt_): new.
7885 (gimple_opt_pass): Convert to subclass of opt_pass.
7886 (gimple_opt_pass::gimple_opt_pass): New.
7887 (rtl_opt_pass): Convert to subclass of opt_pass.
7888 (rtl_opt_pass::rtl_opt_pass): New.
7889 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7890 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7891 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7892 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7893 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7894 invocation of pass_mode_switching to reflect this moving from a
7895 global to a member of gcc::pass_manager.
7896 (ix86_option_override): Rework how pass_insert_vzeroupper is
7897 added to the pass_manager to reflect autogenerated changes.
7898 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7900 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7902 PR rtl-optimization/57708
7903 * recog.c (peep2_find_free_register): Validate all regs in a
7906 2013-08-05 Jan Hubicka <jh@suse.cz>
7909 * cgraph.c (verify_cgraph_node): Accept local flags from other
7911 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7912 (function_and_variable_visibility): Likewise.
7913 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7915 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7917 * graph.c (init_graph_slim_pretty_print): Remove.
7918 (print_graph_cfg): Do not call it. Use local pretty printer.
7919 (start_graph_dump): Likewise.
7921 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7923 * gimple-pretty-print.c (buffer): Remove.
7924 (initialized): Likewise.
7925 (maybe_init_pretty_print): Likewise.
7926 (print_gimple_stmt): Do not call it. Use non-static local
7927 pretty_printer variable.
7928 (print_gimple_expr): Likewise.
7929 (print_gimple_seq): Likewise.
7930 (gimple_dump_bb): Likewise.
7932 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7934 * asan.c (asan_pp): Remove.
7935 (asan_pp_initialized): Likewise.
7936 (asan_pp_initialize): Likewise.
7937 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7938 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7939 (asan_add_global): Likewise.
7941 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7943 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7944 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7945 * pretty-print.c (pp_formatted_text_data): Likewise.
7946 (pp_indent): Rename from pp_base_indent.
7947 (pp_format): Rename from pp_base_format.
7948 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7949 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7950 (pp_flush): Rename from pp_base_flush.
7951 (pp_set_line_maximum_length): Rename from
7952 pp_base_set_line_maximum_length.
7953 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7954 (pp_set_prefix): Rename from pp_base_set_prefix.
7955 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7956 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7957 (pp_append_text): Rename from pp_base_append_text.
7958 (pp_formatted_text): Rename from pp_base_formatted_text.
7959 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7960 (pp_remaining_character_count_for_line): Rename from
7961 pp_base_remaining_character_count_for_line.
7962 (pp_newline): Rename from pp_base_newline.
7963 (pp_character): Rename from pp_base_character.
7964 (pp_string): Rename from pp_base_string.
7965 (pp_maybe_space): Rename from pp_base_maybe_space.
7966 * asan.c (asan_pp_string): Adjust.
7967 (asan_emit_stack_protection): Likewise.
7968 (asan_add_global): Likewise.
7969 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7970 * tree-mudflap.c (mf_varname_tree): Likewise.
7971 * tree-pretty-print.c (pp_tree_identifier): Rename from
7972 pp_base_tree_identifier.
7973 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7974 Declare as function.
7976 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7978 * pretty-print.h (pp_bar_bar): New.
7979 (pp_ampersand_ampersand): Likewise.
7980 (pp_less_equal): Likewise.
7981 (pp_greater_equal): Likewise.
7982 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7983 printer functions instead of pp_string or operators and punctuators.
7984 (dump_gimple_call): Likewise.
7985 (dump_gimple_omp_for): Likewise.
7986 (dump_gimple_transaction): Likewise.
7987 (dump_gimple_phi): Likewise.
7988 (pp_gimple_stmt_1): Likewise.
7989 * sched-vis.c (print_insn): Likewise.
7990 * tree-mudflap.c (mf_varname_tree): Likewise.
7991 * tree-pretty-print.c (dump_block_node): Likewise.
7992 (dump_generic_node): Likewise.
7994 2013-08-02 Jan Hubicka <jh@suse.cz>
7996 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7998 * lto-streamer-out.c (tree_is_indexable): Results decls and
7999 parm decls are not indexable.
8000 (DFS_write_tree_body): Do not follow args and results.
8001 (hash_tree): Likewise.
8002 (output_functions): Rearrange so struct function is needed
8003 only when real body is output; be able to also ouptut abstract
8004 functions; output DECL_ARGUMENTS and DECL_RESULT.
8005 (lto_output): When not in WPA, ale store abstract functions.
8006 (write_symbol): Do not care about RESULT_DECL.
8007 (output_symbol_p): Handle correctly sbtract decls.
8008 * lto-streamer-in.c (input_function): Rearrange so struct
8009 function can be NULL at entry; allow streaming of
8010 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
8011 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
8012 sanity check during LTO.
8013 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
8014 RESULT_DECl and DECL_ARGUMENTS.
8015 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8018 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8020 * pretty-print.h (pp_underscore): New.
8022 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8023 printer functions instead of pp_character.
8024 (dump_binary_rhs): Likewise.
8025 (dump_ternary_rhs): Likewise.
8026 (dump_gimple_call_args): Likewise.
8027 (pp_points_to_solution): Likewise.
8028 (dump_gimple_call): Likewise.
8029 (dump_gimple_switch): Likewise.
8030 (dump_gimple_cond): Likewise.
8031 (dump_gimple_bind): Likewise.
8032 (dump_gimple_try): Likewise.
8033 (dump_gimple_omp_for): Likewise.
8034 (dump_gimple_omp_continue): Likewise.
8035 (dump_gimple_omp_single): Likewise.
8036 (dump_gimple_omp_sections): Likewise.
8037 (dump_gimple_omp_block): Likewise.
8038 (dump_gimple_omp_critical): Likewise.
8039 (dump_gimple_transaction): Likewise.
8040 (dump_gimple_asm): Likewise.
8041 (dump_gimple_phi): Likewise.
8042 (dump_gimple_omp_parallel): Likewise.
8043 (dump_gimple_omp_task): Likewise.
8044 (dump_gimple_omp_atomic_load): Likewise.
8045 (dump_gimple_omp_atomic_store): Likewise.
8046 (dump_gimple_mem_ops): Likewise.
8047 (pp_gimple_stmt_1): Likewise.
8048 (pp_cfg_jump): Likewise.
8049 (dump_implicit_edges): Likewise.
8050 (gimple_dump_bb_for_graph): Likewise.
8051 * graph.c (draw_cfg_node): Likewise.
8052 * langhooks.c (lhd_print_error_function): Likewise.
8053 * sched-vis.c (print_exp): Likewise.
8054 (print_value): Likewise.
8055 (print_pattern): Likewise.
8056 (print_insn): Likewise.
8057 (rtl_dump_bb_for_graph): Likewise.
8058 * tree-pretty-print.c (dump_function_declaration): Likewise.
8059 (dump_array_domain): Likewise.
8060 (dump_omp_clause): Likewise.
8061 (dump_location): Likewise.
8062 (dump_generic_node): Likewise.
8063 (print_struct_decl): Likewise.
8064 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8066 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8068 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8069 candidate table when replacing a candidate statement.
8070 (replace_rhs_if_not_dup): Likewise.
8071 (replace_one_candidate): Likewise.
8073 2013-08-02 Jan Hubicka <jh@suse.cz>
8074 Martin Liska <marxin.liska@gmail.com>
8076 * cgraphunit.c (add_new_function): Fix logic when adding from
8078 (assemble_thunk): Rename to ...
8079 (expand_thunk); .. this one; export; get it working with
8080 general functions; make produced gimple valid.
8081 * cgraph.h (expand_thunk): Declare.
8083 2013-08-02 Jan Hubicka <jh@suse.cz>
8085 * ipa-cp.c (gather_context_independent_values): Use
8086 ipa_get_param_move_cost.
8087 (get_replacement_map): Remove PARAM; move parameter folding
8089 (create_specialized_node): Update.
8090 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8091 assert that we have gimple body; update move_cost.
8092 (count_formal_params): Assert that we have gimple body.
8093 (ipa_dump_param): New function.
8094 (ipa_alloc_node_params): Break out from ...
8095 (ipa_initialize_node_params): ... here.
8096 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8097 (ipa_write_node_info): Stream move costs.
8098 (ipa_read_node_info): Read move costs.
8099 (ipa_update_after_lto_read): Do not recompute node params.
8100 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8101 (ipa_get_param): Check we are not in WPA.
8102 (ipa_get_param_move_cost): New.
8103 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8104 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8105 parm numbers to be present.
8107 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8109 PR rtl-optimization/58048
8110 * lra-constraints.c (process_alt_operands): Don't check asm
8111 operand on register.
8113 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8115 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8116 the implied StoreLoad barrier for atomic operations if before.
8118 2013-08-02 Jan Hubicka <jh@suse.cz>
8119 Martin Liska <marxin.liska@gmail.com>
8121 * cgraph.c (cgraph_function_body_availability): Do not check
8123 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
8124 symtab_node_availability): Declare.
8125 * ipa.c (can_replace_by_local_alias): New.
8126 (function_and_variable_visibility): Use it.
8127 * symtab.c (symtab_for_node_and_aliases,
8128 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
8130 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8132 PR rtl-optimization/57963
8133 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
8134 (lra_constraints): Use them.
8136 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
8138 * config/arm/types.md (define_attr "type"): Add "load_acq"
8140 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
8142 (cortex_a53_store1): Likewise.
8144 2013-08-01 Jan Hubicka <jh@suse.cz>
8146 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
8147 partitions are not needed.
8149 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
8151 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
8152 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
8153 MAYBE_NON_Q_CLASS_P where appropriate.
8155 2013-08-01 Jan Hubicka <jh@suse.cz>
8157 * cgraph.h (release_function_body): Declare.
8158 * tree.c (free_lang_data_in_decl): Free, parameters and return values
8159 of unused delcarations.
8161 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8163 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
8164 RTL form when subtracting a constant.
8166 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8168 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
8169 Generate canonical plus rtx with negated immediate instead of minus
8171 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
8173 2013-08-01 Jan Hubicka <jh@suse.cz>
8175 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
8176 (cgraph_release_function_body): Likewise.
8177 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
8178 * cgraph.h (cgrpah_node): Rename abstract_and_needed
8179 to used_as_abstract_origin.
8180 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
8181 symbols used as abstract origins.
8182 * cgraphunit.c (analyze_functions): Update.
8183 * ipa.c (symtab_remove_unreachable_nodes): Recompute
8184 used_as_abstract_origin.
8185 * tree-inline.c (tree_function_versioning): Update
8186 used_as_abstract_origin; be ready for DECL_RESULT and
8187 DECL_ARGUMENTS to be NULL.
8189 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
8190 for abstract functions.
8191 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
8194 2013-08-01 Jan Hubicka <jh@suse.cz>
8196 * profile.c (compute_value_histograms): Fix thinko.
8198 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8200 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
8201 aarch-common-protos.h to extra_headers.
8202 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8203 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
8204 * config/aarch64/t-aarch64 (aarch-common.o): Define.
8206 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8208 * config/aarch64/aarch64.md (define_attr "type"): Delete.
8209 Include "../arm/types.md". Define "type" attribute for all patterns.
8210 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
8213 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
8215 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
8216 to support power8 load fusion.
8217 (fusion_gpr_mem_load): Likewise.
8219 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
8221 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
8222 declarations for power8 load fusion.
8223 (emit_fusion_gpr_load): Likewise.
8225 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
8226 tuning for power8, turn on fusion mode by default. Turn on sign
8227 extending fusion mode if normal fusion mode is on, and we are at
8229 (fusion_gpr_load_p): New function, return true if we can fuse an
8230 addis instruction with a dependent load to a GPR.
8231 (emit_fusion_gpr_load): Emit the instructions for power8 load
8234 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
8235 (VSX load fusion peepholes): New peepholes to fuse together an
8236 addi instruction with a VSX load instruction.
8238 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
8239 peepholes to fuse an addis instruction with a load to a GPR base
8240 register. If we are supporting sign extending fusions, convert
8241 sign extending loads to zero extending loads and add an explicit
8244 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8246 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
8247 aarch-common-protos.h to extra_headers.
8248 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8249 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8250 (arm_early_store_addr_dep): Likewise.
8251 (arm_no_early_alu_shift_dep): Likewise.
8252 (arm_no_early_alu_shift_value_dep): Likewise.
8253 (arm_no_early_mul_dep): Likewise.
8254 (arm_no_early_store_addr_dep): Likewise.
8255 (arm_mac_accumulator_is_mul_result): Likewise.
8256 (arm_mac_accumulator_is_result): Likewise.
8257 * config/arm/aarch-common.c: ... here. New file.
8258 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8260 (arm_early_store_addr_dep): Likewise.
8261 (arm_no_early_alu_shift_dep): Likewise.
8262 (arm_no_early_alu_shift_value_dep): Likewise.
8263 (arm_no_early_mul_dep): Likewise.
8264 (arm_no_early_store_addr_dep): Likewise.
8265 (arm_mac_accumulator_is_mul_result): Likewise.
8266 (arm_mac_accumulator_is_result): Likewise.
8267 * config/arm/aarch-common-protos.h: ... here. New file.
8268 * config/arm/t-arm (aarch-common.o): Define.
8270 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8272 * config/arm/arm.md: Include new file "types.md".
8273 (define_attr "type"): Move from here to ...
8274 (define_attr "mul32"): Likewise.
8275 (define_attr "mul64"): Likewise.
8276 * config/arm/types.md: ... here. New file.
8278 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
8280 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
8281 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
8283 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8285 * gen-pass-instances.awk: Fix offset of substr().
8287 2013-07-31 David Malcolm <dmalcolm@redhat.com>
8289 * Makefile.in (pass-instances.def): New.
8290 (passes.o): Replace dependency on passes.def with one on
8293 * gen-pass-instances.awk: New.
8295 * passes.c (pass_manager::pass_manager): Use pass-instances.def
8296 rather than passes.def, updating local definition of NEXT_PASS
8297 macro to add an extra NUM parameter (currently unused).
8299 2013-07-30 David Malcolm <dmalcolm@redhat.com>
8301 * Makefile.in (PASS_MANAGER_H): New.
8302 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
8303 (passes.o): Likewise.
8304 (statistics.o): Likewise.
8305 (cgraphunit.o): Likewise.
8306 (context.o): Depend on PASS_MANAGER_H.
8308 * pass_manager.h: New.
8310 * cgraphunit.c (cgraph_add_new_function): Update for moves
8311 of globals to fields of pass_manager.
8312 (analyze_function): Likewise.
8313 (expand_function): Likewise.
8314 (ipa_passes): Likewise.
8315 (compile): Likewise.
8317 * context.c (context::context): New.
8318 * context.h (context::context): New.
8319 (context::get_passes): New.
8320 (context::passes_): New.
8322 * lto-cgraph.c (input_node): Update for moves of globals to
8323 fields of pass_manager.
8325 * passes.c (all_passes): Remove, in favor of a field of the
8326 same name within the new class pass_manager.
8327 (all_small_ipa_passes): Likewise.
8328 (all_lowering_passes): Likewise.
8329 (all_regular_ipa_passes): Likewise.
8330 (all_late_ipa_passes): Likewise.
8331 (all_lto_gen_passes): Likewise.
8332 (passes_by_id): Likewise.
8333 (passes_by_id_size): Likewise.
8334 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8335 the new class pass_manager.
8336 (set_pass_for_id): Convert to...
8337 (pass_manager::set_pass_for_id): ...method.
8338 (get_pass_for_id): Convert to...
8339 (pass_manager::get_pass_for_id): ...method.
8340 (register_one_dump_file): Move body of implementation into...
8341 (pass_manager::register_one_dump_file): ...here.
8342 (register_dump_files_1): Convert to...
8343 (pass_manager::register_dump_files_1): ...method.
8344 (register_dump_files): Convert to...
8345 (pass_manager::register_dump_files): ...method.
8346 (create_pass_tab): Update for moves of globals to fields of
8348 (dump_passes): Move body of implementation into...
8349 (pass_manager::dump_passes): ...here.
8350 (register_pass): Move body of implementation into...
8351 (pass_manager::register_pass): ...here.
8352 (init_optimization_passes): Convert into...
8353 (pass_manager::pass_manager): ...constructor for new
8354 pass_manager class, and initialize the pass_lists array.
8355 (check_profile_consistency): Update for moves of globals to
8356 fields of pass_manager.
8357 (dump_profile_report): Move body of implementation into...
8358 (pass_manager::dump_profile_report): ...here.
8359 (ipa_write_summaries_1): Update for moves of pass lists from
8360 being globals to fields of pass_manager.
8361 (ipa_write_optimization_summaries): Likewise.
8362 (ipa_read_summaries): Likewise.
8363 (ipa_read_optimization_summaries): Likewise.
8364 (execute_all_ipa_stmt_fixups): Likewise.
8366 * statistics.c (statistics_fini): Update for moves of globals to
8367 fields of pass_manager.
8369 * toplev.c (general_init): Replace call to
8370 init_optimization_passes with construction of the pass_manager
8373 * tree-pass.h (all_passes): Remove, in favor of a field of the
8374 same name within the new class pass_manager.
8375 (all_small_ipa_passes): Likewise.
8376 (all_lowering_passes): Likewise.
8377 (all_regular_ipa_passes): Likewise.
8378 (all_lto_gen_passes): Likewise.
8379 (all_late_ipa_passes): Likewise.
8380 (passes_by_id): Likewise.
8381 (passes_by_id_size): Likewise.
8382 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8383 the new class pass_manager.
8384 (get_pass_for_id): Remove.
8386 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8388 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
8391 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8393 * arm.md (mulhi3): New expand pattern.
8395 2013-07-30 Jan Hubicka <jh@suse.cz>
8396 Martin Liska <marxin.liska@gmail.com>
8398 * profile.c (compute_value_histograms): Do not ICE when
8399 there is mismatch only on some counters.
8401 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8403 PR rtl-optimization/57637
8404 * function.c (move_insn_for_shrink_wrap): Also check the
8405 GEN set of the LIVE problem for the liveness analysis
8406 if it exists, otherwise give up.
8408 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8410 PR tree-optimization/57993
8411 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
8412 replaced statement in the candidate table.
8413 (phi_add_costs): Return infinite cost when the hidden basis does
8414 not dominate all phis on which the candidate is dependent.
8415 (replace_one_candidate): Record replaced statement in the
8418 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
8420 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
8421 (ashlv2si3): New expander.
8422 (*ashlv2si3_i): New define_insn_and_split.
8423 * predicates.md (float_operation): Allow patterns with three
8426 PR rtl-optimization/58021
8427 * mode-switching.c (create_pre_exit): Always split off preceding
8428 insns if we are not at the basic block head.
8430 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
8432 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
8433 (UCLIBC_DYNAMIC_LINKER): New macro.
8434 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
8436 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
8437 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
8439 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
8440 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
8441 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
8442 for SF and DF modes. Use ieee_quad_format for TF mode.
8443 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
8444 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
8445 (mips_option_override): Handle `-mnan=legacy'.
8446 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
8447 `-mabs=2008' and `-mnan=2008'.
8448 (OPTION_DEFAULT_SPECS): Add "nan" default.
8449 (ASM_SPEC): Handle `-mnan='.
8450 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
8451 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
8452 comment accordingly.
8453 (neg<mode>2): Likewise.
8454 * config/mips/mips.opt (mabs, mnan): New options.
8455 * doc/install.texi (Configuration): Document `--with-nan=' option.
8456 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
8458 (MIPS Options): Document them.
8459 * config.gcc <mips*-*-*>: Handle `--with-nan='.
8460 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
8461 * configure: Regenerate.
8462 * config.in: Regenerate.
8464 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8466 * config/i386/i386.md (float post-reload splitters): Do not check
8467 for subregs of SSE registers.
8469 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8470 H.J. Lu <hongjiu.lu@intel.com>
8474 * config/i386/i386.md (post-reload splitter
8475 to avoid partial SSE reg dependency stalls): New pattern.
8477 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
8479 * config/s390/s390.md ("movcc"): Swap load and store instructions.
8481 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
8483 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
8484 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
8486 2013-07-26 Cary Coutant <ccoutant@google.com>
8488 * dwarf2out.c (die_checksum_ordered): Don't include template
8489 instantiations in signature.
8490 (is_template_parameter): New function.
8491 (is_template_instantiation): New function.
8492 (generate_skeleton_bottom_up): Don't include template instantiations
8494 (generate_skeleton): Likewise.
8495 (break_out_comdat_types): Move recursive call to break out nested
8497 (prune_unused_types_mark_generic_parms_dies): Call
8498 is_template_parameter.
8500 2013-07-26 Ian Bolton <ian.bolton@arm.com>
8502 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
8503 uses vector registers.
8504 * config/aarch64/iterators.md: Add attributes rtn and vas.
8506 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8507 Richard Earnshaw <richard.earnshaw@arm.com>
8509 * combine.c (simplify_comparison): Re-canonicalize operands
8511 * config/arm/arm.md (movcond_addsi): New splitter.
8513 2013-07-25 Sterling Augustine <saugustine@google.com>
8515 * dwarf2out.c (size_of_pubnames): Move code to...
8516 (include_pubname_in_output): ...here. New.
8517 (want_pubnames): Rearrange.
8518 (output_pubnames): Call include_pubname_in_output. Move assertion.
8520 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8522 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
8524 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8527 * doc/extend.texi: Remove obsolete builtins. Fix
8528 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
8530 2013-07-25 Jan Hubicka <jh@suse.cz>
8532 * cgraph.c (release_function_body): Break out from ...
8533 (cgraph_release_function_body): ... this one; also release DECL_RESULT
8535 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
8536 old_tree in the map.
8537 (create_specialized_node): Update.
8538 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8540 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8541 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8542 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8543 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8546 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8548 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8549 addsi3_carryin_alt2_<optab>): Correct output template.
8551 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8553 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8554 Adjust for arm_restrict_it.
8555 Remove trailing whitespace.
8557 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8559 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8560 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8562 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8564 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8566 PR rtl-optimization/57960
8567 * lra-constraints.c (process_alt_operands): Use the right mode
8568 when checking strict_low.
8570 2013-07-25 Jan Hubicka <jh@suse.cz>
8572 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8573 * cgraph.c (cgraph_remove_node): Do not release function body
8574 when in cgraph streaming.
8575 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8576 in other partitions are not considered reachable; fix handling of
8579 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8581 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8583 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8588 * config/arm/arm.md ("*sibcall_insn): Replace use of
8589 Ss with US. Adjust output for v5 and v4t.
8590 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8592 * config/arm/constraints.md ("Ss"): Rename to US.
8594 2013-07-25 Terry Guo <terry.guo@arm.com>
8596 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8597 shift_add/shift_sub0/shift_sub1 RTXs.
8599 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8600 Anton Blanchard <anton@au1.ibm.com>
8602 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8603 (altivec_vpks<VI_char>ss): Likewise.
8604 (altivec_vpks<VI_char>us): Likewise.
8605 (altivec_vpku<VI_char>us): Likewise.
8606 (altivec_vpku<VI_char>um): Likewise.
8608 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8610 Introduce context class.
8612 * Makefile.in (CONTEXT_H): New.
8613 (OBJS): Add context.o.
8614 (toplev.o): Add CONTEXT_H to dependencies.
8617 * toplev.c (general_init): Create the singleton gcc::context instance.
8623 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8625 PR rtl-optimization/57968
8626 * mode-switching.c (create_pre_exit): Allow instructions that
8627 don't set a return register to need a non-exit mode.
8629 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8630 Anton Blanchard <anton@au1.ibm.com>
8632 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8633 operands to vperm for little endian.
8634 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8635 of lvsl to create the control mask for a vperm for little endian.
8637 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8638 Anton Blanchard <anton@au1.ibm.com>
8640 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8641 two operands for little-endian.
8643 2013-07-23 Steve Ellcey <sellcey@mips.com>
8645 * config/mips/mips.c (mips_case_values_threshold): New.
8646 (TARGET_CASE_VALUES_THRESHOLD): Define.
8648 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8649 Anton Blanchard <anton@au1.ibm.com>
8651 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8652 selection of field for vector splat in little endian mode.
8654 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8656 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8657 expanders to rs6000.md.
8658 (ior<mode>3): Likewise.
8659 (and<mode>3): Likewise.
8660 (one_cmpl<mode>2): Likewise.
8661 (nor<mode>3): Likewise.
8662 (andc<mode>3): Likewise.
8663 (eqv<mode>3): Likewise.
8664 (nand<mode>3): Likewise.
8665 (orc<mode>3): Likewise.
8667 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8670 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8671 to split multi-word logical operations.
8672 (rs6000_split_logical_di): Likewise.
8673 (rs6000_split_logical): Likewise.
8675 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8676 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8677 and allow TImode operations in 32-bit.
8678 (vsx_and<mode>3_64bit): Likewise.
8679 (vsx_ior<mode>3_32bit): Likewise.
8680 (vsx_ior<mode>3_64bit): Likewise.
8681 (vsx_xor<mode>3_32bit): Likewise.
8682 (vsx_xor<mode>3_64bit): Likewise.
8683 (vsx_one_cmpl<mode>2_32bit): Likewise.
8684 (vsx_one_cmpl<mode>2_64bit): Likewise.
8685 (vsx_nor<mode>3_32bit): Likewise.
8686 (vsx_nor<mode>3_64bit): Likewise.
8687 (vsx_andc<mode>3_32bit): Likewise.
8688 (vsx_andc<mode>3_64bit): Likewise.
8689 (vsx_eqv<mode>3_32bit): Likewise.
8690 (vsx_eqv<mode>3_64bit): Likewise.
8691 (vsx_nand<mode>3_32bit): Likewise.
8692 (vsx_nand<mode>3_64bit): Likewise.
8693 (vsx_orc<mode>3_32bit): Likewise.
8694 (vsx_orc<mode>3_64bit): Likewise.
8696 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8697 logical types in GPRs.
8699 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8700 logical insns to rs6000.md, and allow TImode operations in
8702 (altivec_ior<mode>3): Likewise.
8703 (altivec_xor<mode>3): Likewise.
8704 (altivec_one_cmpl<mode>2): Likewise.
8705 (altivec_nor<mode>3): Likewise.
8706 (altivec_andc<mode>3): Likewise.
8708 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8709 attributes for moving the 128-bit logical operations into
8711 (BOOL_REGS_OUTPUT): Likewise.
8712 (BOOL_REGS_OP1): Likewise.
8713 (BOOL_REGS_OP2): Likewise.
8714 (BOOL_REGS_UNARY): Likewise.
8715 (BOOL_REGS_AND_CR0): Likewise.
8716 (one_cmpl<mode>2): Add support for DI logical operations on
8717 32-bit, splitting the operations to 32-bit.
8721 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8722 changes to combine the 32/64-bit code, allow logical operations on
8723 TI mode in 32-bit, and to use similar match_operator patterns like
8724 scalar mode uses. Combine the Altivec and VSX code for logical
8725 operations, and move it here.
8726 (ior<mode>3, 128-bit types): Likewise.
8727 (xor<mode>3, 128-bit types): Likewise.
8728 (one_cmpl<mode>3, 128-bit types): Likewise.
8729 (nor<mode>3, 128-bit types): Likewise.
8730 (andc<mode>3, 128-bit types): Likewise.
8731 (eqv<mode>3, 128-bit types): Likewise.
8732 (nand<mode>3, 128-bit types): Likewise.
8733 (orc<mode>3, 128-bit types): Likewise.
8734 (and<mode>3_internal): Likewise.
8735 (bool<mode>3_internal): Likewise.
8736 (boolc<mode>3_internal1): Likewise.
8737 (boolc<mode>3_internal2): Likewise.
8738 (boolcc<mode>3_internal1): Likewise.
8739 (boolcc<mode>3_internal2): Likewise.
8740 (eqv<mode>3_internal1): Likewise.
8741 (eqv<mode>3_internal2): Likewise.
8742 (one_cmpl1<mode>3_internal): Likewise.
8744 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8746 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8747 Rename flag_stack_usage to flag_stack_usage_info.
8749 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8751 * config/microblaze/sync.md: New file.
8752 * config/microblaze/microblaze.md: Include sync.md
8753 * config/microblaze/microblaze.c: Add print_operand 'y'.
8754 * config/microblaze/constraints.md: Add memory_contraint
8755 'Q' which is a single register.
8757 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8759 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8761 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8763 * reload.c (find_reloads): Exit loop once we find this operand
8764 cannot be reloaded somehow for this alternative.
8766 * reload.c (find_reloads): Exit loop once we find a hard register.
8768 * rtlanal.c (computed_jump_p): Exit loop once we find label
8771 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8773 * cfgloopmanip.c (remove_path): Exit loop after setting
8776 * gensupport.c (subst_dup): Avoid loop if code is not
8777 MATCH_DUP nor MATCH_OP_DUP.
8779 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8781 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8783 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8785 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8786 true for SP_REGNUM if mode == ptr_mode.
8787 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8788 with value R0_REGNUM + 31.
8790 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8792 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8793 pad pointer-typed argument downward.
8795 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8797 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8798 and __ILP32__ when the ILP32 model is in use.
8800 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8802 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8803 (aarch64_load_symref_appropriately): In the case of
8804 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8805 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8806 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8807 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8808 if 'mode' doesn't equal to 'ptr_mode'.
8809 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8810 'vcall_offset'; change to call aarch64_emit_move differently depending
8811 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8812 to calculate the upper bound of 'vcall_offset'.
8813 (aarch64_cannot_force_const_mem): Change to also return true if
8815 (aarch64_legitimize_reload_address): In the case of large
8816 displacements, add new local variable 'xmode' and an assertion
8817 based on it; change to use 'xmode' to generate the new rtx and
8819 (aarch64_asm_trampoline_template): Change to generate the template
8820 differently depending on TARGET_ILP32 or not; change to use
8821 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8822 (aarch64_trampoline_size): Removed.
8823 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8824 and replace immediate literals with it. Change to use 'ptr_mode'
8825 instead of 'DImode' and call convert_memory_address if the mode
8826 of 'fnaddr' doesn't equal to 'ptr_mode'.
8827 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8829 (aarch64_elf_asm_destructor): Likewise.
8830 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8831 on TARGET_ILP32 instead of aarch64_trampoline_size.
8832 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8833 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8834 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8835 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8836 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8837 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8838 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8839 depending on the value of 'mode'.
8840 (add_losym_<mode>): New.
8841 (ldr_got_small_<mode>): New, based on ldr_got_small.
8842 (ldr_got_small): Remove.
8843 (ldr_got_small_sidi): New.
8844 * config/aarch64/iterators.md (P): New.
8845 (PTR): Change to 'ptr_mode' in the condition.
8847 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8849 * config.gcc (aarch64*-*-*): Support --with-abi.
8850 (aarch64*-*-elf): Support --with-multilib-list.
8851 (aarch64*-*-linux*): Likewise.
8852 (supported_defaults): Add abi to aarch64*-*-*.
8853 * configure.ac: Mention AArch64 for --with-multilib-list.
8854 * configure: Re-generated.
8855 * config/aarch64/biarchilp32.h: New file.
8856 * config/aarch64/biarchlp64.h: New file.
8857 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8859 (MULTILIB_DEFAULTS): Ditto.
8860 (DRIVER_SELF_SPECS): Ditto.
8861 (ASM_SPEC): Update to also substitute -mabi.
8862 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8863 file whose name depends on -mabi= and -mbig-endian.
8864 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8866 (POINTER_SIZE): New define.
8867 (POINTERS_EXTEND_UNSIGNED): Ditto.
8868 (enum aarch64_abi_type): New enumeration tag.
8869 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8870 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8871 (TARGET_ILP32): New define.
8872 * config/aarch64/aarch64.opt (mabi): New.
8874 (ilp32, lp64): New values for -mabi.
8875 * config/aarch64/t-aarch64 (comma): New define.
8876 (MULTILIB_OPTIONS): Ditto.
8877 (MULTILIB_DIRNAMES): Ditto.
8878 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8879 * doc/invoke.texi: Document -mabi for AArch64.
8881 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8883 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8885 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8886 Anton Blanchard <anton@au1.ibm.com>
8888 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8889 endianness when selecting field to splat.
8891 2013-07-22 Eric Christopher <echristo@gmail.com>
8893 * dwarf2out.c (die_odr_checksum): New function to use
8894 CHECKSUM_ macros and ULEB128 for DIE tag.
8895 (generate_type_signature): Use.
8897 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8899 * config.gcc (sparc*-*-*): Accept leon3 processor.
8900 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8901 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8902 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8903 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8904 (mfix-ut699): Adjust comment.
8905 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8906 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8907 (CPP_CPU_SPEC): Likewise.
8908 (ASM_CPU_SPEC): Likewise.
8909 * config/sparc/sparc.c (leon3_cost): New constant.
8910 (sparc_option_override): Add leon3 support.
8911 (mem_ref): New function.
8912 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8913 (sparc_do_work_around_errata): Look into the instruction in the delay
8914 slot and adjust accordingly. Add fix for the data cache nullify issues
8915 of the UT699. Change insertion position for the NOP.
8916 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8917 (leon3_load): New reservation.
8918 (leon_store): Bump latency to 2.
8919 (grfpu): New automaton.
8920 (grfpu_alu): New unit.
8921 (grfpu_ds): Likewise.
8922 (leon_fp_alu): Adjust.
8923 (leon_fp_mult): Delete.
8924 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8925 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8926 * config/sparc/sparc.md (cpu): Add leon3.
8927 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8929 (atomic_test_and_set): Likewise.
8932 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8934 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8935 default for R5900 targets.
8936 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8937 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8938 * config/mips/mips.c (mips_option_override): Report an error for
8939 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8940 for -march=r5900 -mhard-float.
8942 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8944 * df-problems.c (can_move_insns_across): Exit loop once we
8945 find a non-fixed, non-global register.
8947 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8950 * omega.c (omega_eliminate_red): Break after setting red_found.
8951 (omega_problem_has_red_equations): Similarly after setting found.
8952 (omega_query_variable): Similarly after setting coupled.
8954 2013-07-22 Marek Polacek <polacek@redhat.com>
8956 * gimplify.c: Don't include gimple.h twice.
8958 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8960 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8961 instead of TARGET_THUMB1.
8962 (Pz): New constraint.
8963 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8965 (compare_negsi_si): Likewise.
8966 (compare_addsi2_op0): Likewise.
8967 (compare_addsi2_op1): Likewise.
8968 (addsi3_carryin_<optab>): Likewise.
8969 (addsi3_carryin_alt2_<optab>): Likewise.
8970 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8971 for arm_restrict_it.
8972 (subsi3_carryin): Likewise.
8973 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8974 (minmax_arithsi): Disable for arm_restrict_it.
8975 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8976 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8977 (satsi_<SAT:code>_shift): Likewise.
8978 (arm_shiftsi3): Add alternative for 16-bit encoding.
8979 (arm32_movhf): Disable for arm_restrict_it.
8980 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8981 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8983 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8985 * config/arm/arm.md (attribute "insn"): Delete.
8986 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8987 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8988 (not_shiftsi): Update for attribute change.
8989 (not_shiftsi_compare0): Likewise.
8990 (not_shiftsi_compare0_scratch): Likewise.
8991 (arm_one_cmplsi2): Likewise.
8992 (thumb1_one_cmplsi2): Likewise.
8993 (notsi_compare0): Likewise.
8994 (notsi_compare0_scratch): Likewise.
8995 (thumb1_movdi_insn): Likewise.
8996 (arm_movsi_insn): Likewise.
8997 (movhi_insn_arch4): Likewise.
8998 (movhi_bytes): Likewise.
8999 (arm_movqi_insn): Likewise.
9000 (thumb1_movqi_insn): Likewise.
9001 (arm32_movhf): Likewise.
9002 (thumb1_movhf): Likewise.
9003 (arm_movsf_soft_insn): Likewise.
9004 (thumb1_movsf_insn): Likewise.
9005 (thumb_movdf_insn): Likewise.
9006 (movsicc_insn): Likewise.
9007 (movsfcc_soft_insn): Likewise.
9008 (and_scc): Likewise.
9009 (cond_move): Likewise.
9010 (if_move_not): Likewise.
9011 (if_not_move): Likewise.
9012 (if_shift_move): Likewise.
9013 (if_move_shift): Likewise.
9014 (if_shift_shift): Likewise.
9015 (if_not_arith): Likewise.
9016 (if_arith_not): Likewise.
9017 (cond_move_not): Likewise.
9018 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9019 (neon_mov<mode>): Likewise.
9020 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9021 (thumb2_movsi_vfp): Likewise.
9022 (movsf_vfp): Likewise.
9023 (thumb2_movsf_vfp): Likewise.
9024 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9026 (cortexa7_older_only): Likewise.
9027 (cortexa7_younger): Likewise.
9028 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9029 (1020alu_shift_op): Likewise.
9030 (1020alu_shift_reg_op): Likewise.
9031 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9032 (alu_shift_op): Likewise.
9033 (alu_shift_reg_op): Likewise.
9034 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9035 (11_alu_shift_op): Likewise.
9036 (11_alu_shift_reg_op): Likewise.
9037 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9038 (9_alu_shift_reg_op): Likewise.
9039 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9041 (cortex_a15_alu_shift): Likewise.
9042 (cortex_a15_alu_shift_reg): Likewise.
9043 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9044 (cortex_a5_alu_shift): Likewise.
9045 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9047 (cortex_a53_alu_shift): Likewise.
9048 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9050 (cortex_a7_alu_reg): Likewise.
9051 (cortex_a7_alu_shift): Likewise.
9052 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9053 (cortex_a8_alu_shift): Likewise.
9054 (cortex_a8_alu_shift_reg): Likewise.
9055 (cortex_a8_mov): Likewise.
9056 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9057 (cortex_a9_dp_shift): Likewise.
9058 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9059 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9060 (cortex_r4_mov): Likewise.
9061 (cortex_r4_alu_shift): Likewise.
9062 (cortex_r4_alu_shift_reg): Likewise.
9063 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9064 (526_alu_shift_op): Likewise.
9065 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9066 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9067 (626te_alu_shift_op): Likewise.
9068 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9069 (726te_alu_op): Likewise.
9070 (726te_alu_shift_op): Likewise.
9071 (726te_alu_shift_reg_op): Likewise.
9072 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9073 (mp626_alu_shift_op): Likewise.
9074 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9075 (pj4_alu_e1_conds): Likewise.
9076 (pj4_alu): Likewise.
9077 (pj4_alu_conds): Likewise.
9078 (pj4_shift): Likewise.
9079 (pj4_shift_conds): Likewise.
9080 (pj4_alu_shift): Likewise.
9081 (pj4_alu_shift_conds): Likewise.
9083 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9085 * config/arm/predicates.md (shiftable_operator_strict_it):
9087 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9088 Disable cond_exec version for arm_restrict_it.
9089 (thumb2_smaxsi3): Convert to generate cond_exec.
9090 (thumb2_sminsi3): Likewise.
9091 (thumb32_umaxsi3): Likewise.
9092 (thumb2_uminsi3): Likewise.
9093 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9094 (thumb2_neg_abssi2): Likewise.
9095 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9096 (thumb2_movsicc_insn): Adjust alternatives.
9097 (thumb2_mov_negscc): Disable for arm_restrict_it.
9098 (thumb2_mov_negscc_strict_it): New pattern.
9099 (thumb2_mov_notscc_strict_it): New pattern.
9100 (thumb2_mov_notscc): Disable for arm_restrict_it.
9101 (thumb2_ior_scc): Likewise.
9102 (thumb2_ior_scc_strict_it): New pattern.
9103 (thumb2_cond_move): Adjust for arm_restrict_it.
9104 (thumb2_cond_arith): Disable for arm_restrict_it.
9105 (thumb2_cond_arith_strict_it): New pattern.
9106 (thumb2_cond_sub): Adjust for arm_restrict_it.
9107 (thumb2_movcond): Likewise.
9108 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9109 (thumb2_zero_extendhisi2_v6): Likewise.
9110 (thumb2_zero_extendqisi2_v6): Likewise.
9111 (orsi_notsi_si): Likewise.
9112 (orsi_not_shiftsi_si): Likewise.
9114 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9116 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9117 instruction sequence is 1 byte shorter.
9119 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9121 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9122 it is not needed after split.
9124 2013-07-20 Iain Sandoe <iain@codesourcery.com>
9127 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
9128 second label for nonlocal goto receivers. Don't output pic base labels
9129 unless we're producing PIC; mark that action unreachable().
9130 (ix86_save_reg): If the function contains a nonlocal label, save the
9132 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
9133 * config/darwin.c (emitted_pic_label_num): New GTY.
9134 (update_pic_label_number_if_needed): New.
9135 (machopic_output_function_base_name): Adjust for nonlocal receiver
9137 (machopic_should_output_picbase_label): New.
9138 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
9139 (nonlocal_goto_receiver): New insn and split.
9141 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
9143 * config/aarch64/aarch64-builtins.c
9144 (aarch64_fold_builtin): Fold abs in all modes.
9145 * config/aarch64/aarch64-simd-builtins.def
9146 (abs): Enable for all modes.
9147 * config/aarch64/arm_neon.h
9148 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
9149 (vabs_f64): Add missing intrinsic.
9151 2013-07-19 Ian Bolton <ian.bolton@arm.com>
9153 * config/aarch64/arm_neon.h (vabs_s64): New function
9155 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
9158 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
9159 * config/avr/avr.md (adjust_len): Add `round'.
9160 * config/avr/avr-protos.h (avr_out_round): New prototype.
9161 (avr_out_plus): Add `out_label' argument.
9162 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
9163 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
9164 Handle the case where `insn' is just a pattern.
9165 (avr_out_bitop): Handle the case where `insn' is just a pattern.
9166 (avr_out_round): New function.
9167 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
9169 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
9171 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9172 Add check for flag_stack_usage to handle -fstack-usage support
9174 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
9176 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
9177 interaction for new Power8 flags and VSX.
9179 2013-07-18 Sriraman Tallam <tmsriram@google.com>
9182 * tree-inline.c (expand_call_inline): Emit errors during
9183 early_inlining only if optimization is not turned on.
9185 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9189 * passes.c (init_optimization_passes): Move the construction of
9190 the pass hierarchy into a new passes.def file.
9192 * Makefile.in (passes.o): Add dependency on passes.def.
9194 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9196 * passes.c (init_optimization_passes): Introduce macros for
9197 constructing the tree of passes (INSERT_PASSES_AFTER,
9198 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
9199 TERMINATE_PASS_LIST).
9201 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
9202 Wei Mi <wmi@google.com>
9204 PR rtl-optimization/57878
9205 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
9207 (reload_pseudo_compare_func): Check nregs first for reload
9210 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9212 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
9214 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
9216 * read-rtl.c (validate_const_int): Once an invalid character is
9217 seen, quit the loop.
9219 * gengtype.c (write_roots): Similarly once we find the "deletable"
9220 or "if_marked" option.
9222 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9224 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
9225 "xtab" and "sat". Move value "clz" from here to ...
9226 (attriubte "type"): ... here.
9227 (satsi_<SAT:code>): Delete "insn" attribute.
9228 (satsi_<SAT:code>_shift): Likewise.
9229 (arm_zero_extendqisi2addsi): Likewise.
9230 (arm_extendqisi2addsi): Likewise.
9231 (clzsi2): Update for attribute changes.
9232 (rbitsi2): Likewise.
9233 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
9235 (arm_usatsihi): Likewise.
9236 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9238 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9240 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
9241 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
9242 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
9243 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
9244 in alphabetical order.
9245 (attribute "core_cycles"): Update for attribute changes.
9246 (arm_addsi3): Likewise.
9247 (addsi3_compare0): Likewise.
9248 (addsi3_compare0_scratch): Likewise.
9249 (addsi3_compare_op1): Likewise.
9250 (addsi3_compare_op2): Likewise.
9251 (compare_addsi2_op0): Likewise.
9252 (compare_addsi2_op1): Likewise.
9253 (addsi3_carryin_shift_<optab>): Likewise.
9254 (subsi3_carryin_shift): Likewise.
9255 (rsbsi3_carryin_shift): Likewise.
9256 (arm_subsi3_insn): Likewise.
9257 (subsi3_compare0): Likewise.
9258 (subsi3_compare): Likewise.
9259 (arm_andsi3_insn): Likewise.
9260 (thumb1_andsi3_insn): Likewise.
9261 (andsi3_compare0): Likewise.
9262 (andsi3_compare0_scratch): Likewise.
9263 (zeroextractsi_compare0_scratch
9264 (andsi_not_shiftsi_si): Likewise.
9265 (iorsi3_insn): Likewise.
9266 (iorsi3_compare0): Likewise.
9267 (iorsi3_compare0_scratch): Likewise.
9268 (arm_xorsi3): Likewise.
9269 (thumb1_xorsi3_insn): Likewise.
9270 (xorsi3_compare0): Likewise.
9271 (xorsi3_compare0_scratch): Likewise.
9272 (satsi_<SAT:code>_shift): Likewise.
9274 (arm_shiftsi3): Likewise.
9275 (shiftsi3_compare0): Likewise.
9276 (not_shiftsi): Likewise.
9277 (not_shiftsi_compare0): Likewise.
9278 (not_shiftsi_compare0_scratch): Likewise.
9279 (arm_one_cmplsi2): Likewise.
9280 (thumb_one_complsi2): Likewise.
9281 (notsi_compare0): Likewise.
9282 (notsi_compare0_scratch): Likewise.
9283 (thumb1_zero_extendhisi2): Likewise.
9284 (arm_zero_extendhisi2): Likewise.
9285 (arm_zero_extendhisi2_v6): Likewise.
9286 (arm_zero_extendhisi2addsi): Likewise.
9287 (thumb1_zero_extendqisi2): Likewise.
9288 (thumb1_zero_extendqisi2_v6): Likewise.
9289 (arm_zero_extendqisi2): Likewise.
9290 (arm_zero_extendqisi2_v6): Likewise.
9291 (arm_zero_extendqisi2addsi): Likewise.
9292 (thumb1_extendhisi2): Likewise.
9293 (arm_extendhisi2): Likewise.
9294 (arm_extendhisi2_v6): Likewise.
9295 (arm_extendqisi): Likewise.
9296 (arm_extendqisi_v6): Likewise.
9297 (arm_extendqisi2addsi): Likewise.
9298 (thumb1_extendqisi2): Likewise.
9299 (thumb1_movdi_insn): Likewise.
9300 (arm_movsi_insn): Likewise.
9301 (movsi_compare0): Likewise.
9302 (movhi_insn_arch4): Likewise.
9303 (movhi_bytes): Likewise.
9304 (arm_movqi_insn): Likewise.
9305 (thumb1_movqi_insn): Likewise.
9306 (arm32_movhf): Likewise.
9307 (thumb1_movhf): Likewise.
9308 (arm_movsf_soft_insn): Likewise.
9309 (thumb1_movsf_insn): Likewise.
9310 (movdf_soft_insn): Likewise.
9311 (thumb_movdf_insn): Likewise.
9312 (arm_cmpsi_insn): Likewise.
9313 (cmpsi_shiftsi): Likewise.
9314 (cmpsi_shiftsi_swp): Likewise.
9315 (arm_cmpsi_negshiftsi_si): Likewise.
9316 (movsicc_insn): Likewise.
9317 (movsfcc_soft_insn): Likewise.
9318 (arith_shiftsi): Likewise.
9319 (arith_shiftsi_compare0
9320 (arith_shiftsi_compare0_scratch
9321 (sub_shiftsi): Likewise.
9322 (sub_shiftsi_compare0
9323 (sub_shiftsi_compare0_scratch
9324 (and_scc): Likewise.
9325 (cond_move): Likewise.
9326 (if_plus_move): Likewise.
9327 (if_move_plus): Likewise.
9328 (if_move_not): Likewise.
9329 (if_not_move): Likewise.
9330 (if_shift_move): Likewise.
9331 (if_move_shift): Likewise.
9332 (if_shift_shift): Likewise.
9333 (if_not_arith): Likewise.
9334 (if_arith_not): Likewise.
9335 (cond_move_not): Likewise.
9336 (thumb1_ashlsi3): Set type attribute.
9337 (thumb1_ashrsi3): Likewise.
9338 (thumb1_lshrsi3): Likewise.
9339 (thumb1_rotrsi3): Likewise.
9340 (shiftsi3_compare0_scratch): Likewise.
9341 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
9342 (neon_mov<mode>): Likewise.
9343 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
9345 (thumb2_movsi_insn): Likewise.
9346 (thumb2_cmpsi_neg_shiftsi): Likewise.
9347 (thumb2_extendqisi_v6): Likewise.
9348 (thumb2_zero_extendhisi2_v6): Likewise.
9349 (thumb2_zero_extendqisi2_v6): Likewise.
9350 (thumb2_shiftsi3_short): Likewise.
9351 (thumb2_addsi3_compare0_scratch): Likewise.
9352 (orsi_not_shiftsi_si): Likewise.
9353 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9354 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
9356 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9357 (1020alu_shift_op): Likewise.
9358 (1020alu_shift_reg_op): Likewise.
9359 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9360 (alu_shift_op): Likewise.
9361 (alu_shift_reg_op): Likewise.
9362 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9363 (11_alu_shift_op): Likewise.
9364 (11_alu_shift_reg_op): Likewise.
9365 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9366 (9_alu_shift_reg_op): Likewise.
9367 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9369 (cortex_a15_alu_shift): Likewise.
9370 (cortex_a15_alu_shift_reg): Likewise.
9371 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
9373 (cortex_a5_alu_shift): Likewise.
9374 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
9376 (cortex_a53_alu_shift): Likewise.
9377 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9379 (cortex_a7_alu_reg): Likewise.
9380 (cortex_a7_alu_shift): Likewise.
9381 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
9383 (cortex_a8_alu_shift): Likewise.
9384 (cortex_a8_alu_shift_reg): Likewise.
9385 (cortex_a8_mov): Likewise.
9386 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9387 (cortex_a9_dp_shift): Likewise.
9388 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
9390 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
9392 (cortex_r4_mov): Likewise.
9393 (cortex_r4_alu_shift): Likewise.
9394 (cortex_r4_alu_shift_reg): Likewise.
9395 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9396 (526_alu_shift_op): Likewise.
9397 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9398 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9399 (626te_alu_shift_op): Likewise.
9400 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
9401 (726te_alu_op): Likewise.
9402 (726te_alu_shift_op): Likewise.
9403 (726te_alu_shift_reg_op): Likewise.
9404 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9405 (mp626_alu_shift_op): Likewise.
9406 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
9407 (pj4_alu_e1_conds): Likewise.
9408 (pj4_alu): Likewise.
9409 (pj4_alu_conds): Likewise.
9410 (pj4_shift): Likewise.
9411 (pj4_shift_conds): Likewise.
9412 (pj4_alu_shift): Likewise.
9413 (pj4_alu_shift_conds): Likewise.
9414 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9416 (cortexa7_older_only): Likewise.
9417 (cortexa7_younger): Likewise.
9419 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9421 * ipa-pure-const.c (generate_summary): Rename to...
9422 (pure_const_generate_summary): ... this.
9424 2013-07-17 Iain Sandoe <iain@codesourcery.com>
9426 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
9428 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
9431 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
9434 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9436 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
9437 enabled without -march=zEC12.
9438 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
9441 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9443 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
9444 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
9445 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
9446 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
9447 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
9448 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
9449 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
9451 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
9452 and ISA_HAS_NMADD3_NMSUB3.
9453 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
9454 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
9455 (nmsub4<mode>, nmsub3<mode>): Likewise.
9456 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
9458 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9460 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
9461 TARGET_MIPS5400 checking.
9463 2013-07-16 Jakub Jelinek <jakub@redhat.com>
9464 Peter Bergner <bergner@vnet.ibm.com>
9466 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
9467 registers in the comment.
9468 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
9469 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
9470 rather than FIRST_PSEUDO_REGISTERS.
9472 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
9474 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
9475 enable extra ISA flags with TARGET_HTM.
9477 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9479 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
9482 2013-07-15 Cong Hou <congh@google.com>
9484 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
9485 in compare function for sorting.
9487 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
9489 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
9490 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
9491 * config/rs6000/rs6000.opt: Add -mhtm option.
9492 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
9493 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
9494 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9495 __HTM__ if the HTM instructions are available.
9496 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
9497 htm_spr_reg_operand): New define_predicates.
9498 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
9499 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
9501 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
9502 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
9503 HTM builtin functions.
9504 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
9505 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
9506 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
9507 (rs6000_builtin_mask_calculate): Likewise.
9508 (rs6000_option_override_internal): Likewise.
9509 (bdesc_htm): Add new HTM builtin support.
9510 (htm_spr_num): New function.
9511 (htm_spr_regno): Likewise.
9512 (rs6000_htm_spr_icode): Likewise.
9513 (htm_expand_builtin): Likewise.
9514 (htm_init_builtins): Likewise.
9515 (rs6000_expand_builtin): Add support for HTM builtin functions.
9516 (rs6000_init_builtins): Likewise.
9517 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
9519 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
9520 (TARGET_HTM, MASK_HTM): Define macros.
9521 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
9522 (FIXED_REGISTERS): Likewise.
9523 (CALL_USED_REGISTERS): Likewise.
9524 (CALL_REALLY_USED_REGISTERS): Likewise.
9525 (REG_ALLOC_ORDER): Likewise.
9526 (enum reg_class): Likewise.
9527 (REG_CLASS_NAMES): Likewise.
9528 (REG_CLASS_CONTENTS): Likewise.
9529 (REGISTER_NAMES): Likewise.
9530 (ADDITIONAL_REGISTER_NAMES): Likewise.
9531 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
9532 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
9533 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
9534 * config/rs6000/htm.md: New file.
9535 * config/rs6000/htmintrin.h: New file.
9536 * config/rs6000/htmxlintrin.h: New file.
9538 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9540 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9541 Define SYMBOL_TINY_GOT, update comment.
9542 * config/aarch64/aarch64.c
9543 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9544 (aarch64_expand_mov_immediate): Likewise.
9545 (aarch64_print_operand): Likewise.
9546 (aarch64_classify_symbol): Likewise.
9547 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9548 (ldr_got_tiny): Define.
9550 2013-07-13 Tobias Grosser <tobias@grosser.es>
9552 PR tree-optimization/54094
9553 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9554 scheduling dimension for the parallelism check from the polyhedral
9555 information in the AST.
9556 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9559 2013-07-13 Jason Merrill <jason@redhat.com>
9561 * print-tree.c (debug_vec_tree): Use debug_raw.
9562 (debug_raw (vec<tree, va_gc> &)): New.
9563 (debug_raw (vec<tree, va_gc> *)): New.
9564 * tree.h: Declare them.
9566 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9568 * ifcvt.c (ifcvt_after_combine): New static variable.
9569 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9571 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9572 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9573 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9575 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9577 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9579 2013-07-12 Michael Matz <matz@suse.de>
9582 * convert.c (convert_to_real): Reject non-float inner types.
9584 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9586 * config/aarch64/aarch64-protos.h
9587 (aarch64_simd_immediate_valid_for_move): Remove.
9588 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9589 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9591 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9593 2013-07-11 Steve Ellcey <sellcey@mips.com>
9595 * config/mips/mips.c (mips_conditional_register_usage): Do not
9596 use t[0-7] registers in MIPS16 mode when optimizing for size.
9598 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9600 * config/i386/i386.c (dispatch_function_versions): Fix array
9601 indexing of function_version_info to match actual_versions.
9603 2013-07-11 Teresa Johnson <tejohnson@google.com>
9605 * vec.h (struct va_gc): Move release out-of-line.
9606 (va_gc::release): Call ggc_free on released vec.
9608 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9610 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9611 Require GOT register as additional operand in UNSPEC.
9612 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9613 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9614 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9615 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9616 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9617 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9618 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9620 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9623 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9624 name seen by assembler/linker rather if available.
9626 2013-07-11 Andreas Schwab <schwab@suse.de>
9628 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9630 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9632 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9634 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9636 * doc/tm.texi.in: Move hook documentation to ....
9637 * target.def: ... here.
9639 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9641 * doc/tm.texi: Regenerate.
9643 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9646 * doc/invoke.texi: Document -Wconditionally-supported.
9648 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9651 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9654 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9657 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9658 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9659 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9660 Remove duplicate devices.
9661 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9662 * config/avr/t-multilib: Regenerate.
9663 * config/avr/avr-tables.opt: Regenerate.
9664 * doc/avr-mmcu.texi: Regenerate.
9666 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9669 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9671 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9673 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9674 the cost of MULT when optimizing for size.
9676 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9678 * config/cr16/cr16-protos.h: Don't include target.h.
9680 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9682 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9683 adjust register size for TDmode and TFmode for VSX registers.
9685 2013-07-08 Kai Tietz <ktietz@redhat.com>
9688 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9689 hook_bool_const_tree_true.
9691 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9693 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9694 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9695 * config/s390/s390.md: Define FPR*_REGNUM constants.
9696 Fix FPR2_REGNUM constant (18 -> 17).
9697 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9698 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9699 ("trunc<BFP:mode><DFP_ALL:mode>2")
9700 ("trunc<DFP_ALL:mode><BFP:mode>2")
9701 ("*extend<BFP:mode><DFP_ALL:mode>2")
9702 ("*extend<DFP_ALL:mode><BFP:mode>2")
9703 ("extend<BFP:mode><DFP_ALL:mode>2")
9704 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9707 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9709 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9711 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9713 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9714 and cfun_fpr_bit_p to cfun_fpr_save_p.
9715 (s390_frame_area, s390_register_info, s390_frame_info)
9716 (s390_emit_prologue, s390_emit_epilogue)
9717 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9719 * config/s390/s390.h: Define *_REGNUM macros for floating point
9722 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9724 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9726 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9728 PR rtl-optimization/57786
9729 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9730 and break out of the loop when it is set to false.
9732 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9735 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9736 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9737 (const_int 63)) 0)).
9738 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9739 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9740 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9742 PR rtl-optimization/57829
9743 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9744 mask bits outside of mode are just sign-extension from mode to HWI.
9746 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9748 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9749 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9750 adjust_address instead of change_address to keep info about alignment.
9751 (emit_strmov): Remove.
9752 (emit_memmov): New function.
9753 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9754 (expand_movmem_epilogue): Likewise and return updated rtx for
9756 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9757 destination and source.
9758 (decide_alignment): Refactor, handle vector_loop.
9759 (ix86_expand_movmem): Likewise.
9760 (ix86_expand_setmem): Likewise.
9761 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9763 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9765 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9766 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9768 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9770 * config/i386/sse.md (sse_movlhps): Change alternative 3
9771 of operand 2 to "m".
9773 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9776 * config/i386/sse.md (iptr): New mode attribute.
9777 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9778 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9779 (<sse>_vmmul<mode>3): Ditto.
9780 (<sse>_vmdiv<mode>3): Ditto.
9781 (sse_vmrcpv4sf2): Ditto.
9782 (<sse>_vmsqrt<mode>2): Ditto.
9783 (sse_vmrsqrtv4sf2): Ditto.
9784 (<sse>_vm<code><mode>3): Ditto.
9785 (avx_vmcmp<mode>3): Ditto.
9786 (<sse>_vmmaskcmp<mode>3): Ditto.
9787 (<sse>_comi): Ditto.
9788 (<sse>_ucomi): Ditto.
9789 (*xop_vmfrcz_<mode>): Ditto.
9790 (*fmai_fmadd_<mode>): Ditto.
9791 (*fmai_fmsub_<mode>): Ditto.
9792 (*fmai_fnmadd_<mode>): Ditto.
9793 (*fmai_fnmsub_<mode>): Ditto.
9794 (*fma4i_vmfmadd_<mode>): Ditto.
9795 (*fma4i_vmfmsub_<mode>): Ditto.
9796 (*fma4i_vmfnmadd_<mode>): Ditto.
9797 (*fma4i_vmfnmsub_<mode>): Ditto.
9798 (*xop_vmfrcz_<mode>): Ditto.
9799 (sse_cvtps2pi): Ditto.
9800 (sse_cvttps2pi): Ditto.
9801 (sse_cvtss2si): Ditto.
9802 (sse_cvtss2si_2): Ditto.
9803 (sse_cvtss2siq_2): Ditto.
9804 (sse_cvttss2si): Ditto.
9805 (sse_cvttss2siq): Ditto.
9806 (sse_cvtsd2si): Ditto.
9807 (sse_cvtsd2si_2): Ditto.
9808 (sse_cvtsd2siq_2): Ditto.
9809 (sse_cvttsd2si): Ditto.
9810 (sse_cvttsd2siq): Ditto.
9811 (sse_cvtsd2ss): Ditto.
9812 (sse_cvtss2sd): Ditto.
9813 (avx2_pbroadcast<mode>): Ditto.
9814 (avx2_pbroadcast<mode>_1): Ditto.
9815 (*avx_vperm_broadcast_v4sf): Ditto.
9817 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9818 (sse_movlhps): Ditto.
9819 (sse_storehps): Ditto.
9820 (sse_loadhps): Ditto.
9821 (sse_storelps): Ditto.
9822 (sse_loadlps): Ditto.
9823 (*vec_concatv4sf): Ditto.
9824 (*vec_interleave_highv2df): Ditto.
9825 (*vec_interleave_lowv2df): Ditto.
9826 (*vec_extractv2df_1_sse): Ditto.
9827 (*vec_extractv2df_0_sse): Ditto.
9828 (sse2_storelpd): Ditto.
9829 (sse2_loadlpd): Ditto.
9830 (sse2_movsd): Ditto.
9831 (*vec_concatv4si): Ditto.
9832 (vec_concatv2di): Ditto.
9834 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9835 for Intel asm dialect.
9836 (mmx_punpcklwd): Ditto.
9837 (mmx_punpckldq): Ditto.
9839 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9840 for intel assembler dialect.
9842 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9845 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9846 for BUILT_IN_C{LZ,LRSB}*.
9847 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9848 * tree-vrp.c (extract_range_basic): Handle
9849 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9850 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9851 fall thru to code calling set_value*.
9852 * builtins.c (expand_builtin): Remove *IMAX cases.
9853 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9854 if width is bigger than 2*HWI.
9856 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9858 PR rtl-optimization/55342
9859 * lra-int.h (lra_subreg_reload_pseudos): New.
9860 * lra.c: Add undoing optional reloads to the block diagram.
9861 (lra_subreg_reload_pseudos): New.
9862 (lra_optional_reload_pseudos): Change comments.
9863 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9864 lra_optional_reload_pseudos after undo transformations.
9865 * lra-assigns.c (pseudo_prefix_title): New.
9866 (lra_setup_reg_renumber): Use it.
9867 (spill_for): Ditto. Check subreg reload pseudos too.
9868 (assign_by_spills): Consider subreg reload pseudos too.
9869 * lra-constraints.c (simplify_operand_subreg): Use
9870 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9871 (curr_insn_transform): Recognize and do optional reloads.
9872 (undo_optional_reloads): New.
9873 (lra_undo_inheritance): Call undo_optional_reloads.
9875 2013-07-05 Thomas Quinot <quinot@adacore.com>
9877 * tree-complex.c (expand_complex_operations_1): Fix typo.
9879 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9881 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9882 (tune_params): New member 'const vec_costs'.
9883 * config/aarch64/aarch64.c (generic_vector_cost): New.
9884 (generic_tunings): New member 'generic_vector_cost'.
9885 (aarch64_builtin_vectorization_cost): New.
9886 (aarch64_add_stmt_cost): New.
9887 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9888 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9890 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9893 * config/i386/predicates.md (vsib_address_operand): Disallow
9894 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9896 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9899 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9900 expand_builtin_setjmp_receiver.
9901 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9902 with NULL for the label parameter.
9903 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9904 the frame-pointer. Adjust comments.
9905 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9906 only if LABEL is non-NULL.
9908 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9910 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9911 (ARM_ABI_AAPCS64): Ditto.
9913 (ARM_DEFAULT_ABI): Ditto.
9915 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9917 * config/aarch64/aarch64-builtins.c
9918 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9919 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9921 * config/aarch64/aarch64-simd.md
9922 (aarch64_ld1<VALL:mode>): New.
9923 (aarch64_st1<VALL:mode>): Likewise.
9924 * config/aarch64/arm_neon.h
9925 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9927 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9929 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9931 (ix86_option_override_internal):Turn on all -mavx target flags by
9932 default as they are dependent on AVX anyway.
9934 2013-07-02 Cary Coutant <ccoutant@google.com>
9936 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9938 (loc_checksum_ordered): Likewise.
9939 (hash_loc_operands): Remove inline keyword.
9941 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9943 PR tree-optimization/57741
9944 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9945 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9946 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9947 Allow REAL_CST step_exprs if flag_associative_math.
9948 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9950 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9952 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9954 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9956 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9958 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9960 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9962 (iorsi3_insn): Likewise.
9963 (arm_xorsi3): Likewise.
9965 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9967 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9969 (attribute "type"): ... here, and prefix with "wmmx_".
9970 (attribute "core_cycles"): Update for attribute changes.
9971 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9972 (tbcstv4hi): Likewise.
9973 (tbcstv2si): Likewise.
9974 (iwmmxt_iordi3): Likewise.
9975 (iwmmxt_xordi3): Likewise.
9976 (iwmmxt_anddi3): Likewise.
9977 (iwmmxt_nanddi3): Likewise.
9978 (iwmmxt_arm_movdi): Likewise.
9979 (iwmmxt_movsi_insn): Likewise.
9980 (mov<mode>_internal): Likewise.
9981 (and<mode>3_iwmmxt): Likewise.
9982 (ior<mode>3_iwmmxt): Likewise.
9983 (xor<mode>3_iwmmxt): Likewise.
9984 (add<mode>3_iwmmxt): Likewise.
9985 (ssaddv8qi3): Likewise.
9986 (ssaddv4hi3): Likewise.
9987 (ssaddv2si3): Likewise.
9988 (usaddv8qi3): Likewise.
9989 (usaddv4hi3): Likewise.
9990 (usaddv2si3): Likewise.
9991 (sub<mode>3_iwmmxt): Likewise.
9992 (sssubv8qi3): Likewise.
9993 (sssubv4hi3): Likewise.
9994 (sssubv2si3): Likewise.
9995 (ussubv8qi3): Likewise.
9996 (ussubv4hi3): Likewise.
9997 (ussubv2si3): Likewise.
9998 (mulv4hi3_iwmmxt): Likewise.
9999 (smulv4hi3_highpart): Likewise.
10000 (umulv4hi3_highpart): Likewise.
10001 (iwmmxt_wmacs): Likewise.
10002 (iwmmxt_wmacsz): Likewise.
10003 (iwmmxt_wmacu): Likewise.
10004 (iwmmxt_wmacuz): Likewise.
10005 (iwmmxt_clrdi): Likewise.
10006 (iwmmxt_clrv8qi): Likewise.
10007 (iwmmxt_clr4hi): Likewise.
10008 (iwmmxt_clr2si): Likewise.
10009 (iwmmxt_uavgrndv8qi3): Likewise.
10010 (iwmmxt_uavgrndv4hi3): Likewise.
10011 (iwmmxt_uavgv8qi3): Likewise.
10012 (iwmmxt_uavgv4hi3): Likewise.
10013 (iwmmxt_tinsrb): Likewise.
10014 (iwmmxt_tinsrh): Likewise.
10015 (iwmmxt_tinsrw): Likewise.
10016 (iwmmxt_textrmub): Likewise.
10017 (iwmmxt_textrmsb): Likewise.
10018 (iwmmxt_textrmuh): Likewise.
10019 (iwmmxt_textrmsh): Likewise.
10020 (iwmmxt_textrmw): Likewise.
10021 (iwmxxt_wshufh): Likewise.
10022 (eqv8qi3): Likewise.
10023 (eqv4hi3): Likewise.
10024 (eqv2si3): Likewise.
10025 (gtuv8qi3): Likewise.
10026 (gtuv4hi3): Likewise.
10027 (gtuv2si3): Likewise.
10028 (gtv8qi3): Likewise.
10029 (gtv4hi3): Likewise.
10030 (gtv2si3): Likewise.
10031 (smax<mode>3_iwmmxt): Likewise.
10032 (umax<mode>3_iwmmxt): Likewise.
10033 (smin<mode>3_iwmmxt): Likewise.
10034 (umin<mode>3_iwmmxt): Likewise.
10035 (iwmmxt_wpackhss): Likewise.
10036 (iwmmxt_wpackwss): Likewise.
10037 (iwmmxt_wpackdss): Likewise.
10038 (iwmmxt_wpackhus): Likewise.
10039 (iwmmxt_wpackwus): Likewise.
10040 (iwmmxt_wpackdus): Likewise.
10041 (iwmmxt_wunpckihb): Likewise.
10042 (iwmmxt_wunpckihh): Likewise.
10043 (iwmmxt_wunpckihw): Likewise.
10044 (iwmmxt_wunpckilb): Likewise.
10045 (iwmmxt_wunpckilh): Likewise.
10046 (iwmmxt_wunpckilw): Likewise.
10047 (iwmmxt_wunpckehub): Likewise.
10048 (iwmmxt_wunpckehuh): Likewise.
10049 (iwmmxt_wunpckehuw): Likewise.
10050 (iwmmxt_wunpckehsb): Likewise.
10051 (iwmmxt_wunpckehsh): Likewise.
10052 (iwmmxt_wunpckehsw): Likewise.
10053 (iwmmxt_wunpckelub): Likewise.
10054 (iwmmxt_wunpckeluh): Likewise.
10055 (iwmmxt_wunpckeluw): Likewise.
10056 (iwmmxt_wunpckelsb): Likewise.
10057 (iwmmxt_wunpckelsh): Likewise.
10058 (iwmmxt_wunpckelsw): Likewise.
10059 (ror<mode>3): Likewise.
10060 (ashr<mode>3_iwmmxt): Likewise.
10061 (lshr<mode>3_iwmmxt): Likewise.
10062 (ashl<mode>3_iwmmxt): Likewise.
10063 (ror<mode>3_di): Likewise.
10064 (ashr<mode>3_di): Likewise.
10065 (lshr<mode>3_di): Likewise.
10066 (ashl<mode>3_di): Likewise.
10067 (iwmmxt_wmadds): Likewise.
10068 (iwmmxt_wmaddu): Likewise.
10069 (iwmmxt_tmia): Likewise.
10070 (iwmmxt_tmiaph): Likewise.
10071 (iwmmxt_tmiabb): Likewise.
10072 (iwmmxt_tmiatb): Likewise.
10073 (iwmmxt_tmiabt): Likewise.
10074 (iwmmxt_tmiatt): Likewise.
10075 (iwmmxt_tmovmskb): Likewise.
10076 (iwmmxt_tmovmskh): Likewise.
10077 (iwmmxt_tmovmskw): Likewise.
10078 (iwmmxt_waccb): Likewise.
10079 (iwmmxt_wacch): Likewise.
10080 (iwmmxt_waccw): Likewise.
10081 (iwmmxt_waligni): Likewise.
10082 (iwmmxt_walignr): Likewise.
10083 (iwmmxt_walignr0): Likewise.
10084 (iwmmxt_walignr1): Likewise.
10085 (iwmmxt_walignr2): Likewise.
10086 (iwmmxt_walignr3): Likewise.
10087 (iwmmxt_wsadb): Likewise.
10088 (iwmmxt_wsadh): Likewise.
10089 (iwmmxt_wsadbz): Likewise.
10090 (iwmmxt_wsadhz): Likewise.
10091 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10092 (iwmmxt_wabsdiffb): Likewise.
10093 (iwmmxt_wabsdiffh): Likewise.
10094 (iwmmxt_wabsdiffw): Likewise.
10095 (iwmmxt_waddsubhx): Likewise
10096 (iwmmxt_wsubaddhx): Likewise.
10097 (addc<mode>3): Likewise.
10098 (iwmmxt_avg4): Likewise.
10099 (iwmmxt_avg4r): Likewise.
10100 (iwmmxt_wmaddsx): Likewise.
10101 (iwmmxt_wmaddux): Likewise.
10102 (iwmmxt_wmaddsn): Likewise.
10103 (iwmmxt_wmaddun): Likewise.
10104 (iwmmxt_wmulwsm): Likewise.
10105 (iwmmxt_wmulwum): Likewise.
10106 (iwmmxt_wmulsmr): Likewise.
10107 (iwmmxt_wmulumr): Likewise.
10108 (iwmmxt_wmulwsmr): Likewise.
10109 (iwmmxt_wmulwumr): Likewise.
10110 (iwmmxt_wmulwl): Likewise.
10111 (iwmmxt_wqmulm): Likewise.
10112 (iwmmxt_wqmulwm): Likewise.
10113 (iwmmxt_wqmulmr): Likewise.
10114 (iwmmxt_wqmulwmr): Likewise.
10115 (iwmmxt_waddbhusm): Likewise.
10116 (iwmmxt_waddbhusl): Likewise.
10117 (iwmmxt_wqmiabb): Likewise.
10118 (iwmmxt_wqmiabt): Likewise.
10119 (iwmmxt_wqmiatb): Likewise.
10120 (iwmmxt_wqmiatt): Likewise.
10121 (iwmmxt_wqmiabbn): Likewise.
10122 (iwmmxt_wqmiabtn): Likewise.
10123 (iwmmxt_wqmiatbn): Likewise.
10124 (iwmmxt_wqmiattn): Likewise.
10125 (iwmmxt_wmiabb): Likewise.
10126 (iwmmxt_wmiabt): Likewise.
10127 (iwmmxt_wmiatb): Likewise.
10128 (iwmmxt_wmiatt): Likewise.
10129 (iwmmxt_wmiabbn): Likewise.
10130 (iwmmxt_wmiabtn): Likewise.
10131 (iwmmxt_wmiatbn): Likewise.
10132 (iwmmxt_wmiattn): Likewise.
10133 (iwmmxt_wmiawbb): Likewise.
10134 (iwmmxt_wmiawbt): Likewise.
10135 (iwmmxt_wmiawtb): Likewise.
10136 (iwmmxt_wmiawtt): Likewise.
10137 (iwmmxt_wmiawbbn): Likewise.
10138 (iwmmxt_wmiawbtn): Likewise.
10139 (iwmmxt_wmiawtbn): Likewise.
10140 (iwmmxt_wmiawttn): Likewise.
10141 (iwmmxt_wmerge): Likewise.
10142 (iwmmxt_tandc<mode>3): Likewise.
10143 (iwmmxt_torc<mode>3): Likewise.
10144 (iwmmxt_torvsc<mode>3): Likewise.
10145 (iwmmxt_textrc<mode>3): Likewise.
10146 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
10147 (wmmxt_pack): Likewise.
10148 (wmmxt_mult_c1): Likewise.
10149 (wmmxt_mult_c2): Likewise.
10150 (wmmxt_alu_c1): Likewise.
10151 (wmmxt_alu_c2): Likewise.
10152 (wmmxt_alu_c3): Likewise.
10153 (wmmxt_transfer_c1): Likewise.
10154 (wmmxt_transfer_c2): Likewise.
10155 (wmmxt_transfer_c3): Likewise.
10156 (marvell_f_iwmmxt_wstr): Likewise.
10157 (marvell_f_iwmmxt_wldr): Likewise.
10159 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
10161 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
10163 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10166 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10167 * lra-constraints.c (need_for_split_p): Check call used hard regs
10168 living through calls.
10170 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
10171 call used regs for call insn.
10173 2013-06-28 Jakub Jelinek <jakub@redhat.com>
10176 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
10177 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
10179 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10181 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
10183 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10185 * lra-constraints.c (need_for_split_p): Check call used hard regs
10186 living through calls.
10188 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10191 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
10192 to tie with any other modes. Eliminate Altivec vector mode tests,
10193 since these are a subset of ALTIVEC or VSX vector modes. Simplify
10194 code, to return 0 if testing MODE2 for a condition, if we've
10195 already tested MODE1 for the same condition.
10197 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10199 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
10202 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10204 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10205 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
10207 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10209 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
10211 (aarch64_symbolic_constant_p): Remove.
10212 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
10213 static. Fix line length and white space.
10214 (aarch64_symbolic_constant_p): Remove.
10215 * config/aarch64/predicates.md (aarch64_valid_symref):
10216 Use aarch64_classify_symbol_expression.
10218 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10220 * config/arm/constraints.md (Ts): New constraint.
10221 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
10223 (compare_scc): Use "Ts" constraint for operand 0.
10224 (ior_scc_scc): Likewise.
10225 (and_scc_scc): Likewise.
10226 (and_scc_scc_nodom): Likewise.
10227 (ior_scc_scc_cmp): Likewise for operand 7.
10228 (and_scc_scc_cmp): Likewise.
10229 * config/arm/thumb2.md (thumb2_movsi_insn):
10230 Add alternatives for 16-bit encodings.
10231 (thumb2_movhi_insn): Likewise.
10232 (thumb2_movsicc_insn): Likewise.
10233 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
10234 (thumb2_negscc): Use "Ts" constraint.
10235 Move mvn instruction outside cond_exec block.
10236 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
10237 for 16-bit encodings.
10239 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10241 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
10243 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
10244 (mulsi3subsi): Likewise.
10245 (mulsidi3adddi): Likewise.
10246 (mulsidi3_v6): Likewise.
10247 (umulsidi3_v6): Likewise.
10248 (umulsidi3adddi_v6): Likewise.
10249 (smulsi3_highpart_v6): Likewise.
10250 (umulsi3_highpart_v6): Likewise.
10251 (mulhisi3tb): Likewise.
10252 (mulhisi3bt): Likewise.
10253 (mulhisi3tt): Likewise.
10254 (maddhisi4): Likewise.
10255 (maddhisi4tb): Likewise.
10256 (maddhisi4tt): Likewise.
10257 (maddhidi4): Likewise.
10258 (maddhidi4tb): Likewise.
10259 (maddhidi4tt): Likewise.
10260 (zeroextractsi_compare0_scratch): Likewise.
10261 (insv_zero): Likewise.
10262 (insv_t2): Likewise.
10263 (anddi_notzesidi_di): Likewise.
10264 (anddi_notsesidi_di): Likewise.
10265 (andsi_notsi_si): Likewise.
10266 (iordi_zesidi_di): Likewise.
10267 (xordi_zesidi_di): Likewise.
10268 (andsi_iorsi3_notsi): Likewise.
10269 (smax_0): Likewise.
10270 (smax_m1): Likewise.
10271 (smin_0): Likewise.
10272 (not_shiftsi): Likewise.
10273 (unaligned_loadsi): Likewise.
10274 (unaligned_loadhis): Likewise.
10275 (unaligned_loadhiu): Likewise.
10276 (unaligned_storesi): Likewise.
10277 (unaligned_storehi): Likewise.
10278 (extv_reg): Likewise.
10279 (extzv_t2): Likewise.
10280 (divsi3): Likewise.
10281 (udivsi3): Likewise.
10282 (arm_zero_extendhisi2addsi): Likewise.
10283 (arm_zero_extendqisi2addsi): Likewise.
10284 (compareqi_eq0): Likewise.
10285 (arm_extendhisi2_v6): Likewise.
10286 (arm_extendqisi2addsi): Likewise.
10287 (arm_movt): Likewise.
10288 (thumb2_ldrd): Likewise.
10289 (thumb2_ldrd_base): Likewise.
10290 (thumb2_ldrd_base_neg): Likewise.
10291 (thumb2_strd): Likewise.
10292 (thumb2_strd_base): Likewise.
10293 (thumb2_strd_base_neg): Likewise.
10294 (arm_negsi2): Add alternative for 16-bit encoding.
10295 (arm_one_cmplsi2): Likewise.
10297 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10299 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
10300 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
10301 (movdfcc): Likewise.
10302 * config/arm/vfp.md (*thumb2_movsf_vfp):
10303 Disable predication for arm_restrict_it.
10304 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
10305 (*thumb2_movdfcc_vfp): Likewise.
10306 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
10307 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
10308 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
10309 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
10310 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
10311 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
10312 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
10313 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
10314 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
10315 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
10316 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
10317 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
10318 Disable predication for arm_restrict_it.
10320 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
10322 * config/i386/bmiintrin.h (_bextr_u32): New.
10323 (_bextr_u64): Ditto.
10325 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
10327 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
10328 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
10329 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
10330 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
10331 * config/mips/n32-elf.h: ...this new file.
10333 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
10336 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
10337 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
10339 2013-06-27 Catherine Moore <clm@codesourcery.com>
10341 * config/mips/mips-tables.opt: Regenerate.
10342 * config/mips/mips-cpus.def: Add m14ke and m14kec.
10343 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
10344 * doc/invoke.texi: Add -m14kc.
10346 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10349 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
10350 constraints of operand 1 and 2.
10353 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
10354 to match RTL canonicalization. Swap predicates and
10355 constraints of operand 1 and 2.
10357 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
10359 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
10360 Process OP_INOUT regs for splitting too.
10362 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10364 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
10365 decl before the loop, initialize to NULL.
10366 (vectorizable_load): Initialize ptr_incr to NULL.
10368 2013-06-27 Martin Jambor <mjambor@suse.cz>
10371 * ipa-ref.h (ipa_maybe_record_reference): Declare.
10372 * ipa-ref.c (ipa_maybe_record_reference): New function.
10373 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
10374 * ipa-cp.c (create_specialized_node): Record potential references from
10376 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
10378 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10380 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
10381 parameter 'mode' of type 'enum machine_mode mode'; change to pass
10382 'mode' to force_reg.
10383 (aarch64_add_offset): Update calls to aarch64_force_temporary.
10384 (aarch64_expand_mov_immediate): Likewise.
10386 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10388 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
10389 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
10391 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10393 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
10394 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10395 (struct machine_function): Add tbegin_p.
10396 (s390_canonicalize_comparison): Fold CC mode compares to
10397 conditional jump if possible.
10398 (s390_emit_jump): Return the emitted jump.
10399 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
10400 Handle CCRAWmode compares.
10401 (s390_option_override): Default to -mhtm if available.
10402 (s390_reg_clobbered_rtx): Handle floating point regs as well.
10403 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
10404 FPRs instead of df_regs_ever_live_p.
10405 (s390_optimize_nonescaping_tx): New function.
10406 (s390_init_frame_layout): Extend clobbered_regs array to cover
10408 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
10409 (s390_expand_tbegin): New function.
10410 (enum s390_builtin): New enum definition.
10411 (code_for_builtin): New array definition.
10412 (s390_init_builtins): New function.
10413 (s390_expand_builtin): New function.
10414 (TARGET_INIT_BUILTINS): Define.
10415 (TARGET_EXPAND_BUILTIN): Define.
10416 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
10417 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
10418 (s390_alc_comparison): Likewise.
10419 * config/s390/s390-modes.def: Add CCRAWmode.
10420 * config/s390/s390.h (processor_flags): Add PF_TX.
10421 (TARGET_CPU_HTM): Define macro.
10422 (TARGET_HTM): Define macro.
10423 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
10424 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
10425 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10426 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
10427 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
10428 (TBEGIN_MASK, TBEGINC_MASK): New constants.
10429 ("*cc_to_int"): Move up.
10430 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
10431 constants other than 0.
10432 ("*ccraw_to_int"): New insn and splitter definition.
10433 ("tbegin", "tbegin_nofloat", "tbegin_retry")
10434 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
10435 ("tx_assist"): New expander.
10436 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
10437 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
10438 * config/s390/s390.opt: Add -mhtm option.
10439 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
10440 * config/s390/htmxlintrin.h: New file.
10441 * config/s390/htmintrin.h: New file.
10442 * config/s390/s390intrin.h: New file.
10443 * doc/extend.texi: Document htm builtins.
10444 * config.gcc: Add the new header files to extra_headers.
10446 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
10448 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
10449 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
10451 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
10452 Pat Haugen <pthaugen@us.ibm.com>
10453 Peter Bergner <bergner@vnet.ibm.com>
10455 * config/rs6000/power8.md: New.
10456 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
10457 setting for power8 entry.
10458 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
10459 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
10460 test for Power4/Power5 only.
10461 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
10463 (force_new_group): Adjust comment.
10464 * config/rs6000/rs6000.md: Include power8.md.
10466 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
10468 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
10469 * config/arm/arm-protos.h (arm_max_conditional_execute): New
10471 (tune_params): Update comment.
10472 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
10473 (arm_max_conditional_execute): New function.
10474 (thumb2_final_prescan_insn): Use max_insn_skipped and
10475 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
10477 2013-06-25 Jakub Jelinek <jakub@redhat.com>
10479 PR tree-optimization/57705
10480 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
10481 SSA_NAME step, provided that it is not defined inside the loop.
10482 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
10483 (get_initial_def_for_induction): Handle SSA_NAME IV step.
10485 2013-06-25 Martin Jambor <mjambor@suse.cz>
10487 PR middle-end/57670
10488 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
10489 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
10491 (ipa_note_param_call): Initialize member_ptr flag.
10492 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
10493 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
10494 (ipa_write_indirect_edge_info): Stream member_ptr flag.
10495 (ipa_read_indirect_edge_info): Likewise.
10497 2013-06-25 Richard Biener <rguenther@suse.de>
10499 PR middle-end/56977
10500 * passes.c (init_optimization_passes): Move pass_fold_builtins
10501 and pass_dce earlier with -Og.
10503 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
10505 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
10506 <BIT_FIELD_REF>: Remove trailing TAB.
10507 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
10510 2013-06-24 Martin Jambor <mjambor@suse.cz>
10512 PR tree-optimization/57358
10513 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
10514 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
10515 (ipa_analyze_params_uses): Generate pessimistic info when true.
10517 2013-06-24 Martin Jambor <mjambor@suse.cz>
10519 PR tree-optimization/57539
10520 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
10521 global.inlined_to of the new node to it. All callers changed.
10522 * ipa-inline-transform.c (clone_inlined_nodes): New variable
10523 inlining_into, pass it to cgraph_clone_node.
10524 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
10525 ipa_free_edge_args_substructures.
10526 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
10527 rdesc linked list. Do not assert rdesc edges have inlined caller.
10528 Assert we have found an rdesc in the rdesc list.
10530 2013-06-24 Richard Biener <rguenther@suse.de>
10532 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
10533 (pointer_set_lookup): Declare.
10534 (class pointer_map): New template class implementing a
10535 generic pointer to T map.
10536 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
10537 pointer_map<T>::contains, pointer_map<T>::insert,
10538 pointer_map<T>::traverse): New functions.
10539 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10540 (pointer_set_lookup): New function.
10541 (pointer_set_contains): Use pointer_set_lookup.
10542 (pointer_set_insert): Likewise.
10543 (insert_aux): Remove.
10544 (struct pointer_map_t): Embed a pointer_set_t.
10545 (pointer_map_create): Adjust.
10546 (pointer_map_destroy): Likewise.
10547 (pointer_map_contains): Likewise.
10548 (pointer_map_insert): Likewise.
10549 (pointer_map_traverse): Likewise.
10550 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10551 pointer_map<unsigned> instead of a pointer_map_t.
10552 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10553 (streamer_tree_cache_lookup): Likewise.
10554 (streamer_tree_cache_create): Likewise.
10555 (streamer_tree_cache_delete): Likewise.
10556 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10557 pointer_map<unsigned> instead of a pointer_map_t.
10558 (lto_init_tree_ref_encoder): Adjust.
10559 (lto_destroy_tree_ref_encoder): Likewise.
10560 * lto-section-out.c (lto_output_decl_index): Likewise.
10561 (lto_record_function_out_decl_state): Likewise.
10562 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10564 2013-06-24 Richard Biener <rguenther@suse.de>
10566 PR tree-optimization/57488
10567 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10569 2013-06-24 Alan Modra <amodra@gmail.com>
10571 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10572 (gen_easy_altivec_constant): Likewise.
10573 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10574 easy_vector_constant_msb): Likewise.
10576 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10579 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10580 add missing return true.
10582 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10585 * config/sh/predicates.md (general_extend_operand): Invoke
10586 general_movsrc_operand for memory operands.
10587 (general_movsrc_operand): Allow reg+reg addressing, do not use
10588 general_operand for memory operands.
10590 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10592 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10593 when current target options does not apply.
10594 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10595 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10596 * config/i386/bmiintrin.h: Pass appropriate target
10597 attributes to header.
10598 * config/i386/mmintrin.h: Ditto.
10599 * config/i386/nmmintrin.h: Ditto.
10600 * config/i386/avx2intrin.h: Ditto.
10601 * config/i386/fxsrintrin.h: Ditto.
10602 * config/i386/tbmintrin.h: Ditto.
10603 * config/i386/xsaveintrin.h: Ditto.
10604 * config/i386/f16cintrin.h: Ditto.
10605 * config/i386/xtestintrin.h: Ditto.
10606 * config/i386/xsaveoptintrin.h: Ditto.
10607 * config/i386/bmi2intrin.h: Ditto.
10608 * config/i386/lzcntintrin.h: Ditto.
10609 * config/i386/smmintrin.h: Ditto.
10610 * config/i386/wmmintrin.h: Ditto.
10611 * config/i386/x86intrin.h: Remove all header include guards.
10612 * config/i386/prfchwintrin.h: Ditto.
10613 * config/i386/pmmintrin.h: Ditto.
10614 * config/i386/tmmintrin.h: Ditto.
10615 * config/i386/xmmintrin.h: Ditto.
10616 * config/i386/popcntintrin.h: Ditto.
10617 * config/i386/rdseedintrin.h: Ditto.
10618 * config/i386/ammintrin.h: Ditto.
10619 * config/i386/emmintrin.h: Ditto.
10620 * config/i386/immintrin.h: Remove all header include guards.
10621 * config/i386/fma4intrin.h: Ditto.
10622 * config/i386/lwpintrin.h: Ditto.
10623 * config/i386/xopintrin.h: Ditto.
10624 * config/i386/ia32intrin.h: Ditto.
10625 * config/i386/avxintrin.h: Ditto.
10626 * config/i386/rtmintrin.h: Ditto.
10627 * config/i386/fmaintrin.h: Ditto.
10628 * config/i386/mm3dnow.h: Ditto.
10630 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10632 * common/config/i386/i386-common.c: Handle LZCNT.
10634 2013-06-22 Andi Kleen <ak@linux.intel.com>
10636 * doc/extend.texi: Use __atomic_store_n instead of
10637 __atomic_store in HLE example.
10639 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10641 * config/sh/sh.c: Remove <cstdlib> workaround.
10643 2013-06-21 Andi Kleen <ak@linux.intel.com>
10645 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10647 2013-06-21 Andi Kleen <ak@linux.intel.com>
10649 * doc/extend.texi: Document that __atomic_clear and
10650 __atomic_test_and_set should only be used with bool.
10652 2013-06-20 Jan Hubicka <jh@suse.cz>
10654 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10655 types_same_for_odr.
10656 * tree.c (decls_same_for_odr): New function.
10657 (same_for_edr): New function.
10658 (types_same_for_odr): New function.
10659 (get_binfo_at_offset): Use it.
10660 * tree.h (types_same_for_odr): Declare.
10662 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10663 Jason Merrill <jason@redhat.com>
10665 * system.h: Include <cstdlib> as well as <stdlib.h>.
10667 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10670 * config/i386/i386.c (construct_container): Report error if
10671 long double is used with disabled x87 float returns.
10673 2013-06-20 Jan Hubicka <jh@suse.cz>
10675 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10677 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10679 PR rtl-optimization/57425
10680 PR rtl-optimization/57569
10681 * alias.c (write_dependence_p): Remove parameters mem_mode and
10682 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10683 Changed all callers.
10684 (canon_anti_dependence): Get comments and semantics in sync.
10685 Add parameter mem_canonicalized. Changed all callers.
10686 * rtl.h (canon_anti_dependence): Update prototype.
10688 2013-06-20 Richard Biener <rguenther@suse.de>
10690 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10691 byte case, inline streamer_read_uchar and defer section
10694 2013-06-20 Richard Biener <rguenther@suse.de>
10696 PR tree-optimization/57584
10697 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10698 SSA names into the expanded expression that take part in
10699 abnormal coalescing.
10701 2013-06-19 Sharad Singhai <singhai@google.com>
10703 * gcov.c (print_usage): Handle new option.
10704 (process_args): Ditto.
10705 (get_gcov_intermediate_filename): New function.
10706 (output_intermediate_file): New function.
10707 (output_gcov_file): New function
10708 (generate_results): Handle new option.
10709 (release_function): Relase demangled name.
10710 (read_graph_file): Handle demangled name.
10711 (output_lines): Ditto.
10712 * doc/gcov.texi: Document gcov intermediate format.
10714 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10717 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10718 (lra_emit_add): Use the functions. Add comment about Y as an
10721 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10724 * collect2.c (collect_atexit): New.
10725 (collect_exit): Delete.
10726 (main): Register collect_atexit with atexit.
10727 (collect_wait): Change collect_exit to exit.
10729 * collect2.h (collect_exit): Delete.
10730 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10732 2013-06-19 Wei Mi <wmi@google.com>
10734 PR rtl-optimization/57518
10735 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10736 if regno is used in paradoxical subreg.
10737 (update_equiv_regs): Check pdx_subregs[regno] before
10738 set a reg to be equivalent with a mem.
10740 2013-06-19 Matthias Klose <doko@ubuntu.com>
10743 * file-find.h (find_a_file): Add a mode parameter.
10744 * file-find.c (find_a_file): Likewise.
10745 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10746 with X_OK for the executables.
10747 * collect2.c (main): Call find_a_file with X_OK.
10749 2013-06-19 Steve Ellcey <sellcey@mips.com>
10752 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10753 Use NEXT_INSN instead of next_real_insn.
10755 2013-06-19 Jan Hubicka <jh@suse.cz>
10757 * cgraph.h (const_value_known_p): Replace by ...
10758 (ctor_for_folding): .. this one.
10759 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10760 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10761 * expr.c (expand_expr_real_1): Likewise.
10762 (string_constant): Likewise.
10763 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10764 * ipa.c (process_references): Likewise.
10765 (symtab_remove_unreachable_nodes): Likewise.
10766 * ipa-inline-analysis.c (param_change_prob): Likewise.
10767 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10768 (get_base_constructor): Likwise.
10769 * varpool.c (varpool_remove_node): Likewise.
10770 (varpool_remove_initializer): LIkewise.
10771 (dump_varpool_node): LIkwise.
10772 (const_value_known_p): Rewrite to ...
10773 (ctor_for_folding): ... this one.
10775 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10778 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10779 PERSONALITY in $PATH derived prefixes.
10781 2013-06-19 Jeff Law <law@redhat.com>
10783 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10786 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10787 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10790 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10792 * config/arm/vfp.md: Move VFP instruction classification documentation
10794 * config/arm/arm.md: ... here. Update instruction classification
10797 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10799 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10800 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10801 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10803 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10806 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10807 with NEXT_INSN. Use tablejump_p to check for jump table data
10810 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10813 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10814 that now in C++ the value is correct per the C++ standards.
10816 2013-06-19 Richard Biener <rguenther@suse.de>
10818 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10819 for global context.
10821 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10824 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10827 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10828 with next_active_insn.
10830 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10832 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10833 functions are inlined during failures to flag an error.
10834 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10835 in early inline pass.
10837 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10839 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10842 2013-06-18 Julian Brown <julian@codesourcery.com>
10844 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10845 Permit virtual register pre-reload if !strict.
10846 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10848 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10850 * config/arm/neon.md (movmisalign<mode>): Use
10851 neon_perm_struct_or_reg_operand instead of
10852 neon_struct_or_register_operand.
10853 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10854 neon_permissive_struct_operand instead of neon_struct_operand.
10855 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10856 neon_vector_mem_operand.
10857 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10858 neon_vector_mem_operand.
10859 (neon_permissive_struct_operand): New.
10860 (neon_struct_or_register_operand): Rename to...
10861 (neon_perm_struct_or_reg_operand): This. Adjust call to
10862 neon_vector_mem_operand.
10864 2013-06-18 Richard Biener <rguenther@suse.de>
10866 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10867 * lto-streamer.h: Include pointer-set.h.
10868 (struct lto_decl_slot): Remove.
10869 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10870 Remove next_index entry.
10871 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10872 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10873 (lto_init_tree_ref_encoder): Adjust.
10874 (lto_destroy_tree_ref_encoder): Likewise.
10875 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10876 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10877 (lto_output_decl_index): Adjust.
10878 (lto_new_out_decl_state): Likewise.
10879 (lto_record_function_out_decl_state): Likewise.
10880 * lto-streamer-out.c (copy_function): Likewise.
10882 2013-06-18 Richard Biener <rguenther@suse.de>
10884 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10885 * cgraphunit.c: Include cfgloop.h.
10886 (init_lowered_empty_function): Initialize the loop tree.
10887 (assemble_thunk): Insert new BBs into loops.
10889 2013-06-18 Richard Biener <rguenther@suse.de>
10891 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10892 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10893 the map from cache entry to cache index optional.
10894 (streamer_tree_cache_replace_tree): Adjust accordingly.
10895 (streamer_tree_cache_append): Likewise.
10896 (streamer_tree_cache_delete): Likewise.
10897 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10898 streamer cache map from cache entry to cache index.
10899 * lto-streamer-out.c (create_output_block): Adjust.
10901 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10903 * config/arm/arm.md (attribute "insn"): Move multiplication and
10904 division attributes to...
10905 (attribute "type"): ... here. Remove mult.
10906 (attribute "mul32"): New attribute.
10907 (attribute "mul64"): Add umaal.
10908 (*arm_mulsi3): Update attributes.
10909 (*arm_mulsi3_v6): Likewise.
10910 (*thumb_mulsi3): Likewise.
10911 (*thumb_mulsi3_v6): Likewise.
10912 (*mulsi3_compare0): Likewise.
10913 (*mulsi3_compare0_v6): Likewise.
10914 (*mulsi_compare0_scratch): Likewise.
10915 (*mulsi_compare0_scratch_v6): Likewise.
10916 (*mulsi3addsi): Likewise.
10917 (*mulsi3addsi_v6): Likewise.
10918 (*mulsi3addsi_compare0): Likewise.
10919 (*mulsi3addsi_compare0_v6): Likewise.
10920 (*mulsi3addsi_compare0_scratch): Likewise.
10921 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10922 (*mulsi3subsi): Likewise.
10923 (*mulsidi3adddi): Likewise.
10924 (*mulsi3addsi_v6): Likewise.
10925 (*mulsidi3adddi_v6): Likewise.
10926 (*mulsidi3_nov6): Likewise.
10927 (*mulsidi3_v6): Likewise.
10928 (*umulsidi3_nov6): Likewise.
10929 (*umulsidi3_v6): Likewise.
10930 (*umulsidi3adddi): Likewise.
10931 (*umulsidi3adddi_v6): Likewise.
10932 (*smulsi3_highpart_nov6): Likewise.
10933 (*smulsi3_highpart_v6): Likewise.
10934 (*umulsi3_highpart_nov6): Likewise.
10935 (*umulsi3_highpart_v6): Likewise.
10936 (mulhisi3): Likewise.
10937 (*mulhisi3tb): Likewise.
10938 (*mulhisi3bt): Likewise.
10939 (*mulhisi3tt): Likewise.
10940 (maddhisi4): Likewise.
10941 (*maddhisi4tb): Likewise.
10942 (*maddhisi4tt): Likewise.
10943 (maddhidi4): Likewise.
10944 (*maddhidi4tb): Likewise.
10945 (*maddhidi4tt): Likewise.
10946 (divsi3): Likewise.
10947 (udivsi3): Likewise.
10948 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10949 (thumb2_mulsi_short_compare0): Likewise.
10950 (thumb2_mulsi_short_compare0_scratch): Likewise.
10951 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10952 (1020mult2): Likewise.
10953 (1020mult3): Likewise.
10954 (1020mult4): Likewise.
10955 (1020mult5): Likewise.
10956 (1020mult6): Likewise.
10957 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10959 (cortex_a15_mult64): Likewise.
10960 (cortex_a15_sdiv): Likewise.
10961 (cortex_a15_udiv): Likewise.
10962 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10968 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10969 (pj4_ir_div): Likewise.
10970 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10971 (11_mult2): Likewise.
10972 (11_mult3): Likewise.
10973 (11_mult4): Likewise.
10974 (11_mult5): Likewise.
10975 (11_mult6): Likewise.
10976 (11_mult7): Likewise.
10977 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10978 (cortex_a8_mla): Likewise.
10979 (cortex_a8_mull): Likewise.
10980 (cortex_a8_smulwy): Likewise.
10981 (cortex_a8_smlald): Likewise.
10982 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10983 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10984 (cortex_r4_mul_3): Likewise.
10985 (cortex_r4_mla_4): Likewise.
10986 (cortex_r4_mla_3): Likewise.
10987 (cortex_r4_smlald): Likewise.
10988 (cortex_r4_mull): Likewise.
10989 (cortex_r4_sdiv): Likewise.
10990 (cortex_r4_udiv): Likewise.
10991 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10992 (cortex_a7_idiv): Likewise.
10993 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10994 (9_mult2): Likewise.
10995 (9_mult3): Likewise.
10996 (9_mult4): Likewise.
10997 (9_mult5): Likewise.
10998 (9_mult6): Likewise.
10999 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
11000 (cortex_a53_sdiv): Likewise.
11001 (cortex_a53_udiv): Likewise.
11002 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
11003 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
11004 (mp626_mult2): Likewise.
11005 (mp626_mult3): Likewise.
11006 (mp626_mult4): Likewise.
11007 * config/arm/fa526.md (526_mult1): Update attribute change.
11008 (526_mult2): Likewise.
11009 * config/arm/arm-generic.md (mult): Update attribute change.
11010 (mult_ldsched_strongarm): Likewise.
11011 (mult_ldsched): Likewise.
11012 (multi_cycle): Likewise.
11013 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
11014 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11015 (606te_mult2): Likewise.
11016 (606te_mult3): Likewise.
11017 (606te_mult4): Likewise.
11018 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11019 (cortex_a9_mac16): Likewise.
11020 (cortex_a9_multiply): Likewise.
11021 (cortex_a9_mac): Likewise.
11022 (cortex_a9_multiply_long): Likewise.
11023 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11024 (626te_mult2): Likewise.
11025 (626te_mult3): Likewise.
11026 (626te_mult4): Likewise.
11028 2013-06-18 Richard Biener <rguenther@suse.de>
11031 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11033 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11036 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11037 with next_active_insn.
11039 2013-06-18 Alan Modra <amodra@gmail.com>
11041 * config/rs6000/rs6000.h (enum data_align): New.
11042 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11043 (DATA_ABI_ALIGNMENT): Define.
11044 (CONSTANT_ALIGNMENT): Correct comment.
11045 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11046 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11048 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11050 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11051 ATTRIBUTE_UNUSED marking.
11053 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11055 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11056 alternative and update.
11057 (aarch64_dup_lanedi): Delete.
11058 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11059 * config/aarch64/aarch64-simd-builtins.def: Update.
11061 2013-06-17 Richard Biener <rguenther@suse.de>
11063 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11064 (lto_input_scc): Declare.
11065 (lto_input_tree_1): Likewise.
11066 (struct lto_stats_d): Add num_tree_bodies_output and
11067 num_pickle_refs_output.
11068 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11069 (lto_read_tree_1): Split out from ...
11070 (lto_read_tree): ... this.
11071 (lto_input_scc): New function.
11072 (lto_input_tree_1): Split out from ...
11073 (lto_input_tree): ... this. Handle LTO_tree_scc.
11074 (lto_data_in_create): Create the streamer cache without hashes.
11075 * lto-streamer-out.c (create_output_block): Create the streamer
11076 cache with hashes when not doing WPA.
11077 (lto_write_tree_1): Split out from ...
11078 (lto_write_tree): ... this.
11079 (get_symbol_initial_value): New function.
11080 (lto_output_tree_1): Split out from ...
11081 (lto_output_tree): ... this. Write trees as series of SCCs
11082 using a DFS walk via DFS_write_tree.
11083 (struct sccs, struct scc_entry): New types.
11084 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11085 (DFS_write_tree_body): New function.
11086 (DFS_write_tree): Likewise.
11087 (hash_tree): Likewise.
11088 (scc_entry_compare): Likewise.
11089 (hash_scc): Likewise.
11090 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11091 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11092 TREE_CHAIN as regular reference.
11093 (streamer_read_integer_cst): Remove.
11094 (streamer_get_pickled_tree): Adjust.
11095 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11096 of DECL_EXTERNALs in BLOCK_VARS for now.
11097 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11099 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11100 Add hash value argument and record that if hashes are recorded
11102 (streamer_tree_cache_insert_1): Adjust.
11103 (streamer_tree_cache_insert): Likewise.
11104 (streamer_tree_cache_insert_at): Rename to ...
11105 (streamer_tree_cache_replace_tree): ... this and adjust.
11106 (streamer_tree_cache_append): Adjust.
11107 (record_common_node): Likewise.
11108 (streamer_tree_cache_create): Add argument whether to
11109 record hash values together with trees.
11110 (streamer_tree_cache_delete): Adjust.
11111 * tree-streamer.h (struct streamer_tree_cache_d): Add
11113 (streamer_read_integer_cst): Remove.
11114 (streamer_tree_cache_insert): Adjust.
11115 (streamer_tree_cache_append): Likewise.
11116 (streamer_tree_cache_insert_at): Rename to ...
11117 (streamer_tree_cache_replace_tree): ... this and adjust.
11118 (streamer_tree_cache_create): Add argument whether to record hashes.
11119 (streamer_tree_cache_get): Rename to ...
11120 (streamer_tree_cache_get_tree): ... this.
11121 (streamer_tree_cache_get_hash): New function.
11122 * tree.c (cache_integer_cst): New function.
11123 * tree.h (cache_integer_cst): Declare.
11124 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
11125 * lto-symtab.c (lto_varpool_replace_node): Only release
11126 DECL_INITIAL of non-prevailing decls.
11127 * varpool.c (varpool_remove_initializer): Do not release
11128 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
11130 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11132 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
11133 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
11134 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
11135 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
11136 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
11137 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
11138 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
11139 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
11140 instead of TARGET_64BIT.
11141 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
11142 Require ISA_HAS_<D>DIV.
11144 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
11146 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
11147 (mips*-*-linux*): Move default with_llsc setting to where other
11149 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
11151 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
11152 Likewise. Remove default with_tune setting. Move default float
11153 setting to its own block. Handle with_llsc in the same block as above.
11155 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
11157 PR rtl-optimization/57425
11158 PR rtl-optimization/57569
11159 * alias.c (write_dependence_p): Add new parameters mem_mode,
11160 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
11161 Changed all callers.
11162 (canon_anti_dependence): New function.
11163 * cse.c (check_dependence): Use canon_anti_dependence.
11164 * cselib.c (cselib_invalidate_mem): Likewise.
11165 * rtl.h (canon_anti_dependence): Declare.
11167 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11169 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
11170 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
11171 ".set mips3" for 64-bit targets.
11173 2013-06-15 Dehao Chen <dehao@google.com>
11175 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
11176 * gimple-low.c (gimple_check_call_matching_types): Likewise.
11177 (gimple_check_call_args): Likewise.
11178 * value-prof.c (check_ic_target): Likewise.
11179 * ipa-inline.c (early_inliner): Likewise.
11180 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
11181 * cgraph.c (cgraph_create_edge_1): Likewise.
11182 (cgraph_make_edge_direct): Likewise.
11184 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
11187 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
11188 rs6000_output_move_128bit to handle emitting quad memory
11189 operations. Set attribute length to 8 bytes.
11191 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
11193 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
11195 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
11196 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
11197 (aarch64_<su>mlsl<mode>): Likewise.
11199 2013-06-14 Mike Stump <mikestump@comcast.net>
11201 * Makefile.in (TARGET_H): Add insn-codes.h.
11203 2013-06-14 Alan Modra <amodra@gmail.com>
11205 PR middle-end/57134
11206 PR middle-end/57586
11207 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
11208 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
11209 bitfield expansion when EXPAND_MEMORY.
11210 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
11212 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
11214 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
11215 test for clearing quad memory on 32-bit later.
11217 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11219 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
11220 (fold_negate_expr): Likewise.
11221 (fold_real_zero_addition_p): Handle vectors.
11222 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
11224 2013-06-14 Alan Modra <amodra@gmail.com>
11226 * varasm.c (force_const_mem): Revert 2013-06-07 change.
11228 2013-06-13 Jan Hubicka <jh@suse.cz>
11230 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11231 Local comdats are not externally visible.
11232 * symtab.c (dump_symtab_base): Dump externally visible.
11233 (verify_symtab_base): Verify back links in the symtab hash.
11235 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11237 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
11238 CONVERT_EXPR as equal nodes.
11240 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11242 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
11244 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11246 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
11247 Generalize to complex and vector.
11248 * tree.c (build_all_ones_cst): New function.
11249 * tree.h (build_all_ones_cst): Declare it.
11251 2013-06-13 Alan Modra <amodra@gmail.com>
11253 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11254 * config/rs6000/rs6000.md (signbittf2): New insn.
11255 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11256 (abstf2_internal, cmptf_internal2): Likewise.
11257 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11259 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11260 Pat Haugen <pthaugen@us.ibm.com>
11261 Peter Bergner <bergner@vnet.ibm.com>
11263 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11264 power8 byte, half-word, and quad-word atomic instructions.
11265 (emit_store_conditional): Likewise.
11266 (rs6000_expand_atomic_compare_and_swap): Likewise.
11267 (rs6000_expand_atomic_op): Likewise.
11269 * config/rs6000/sync.md (larx): Add new modes for power8.
11271 (AINT): New mode iterator to include TImode as well as normal
11272 integer modes on power8.
11273 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
11274 that VSX registers are not considered. Use AINT mode iterator
11275 instead of INT1 to allow inclusion of quad word atomic operations
11277 (load_locked<mode>): Likewise.
11278 (store_conditional<mode>): Likewise.
11279 (atomic_compare_and_swap<mode>): Likewise.
11280 (atomic_exchange<mode>): Likewise.
11281 (atomic_nand<mode>): Likewise.
11282 (atomic_fetch_<fetchop_name><mode>): Likewise.
11283 (atomic_nand_fetch<mode>): Likewise.
11284 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
11286 (ATOMIC): On power8, add QImode, HImode modes.
11287 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
11288 modes that promote to SImode.
11289 (load_lockedti): Convert TImode arguments to PTImode, so that we
11290 get a guaranteed even/odd register pair.
11291 (load_lockedpti): Likewise.
11292 (store_conditionalti): Likewise.
11293 (store_conditionalpti): Likewise.
11295 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
11296 atomic load/store instructions.
11299 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
11301 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
11303 (insn_count): New attribute, with most cases extracted from...
11304 (length): ...here. Redefine most cases in terms of insn_count.
11305 (single_insn): Delete.
11306 (can_delay): Use insn_count to check for single instructions.
11307 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
11308 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
11309 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
11310 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
11311 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
11312 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
11313 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
11314 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
11315 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
11316 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
11317 rather than "length".
11318 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
11319 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
11320 Use "insn_count" rather than "length".
11321 * config/mips/mips-dsp.md
11322 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
11323 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
11326 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
11328 PR tree-optimization/57361
11329 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
11331 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
11333 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
11335 (aarch64_simd_combine<mode>): New instruction expansion.
11336 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
11337 function prototype.
11338 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
11339 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
11341 2013-06-12 Jan Hubicka <jh@suse.cz>
11343 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
11344 decl has when in streaming stage.
11345 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
11346 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
11348 2013-06-12 Roland Stigge <stigge@antcom.de>
11351 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
11353 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11355 PR tree-optimization/57537
11356 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
11357 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
11359 2013-06-12 Richard Biener <rguenther@suse.de>
11361 * data-streamer.h (streamer_write_char_stream): CSE
11362 obs->current_pointer.
11363 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
11364 streamer_write_char_stream manually and optimize the resulting loop.
11365 (streamer_write_hwi_stream): Likewise.
11367 2013-06-12 Jan Hubicka <jh@suse.cz>
11369 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
11370 * cgraph.h (varpool_create_empty_node): Declare.
11371 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
11373 * symtab.c (symtab_unregister_node): Be lax about missin entries
11375 (symtab_get_node): Update comment.
11376 * varpool.c (varpool_create_empty_node): Break out from ...
11377 (varpool_node_for_decl): ... here.
11378 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
11380 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
11382 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
11383 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
11384 part. Use straight-line flow at the end.
11385 <COMPONENT_REF>: Remove superfluous else.
11386 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
11388 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11391 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
11392 target hook even for !TREE_PUBLIC decls. If no resolution info
11393 is available, return false for common and external decls.
11395 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11397 * config/rl78/constraints.md (U): New constraint.
11398 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
11401 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
11404 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
11405 to allow returning address to AT_PLATFORM name.
11407 2013-06-11 Jan Hubicka <jh@suse.cz>
11409 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
11410 * cgraph.h (symtab_node_base): Add weakref flag.
11411 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
11412 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
11413 (output_weakrefs): Use weakref flag.
11414 * fold-const.c (simple_operand_p): Handle WEAK.
11415 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
11416 * ipa.c (varpool_externally_visible_p): Drop weakref.
11417 (function_and_variable_visibility): Update comment; fix weakref
11418 sanity checks; do not clear DECL_WEAK on them.
11419 * lto-cgraph.c (lto_output_node): update.
11420 (lto_output_varpool_node): Update.
11421 (input_overwrite_node): Update.
11422 (input_node): Update.
11423 (input_varpool_node): Update.
11424 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
11425 (lto_symtab_merge_symbols): Add sanity check.
11426 (lto_symtab_prevailing_decl): Do not special case weakrefs.
11427 * passes.c (rest_of_decl_compilation): Set static flag, too.
11428 * symtab.c (dump_symtab_base): Dump weakref.
11429 (verify_symtab_base): Sanity check weakrefs.
11430 (symtab_make_decl_local): Remove duplicated code.
11431 (symtab_alias_ultimate_target): Simplify.
11432 * varpool.c (varpool_create_variable_alias): Set weakref flag.
11434 2013-06-11 Tom de Vries <tom@codesourcery.com>
11436 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
11437 sequence_vect == NULL.
11439 2013-06-11 DJ Delorie <dj@redhat.com>
11441 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
11442 (rl78_unwind_word_mode): New.
11444 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11446 * final.c (debug_prefix_maps): Make static.
11448 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11450 * function.c (initial_trampoline): Remove stray copy.
11452 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
11454 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
11456 2013-06-11 Martin Jambor <mjambor@suse.cz>
11458 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
11459 within bounds at the beginning of the function.
11461 2013-06-11 Alan Modra <amodra@gmail.com>
11463 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
11465 (get_named_section): Don't NULL !DECL_P decl.
11467 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
11469 * doc/invoke.texi (core-avx2): Document.
11471 (atom): Updated with MOVBE.
11473 2013-06-11 Richard Biener <rguenther@suse.de>
11475 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
11477 2013-06-11 Anton Blanchard <anton@samba.org>
11479 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
11480 correct shift value in little-endian mode.
11482 2013-06-11 Jakub Jelinek <jakub@redhat.com>
11485 * varasm.c (get_variable_align): Move #endif to the right place.
11487 2013-06-10 Cary Coutant <ccoutant@google.com>
11489 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
11490 for hash so that hash table traversal order is deterministic.
11492 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
11493 Pat Haugen <pthaugen@us.ibm.com>
11494 Peter Bergner <bergner@vnet.ibm.com>
11496 * config/rs6000/vector.md (GPR move splitter): Do not split moves
11497 of vectors in GPRS if they are direct moves or quad word load or
11500 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
11502 (direct_move_p): Likewise.
11503 (quad_load_store_p): Likewise.
11505 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
11506 classes into bins based on the physical register type.
11507 (reg_class_to_reg_type): Likewise.
11508 (IS_STD_REG_TYPE): Likewise.
11509 (IS_FP_VECT_REG_TYPE): Likewise.
11510 (reload_fpr_gpr): Arrays to determine what insn to use if we can
11511 use direct move instructions.
11512 (reload_gpr_vsx): Likewise.
11513 (reload_vsx_gpr): Likewise.
11514 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
11515 information that is a simplification of register classes. Also
11516 precalculate direct move reload helpers.
11517 (direct_move_p): New function to return true if the operation can
11518 be done as a direct move instruciton.
11519 (quad_load_store_p): New function to return true if the operation
11520 is a quad memory operation.
11521 (rs6000_legitimize_address): If quad memory, only allow register
11522 indirect for TImode addresses.
11523 (rs6000_legitimate_address_p): Likewise.
11524 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
11525 (rs6000_reload_register_type): Likewise.
11526 (register_to_reg_type): Return register type.
11527 (rs6000_secondary_reload_simple_move): New helper function for
11528 secondary reload and secondary memory needed to identify anything
11529 that is a simple move, and does not need reloading.
11530 (rs6000_secondary_reload_direct_move): New helper function for
11531 secondary reload to identify cases that can be done with several
11532 instructions via the direct move instructions.
11533 (rs6000_secondary_reload_move): New helper function for secondary
11534 reload to identify moves between register types that can be done.
11535 (rs6000_secondary_reload): Add support for quad memory operations
11536 and for direct move.
11537 (rs6000_secondary_memory_needed): Likewise.
11538 (rs6000_debug_secondary_memory_needed): Change argument names.
11539 (rs6000_output_move_128bit): New function to return the move to
11540 use for 128-bit moves, including knowing about the various
11541 limitations of quad memory operations.
11543 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11544 memory operations. call rs6000_output_move_128bit for the actual
11545 instruciton(s) to generate.
11546 (vsx_movti_64bit): Likewise.
11548 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11549 (UNSPEC_P8V_MTVSRWZ): Likewise.
11550 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11551 (UNSPEC_P8V_MTVSRD): Likewise.
11552 (UNSPEC_P8V_XXPERMDI): Likewise.
11553 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11554 (UNSPEC_FUSION_GPR): Likewise.
11555 (FMOVE128_GPR): New iterator for direct move.
11556 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11557 (f32_sv): Likewise.
11558 (f32_dm): Likewise.
11559 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11560 loads and direct move instructions.
11561 (zero_extendsidi2_lfiwzx): Likewise.
11562 (extendsidi2_lfiwax): Likewise.
11563 (extendsidi2_nocell): Likewise.
11564 (floatsi<mode>2_lfiwax): Likewise.
11565 (lfiwax): Likewise.
11566 (floatunssi<mode>2_lfiwzx): Likewise.
11567 (lfiwzx): Likewise.
11568 (fix_trunc<mode>_stfiwx): Likewise.
11569 (fixuns_trunc<mode>_stfiwx): Likewise.
11570 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11571 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11572 (parity<mode>2_cmpb): Set length/type attr.
11573 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11574 for 'mr.' to fast_compare.
11575 (bpermd_<mode>): Change type attr to popcnt.
11576 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11577 (p8_mtvsrwz_1): Likewise.
11578 (p8_mtvsrwz_2): Likewise.
11579 (reload_fpr_from_gpr<mode>): Likewise.
11580 (p8_mtvsrd_1): Likewise.
11581 (p8_mtvsrd_2): Likewise.
11582 (p8_xxpermdi_<mode>): Likewise.
11583 (reload_vsx_from_gpr<mode>): Likewise.
11584 (reload_vsx_from_gprsf): Likewise.
11585 (p8_mfvsrd_3_<mode>): LIkewise.
11586 (reload_gpr_from_vsx<mode>): Likewise.
11587 (reload_gpr_from_vsxsf): Likewise.
11588 (p8_mfvsrd_4_disf): Likewise.
11589 (multi-word GPR splits): Do not split direct moves or quad memory
11592 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11594 * tree-into-ssa.c (interesting_blocks): Make static.
11596 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11599 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11600 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11601 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11602 (get_variable_align): New function.
11603 (get_variable_section, emit_bss, emit_common,
11604 assemble_variable_contents, place_block_symbol): Use
11605 get_variable_align instead of DECL_ALIGN.
11606 (assemble_noswitch_variable): Add align argument, use it
11607 instead of DECL_ALIGN.
11608 (assemble_variable): Adjust caller. Use get_variable_align
11609 instead of DECL_ALIGN.
11610 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11612 (DATA_ABI_ALIGNMENT): Define.
11613 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11614 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11615 opt is false, only return the psABI mandated alignment increase.
11616 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11617 (DATA_ABI_ALIGNMENT): ... this.
11618 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11619 (DATA_ABI_ALIGNMENT): ... this.
11620 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11621 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11622 (DATA_ABI_ALIGNMENT): ... this.
11623 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11624 * doc/tm.texi: Regenerated.
11626 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11628 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11629 cmp_code to construct REG_EQUAL note.
11631 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11634 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11635 that operands[2] doesn't overlap with operands[0].
11637 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11638 Jan Hubicka <jh@suse.cz>
11640 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11641 hack to mark symbols as used.
11643 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11645 PR rtl-optimization/57559
11646 * lra-constraints.c (process_alt_operands): Don't discourage
11647 memory with known offset for offsetable memory constraint.
11648 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11650 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11652 * varasm.c (struct oc_local_state): Reorder fields.
11653 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11654 and adjust accordingly.
11655 (output_constructor): Reorder initialization code and adjust call to
11656 output_constructor_bitfield.
11658 2013-06-07 Jan Hubicka <jh@suse.cz>
11660 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11662 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11664 * tree-object-size.c (unknown): Make const.
11666 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11668 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11669 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11670 for last alternative in the cpu_facility attribute.
11672 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11675 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11676 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11677 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11678 * config/arm/constraints.md (Dg): New constraint.
11679 * config/arm/neon.md (xordi3_neon): Remove.
11680 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11681 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11683 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11685 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11686 Clean up alternatives.
11688 2013-06-07 Alan Modra <amodra@gmail.com>
11690 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11693 2013-06-07 Alan Modra <amodra@gmail.com>
11695 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11697 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11699 * config/arm/constraints.md (Df): New constraint.
11700 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11701 Correct length attribute for last two alternatives.
11703 2013-06-07 Alan Modra <amodra@gmail.com>
11705 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11706 override user -mfp-in-toc.
11707 (offsettable_ok_by_alignment): Consider just the current access
11708 rather than the whole object, unless BLKmode. Handle
11709 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11710 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11711 for -mcmodel=medium.
11712 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11713 override user -mfp-in-toc or -msum-in-toc. Default to
11714 -mno-fp-in-toc for -mcmodel=medium.
11716 2013-06-06 DJ Delorie <dj@redhat.com>
11718 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11719 TARGET_VALID_POINTER_MODE.
11721 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11722 Pat Haugen <pthaugen@us.ibm.com>
11723 Peter Bergner <bergner@vnet.ibm.com>
11725 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11726 Document new power8 builtins.
11728 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11729 condition code register, to allow 128-bit logical operations to be
11730 done in the VSX or GPR registers.
11731 (nor<mode>3): Use the canonical form for nor.
11732 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11733 vclz*, and vpopcnt* vector instructions.
11734 (nand<mode>3): Likewise.
11735 (orc<mode>3): Likewise.
11736 (clz<mode>2): LIkewise.
11737 (popcount<mode>2): Likewise.
11739 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11740 that only the GPRs are recognized.
11742 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11743 support for new power8 builtins.
11745 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11747 (xscvdpspn): Likewise.
11753 (vpopcnt): Likewise.
11754 (vpopcntb): Likewise.
11755 (vpopcnth): Likewise.
11756 (vpopcntw): Likewise.
11757 (vpopcntd): Likewise.
11759 (vmrgew): Likewise.
11760 (vmrgow): Likewise.
11762 (eqv_v16qi3): Likewise.
11763 (eqv_v8hi3): Likewise.
11764 (eqv_v4si3): Likewise.
11765 (eqv_v2di3): Likewise.
11766 (eqv_v4sf3): Likewise.
11767 (eqv_v2df3): Likewise.
11769 (nand_v16qi3): Likewise.
11770 (nand_v8hi3): Likewise.
11771 (nand_v4si3): Likewise.
11772 (nand_v2di3): Likewise.
11773 (nand_v4sf3): Likewise.
11774 (nand_v2df3): Likewise.
11776 (orc_v16qi3): Likewise.
11777 (orc_v8hi3): Likewise.
11778 (orc_v4si3): Likewise.
11779 (orc_v2di3): Likewise.
11780 (orc_v4sf3): Likewise.
11781 (orc_v2df3): Likewise.
11783 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11784 allow power8 quad mode in 64-bit.
11785 (rs6000_builtin_vectorized_function): Add support to vectorize
11786 ISA 2.07 count leading zeros, population count builtins.
11787 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11788 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11789 (builtin_function_type): Add vgbbd builtin function which takes an
11791 (altivec_expand_vec_perm_const): Add support for new power8 merge
11794 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11795 that does not include TImdoe for use with 32-bit.
11796 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11798 (UNSPEC_VSX_CVDPSPN): Likewise.
11799 (vsx_xscvdpspn): Likewise.
11800 (vsx_xscvspdpn): Likewise.
11801 (vsx_xscvdpspn_scalar): Likewise.
11802 (vsx_xscvspdpn_directmove): Likewise.
11803 (vsx_and<mode>3): Split logical operations into 32-bit and
11804 64-bit. Add support to do logical operations on TImode as well as
11805 VSX vector types. Allow logical operations to be done in either
11806 VSX registers or in general purpose registers in 64-bit mode. Add
11807 splitters if GPRs were used. For AND, add clobber of CCmode to
11808 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11810 (vsx_and<mode>3_32bit): Likewise.
11811 (vsx_and<mode>3_64bit): Likewise.
11812 (vsx_ior<mode>3): Likewise.
11813 (vsx_ior<mode>3_32bit): Likewise.
11814 (vsx_ior<mode>3_64bit): Likewise.
11815 (vsx_xor<mode>3): Likewise.
11816 (vsx_xor<mode>3_32bit): Likewise.
11817 (vsx_xor<mode>3_64bit): Likewise.
11818 (vsx_one_cmpl<mode>2): Likewise.
11819 (vsx_one_cmpl<mode>2_32bit): Likewise.
11820 (vsx_one_cmpl<mode>2_64bit): Likewise.
11821 (vsx_nor<mode>3): Likewise.
11822 (vsx_nor<mode>3_32bit): Likewise.
11823 (vsx_nor<mode>3_64bit): Likewise.
11824 (vsx_andc<mode>3): Likewise.
11825 (vsx_andc<mode>3_32bit): Likewise.
11826 (vsx_andc<mode>3_64bit): Likewise.
11827 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11828 and xxlorc instructions.
11829 (vsx_eqv<mode>3_64bit): Likewise.
11830 (vsx_nand<mode>3_32bit): Likewise.
11831 (vsx_nand<mode>3_64bit): Likewise.
11832 (vsx_orc<mode>3_32bit): Likewise.
11833 (vsx_orc<mode>3_64bit): Likewise.
11835 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11837 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11839 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11840 (p8_vmrgow): Likewise.
11841 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11842 GPRs to be split under VSX.
11843 (p8v_clz<mode>2): Add power8 count leading zero support.
11844 (p8v_popcount<mode>2): Add power8 population count support.
11845 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11848 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11851 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11853 (vec_nand): Likewise.
11854 (vec_vclz): Likewise.
11855 (vec_vclzb): Likewise.
11856 (vec_vclzd): Likewise.
11857 (vec_vclzh): Likewise.
11858 (vec_vclzw): Likewise.
11859 (vec_vgbbd): Likewise.
11860 (vec_vmrgew): Likewise.
11861 (vec_vmrgow): Likewise.
11862 (vec_vpopcnt): Likewise.
11863 (vec_vpopcntb): Likewise.
11864 (vec_vpopcntd): Likewise.
11865 (vec_vpopcnth): Likewise.
11866 (vec_vpopcntw): Likewise.
11868 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11870 PR rtl-optimization/57468
11871 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11874 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11876 PR rtl-optimization/57459
11877 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11878 type when setting live regs.
11880 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11882 * config/s390/s390.opt (mlra): New option.
11883 * config/s390/s390.c (s390_decompose_address): Check displacement
11884 for all registers for LRA.
11885 (s390_secondary_reload): Don't used secondary reloads for LRA.
11886 (s390_lra_p): New function.
11887 (TARGET_LRA_P): Define.
11888 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11889 of attribute cpu_facility to zarch for the last alternative.
11890 (*cmpmem_short): Ditto.
11892 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11894 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11895 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11896 (arm_expand_prologue): Likewise.
11898 2013-06-06 Teresa Johnson <tejohnson@google.com>
11901 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11902 as this is now done by redirect_edge_and_branch_force.
11903 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11904 barriers, and fix interaction with splitting.
11905 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11906 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11907 reflect changes made in the routine.
11908 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11909 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11910 since this is called in cfglayout mode, and replace partition fixup
11911 with assert as that is now done by force_nonfallthru_and_redirect.
11912 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11913 already be marked with region crossing note.
11914 (insert_section_boundary_note): Make non-static, gate on flag
11915 has_bb_partition, rewrite to also check for multiple partitions.
11916 (rest_of_handle_reorder_blocks): Remove call to
11917 insert_section_boundary_note, now done later during free_cfg.
11918 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11919 * bb-reorder.h (insert_section_boundary_note): Declare.
11920 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11921 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11922 invoke insert_section_boundary_note.
11923 (try_redirect_by_replacing_jump): Remove unnecessary
11924 check for region crossing note.
11925 (fixup_partition_crossing): New function.
11926 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11927 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11928 in non-cfglayout mode.
11929 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11930 remove old code that tried to do this. Emit barrier correctly
11931 when we are in cfglayout mode.
11932 (last_bb_in_partition): New function.
11933 (rtl_split_edge): Correctly fixup partition boundaries.
11934 (commit_one_edge_insertion): Remove old code that tried to
11935 fixup region crossing edge since this is now handled in
11936 split_block, and set up insertion point correctly since
11937 block may now end in a jump.
11938 (verify_hot_cold_block_grouping): Guard against checking when not in
11939 linearized RTL mode.
11940 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11942 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11943 rtl_verify_flow_info, so not called in cfglayout mode.
11944 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11945 (fixup_reorder_chain): Remove old code that attempted to fixup region
11946 crossing note as this is now handled in force_nonfallthru_and_redirect.
11947 (duplicate_insn_chain): Don't duplicate switch section notes.
11948 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11950 * basic-block.h (emit_barrier_after_bb): Declare.
11952 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11954 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11955 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11956 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11958 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11960 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11962 * config/arm/ldmstm.md: Regenerate.
11964 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11966 * config/arm/sync.md (atomic_loaddi_1):
11967 Disable predication for arm_restrict_it.
11968 (arm_load_exclusive<mode>): Likewise.
11969 (arm_load_exclusivesi): Likewise.
11970 (arm_load_exclusivedi): Likewise.
11971 (arm_load_acquire_exclusive<mode>): Likewise.
11972 (arm_load_acquire_exclusivesi): Likewise.
11973 (arm_load_acquire_exclusivedi): Likewise.
11974 (arm_store_exclusive<mode>): Likewise.
11975 (arm_store_exclusive<mode>): Likewise.
11976 (arm_store_release_exclusivedi): Likewise.
11977 (arm_store_release_exclusive<mode>): Likewise.
11979 2013-06-06 Richard Biener <rguenther@suse.de>
11981 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11983 (lto_tag_is_tree_code_p): Adjust.
11984 (lto_tag_is_gimple_code_p): Likewise.
11985 (lto_gimple_code_to_tag): Likewise.
11986 (lto_tag_to_gimple_code): Likewise.
11987 (lto_tree_code_to_tag): Likewise.
11988 (lto_tag_to_tree_code): Likewise.
11989 * data-streamer.h (streamer_write_hwi_in_range): Use
11990 uhwi streaming to stream the normalized range.
11991 (streamer_read_hwi_in_range): Likewise.
11993 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11995 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11996 (predicable_short_it): Likewise.
11997 (predicated): Likewise.
11998 (enabled): Handle above.
11999 (define_cond_exec): Set predicated attribute to yes.
12001 2013-06-05 Mike Stump <mikestump@comcast.net>
12003 * gdbinit.in (__FUNCTION__): Add.
12005 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
12007 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
12008 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
12010 2013-06-05 Jan Hubicka <jh@suse.cz>
12012 * varasm.c (mark_decl_referenced): Revert the removal until targets
12015 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12017 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12018 instead of mark_decl_referenced.
12020 2013-06-05 Jan Hubicka <jh@suse.cz>
12022 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12023 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12024 and symtab_used_from_object_file_p.
12025 (cgraph_make_node_local_1): Clear forced_by_abi.
12026 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12027 * cgraph.h (symtab_node_base): Add forced_by_abi.
12028 (decide_is_variable_needed): Remove.
12029 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12030 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12031 (decide_is_symbol_needed): ... this one; handle symbols in general;
12032 always analyze virtuals; honnor forced_by_abi.
12033 (cgraph_finalize_function): Update.
12034 (varpool_finalize_decl): Update.
12035 (symbol_defined_and_needed): Remove.
12036 (analyze_functions): Update.
12037 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12038 output_refs, input_overwrite_node): Handle forced_by_abi.
12039 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12040 (address_taken_from_non_vtable_p): ... this one.
12041 (comdat_can_be_unshared_p_1): New function.
12042 (cgraph_comdat_can_be_unshared_p): Rename to ...
12043 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12044 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12045 (function_and_variable_visibility): Clear forced_by_abi as needed.
12046 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12047 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12048 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12049 * varpool.c (decide_is_variable_needed): Remove.
12051 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12053 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12054 (arm_option_override): Override arm_restrict_it where appropriate.
12055 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12056 * config/arm/arm.opt (mrestrict-it): New command-line option.
12057 * doc/invoke.texi: Document -mrestrict-it.
12059 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12061 * tsan.c (tsan_atomic_table): Make const.
12063 2013-06-05 Richard Biener <rguenther@suse.de>
12065 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12066 index associated with the tree we are supposed to replace.
12067 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12068 TREE_ASM_WRITTEN as zero for everything but SSA names.
12070 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12072 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12074 2013-06-04 Jan Hubicka <jh@suse.cz>
12076 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12077 (input_node, input_varpool_node): Handle correctly external same
12079 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12080 nodes at ltrans stage.
12082 2013-06-04 Jan Hubicka <jh@suse.cz>
12084 * ipa-inline.c (update_caller_keys): Fix availability test.
12085 (update_callee_keys): Likewise.
12086 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12087 to follow ELF standard.
12089 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12091 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12092 (mips64r5900el-*-elf*): New configurations.
12093 * config/mips/mips-cpus.def (r5900): New processor.
12094 * config/mips/mips-tables.opt: Regenerate.
12095 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12096 (mips_issue_rate): Handle PROCESSOR_R5900.
12097 (mips_reorg_process_insns): Force reorder mode for the R5900.
12098 * config/mips/mips.h (TARGET_MIPS5900): Define.
12099 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12101 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12103 * config/mips/mips.md (processor): Add r5900.
12104 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12106 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12108 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12109 into function to generate MOVI instruction.
12110 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12111 (aarch64_preferred_simd_mode): Turn into wrapper.
12112 (aarch64_output_scalar_simd_mov_immediate): New function.
12113 * config/aarch64/aarch64-protos.h: Add prototype for above.
12115 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12117 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12118 element_char member.
12119 (sizetochar): Return signed char.
12120 (aarch64_simd_valid_immediate): Remove elchar and other
12121 unnecessary variables.
12122 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12123 Calculate element_char as required.
12124 * config/aarch64/aarch64-protos.h: Update and move prototype
12125 for aarch64_output_simd_mov_immediate.
12126 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
12129 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12131 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
12132 information completed by aarch64_simd_valid_immediate.
12133 (aarch64_legitimate_constant_p): Update arguments.
12134 (aarch64_simd_valid_immediate): Work with struct rather than many
12136 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
12137 (aarch64_simd_make_constant): Update arguments.
12138 (aarch64_output_simd_mov_immediate): Work with struct rather than
12139 many pointers. Output immediate directly rather than as operand.
12140 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12142 * config/aarch64/constraints.md (Dn): Update arguments.
12144 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12146 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
12148 (aarch64_simd_immediate_valid_for_move): Remove.
12149 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
12150 (aarch64_simd_make_constant): Update call.
12151 (aarch64_output_simd_mov_immediate): Update call.
12152 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12154 * config/aarch64/constraints.md (Dn): Update call.
12156 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12158 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
12159 return type to bool for prototype.
12160 (aarch64_legitimate_constant_p): Check for true instead of not -1.
12161 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
12162 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
12164 2013-06-04 Catherine Moore <clm@codesourcery.com>
12166 * config/mips/mips.opt (meva): New.
12167 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
12168 (ASM_SPEC): Handle -meva.
12169 * doc/invoke.texi (meva): Document.
12171 2013-06-04 Alan Modra <amodra@gmail.com>
12173 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
12176 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12178 * rtl.def: Add extra fourth optional field to define_cond_exec.
12179 * gensupport.c (process_one_cond_exec): Process attributes from
12181 * doc/md.texi: Document fourth field in define_cond_exec.
12183 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
12185 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
12186 out the processing order as in store_bit_field_1.
12188 2013-06-04 Jan Hubicka <jh@suse.cz>
12190 PR middle-end/57500
12191 * cgraphunit.c (cgraph_process_same_body_aliases): Create
12192 non-VAR_DECL node if it does not exist yet.
12194 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
12196 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
12197 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
12198 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
12199 target_cpu_default setting.
12201 2013-06-03 Teresa Johnson <tejohnson@google.com>
12203 * dumpfile.c (opt_info_switch_p): Change -fopt-info
12204 default to -fopt-info=optimized instead of all.
12205 * doc/invoke.texi: Ditto.
12206 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
12207 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
12208 (execute_vect_slp): Emit BB vectorization success under
12209 MSG_OPTIMIZED_LOCATIONS.
12210 * tree-vect-slp.c (vect_slp_transform_bb): Change
12211 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
12212 * tree-vect-loop.c (vect_transform_loop): Ditto.
12214 2013-06-03 Jason Merrill <jason@redhat.com>
12217 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12218 Use TARGET_EXPR for C++.
12220 2013-06-03 Jakub Jelinek <jakub@redhat.com>
12222 PR rtl-optimization/57268
12223 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
12224 if DEBUG_INSN_P (insn).
12227 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12229 PR rtl-optimization/57268
12230 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12231 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12233 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12235 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
12236 (ix86_avoid_lea_for_addr): Likewise.
12237 (exact_dependency_1): Likewise.
12238 (ix86_adjust_cost): Likewise.
12239 (swap_top_of_ready_list): Fix formatting and !reload_completed check
12241 (do_reorder_for_imul): Fix typo, formatting and
12242 !reload_completed check removed.
12243 (ix86_sched_reorder): Fix typo and formatting.
12244 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
12247 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
12249 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12251 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12253 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12254 <STRING_CST>: Likewise.
12255 <VECTOR_CST>: Likewise.
12257 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12258 Mikael Morin <mikael@gcc.gnu.org>
12260 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12261 * config.in: Regenerated.
12262 * configure: Regenerated.
12264 2013-06-01 Jan Hubicka <jh@suse.cz>
12266 PR middle-end/57366
12267 * cgraphunit.c (compile): When weakref is not supported,
12268 set up transparent aliases before final output pass.
12269 * varasm.c (assemble_alias): Do not try to do it here.
12271 2013-06-01 Jan Hubicka <jh@suse.cz>
12273 PR middle-end/57467
12274 * passes.c (for_per_function): Skip unanalyzed functions.
12276 2013-06-01 Jan Hubicka <jh@suse.cz>
12278 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
12279 (lto_symtab_merge_symbols_1): ... this one.
12280 (lto_symtab_merge_cgraph_nodes): Rename to ...
12281 (lto_symtab_merge_symbols): ... this one; simplify.
12282 * cgraph.c (same_body_aliases_done): Rename to ...
12283 (cpp_implicit_aliases_done): ... this one.
12284 (cgraph_create_function_alias): Update.
12285 (cgraph_same_body_alias): Update.
12286 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
12287 (verify_edge_corresponds_to_fndecl): Simplify.
12288 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
12289 (cgraph_node): Remove same_body_alias.
12290 (varpool_node): Remove alias_of and extra_name_alias.
12291 (same_body_aliases_done): Rename to ..
12292 (cpp_implicit_aliases_done): ... this one.
12293 (symtab_alias_ultimate_target): Add default parameter.
12294 (symtab_resolve_alias): New function.
12295 (fixup_same_cpp_alias_visibility): Declare.
12296 (cgraph_function_node): Add default parameter.
12297 (cgraph_node_asm_name): Likewise.
12298 (cgraph_function_or_thunk_node): Add default parameter; do
12299 not ICE when it is NULL.
12300 (varpool_variable_node): Likewise.
12301 * tree-emutls.c (create_emultls_var): Update.
12302 (ipa_lower_emutls): Update.
12303 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
12304 (cgraph_reset_node): Reset alias info.
12305 (cgraph_finalize_function): Update.
12306 (fixup_same_cpp_alias_visibility): Move to symtab.c.
12307 (analyze_function): Simplify.
12308 (cgraph_process_same_body_aliases): Simplify.
12309 (analyze_functions): Fixup same body aliases.
12310 (handle_alias_pairs): Simplify.
12311 (assemble_thunk): Update.
12312 (assemble_thunks_and_aliases): Update.
12313 (output_weakrefs): Rewrite.
12314 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
12315 (lto_output_varpool_node): Likewise.
12316 (compute_ltrans_boundary): Remve assert.
12317 (get_alias_symbol): New functoin.
12318 (input_node): Rewrite alias handling.
12319 (input_varpool_node): Likewise.
12320 * ipa-pure-const.c (propagate_pure_const): Fix formating.
12321 * ipa.c (process_references): Handle weakrefs correctly.
12322 (symtab_remove_unreachable_nodes): Likewise.
12323 * trans-mem.c (get_cg_data): Update.
12324 (ipa_tm_create_version_alias): Update.
12325 (ipa_tm_execute): Update.
12326 * symtab.c (dump_symtab_base): Dump aliases.
12327 (verify_symtab_base): Verify aliases.
12328 (symtab_node_availability): New function.
12329 (symtab_alias_ultimate_target): Simplify.
12330 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
12331 handle all the fixup cases.
12332 (symtab_resolve_alias): New function.
12333 * passes.c (ipa_write_summaries): Handle weakrefs.
12334 * varpool.c (varpool_analyze_node): Simplify.
12335 (assemble_aliases): Update.
12336 (varpool_create_variable_alias): Simplify.
12337 (varpool_extra_name_alias): Simplify.
12338 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
12339 (lto_symtab_merge_symbols): ... this one.
12341 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
12344 PR rtl-optimization/57268
12345 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12346 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12348 2013-06-01 Tobias Burnus <burnus@net-b.de>
12350 Partially reverted:
12351 2013-05-31 Tobias Burnus <burnus@net-b.de>
12353 PR middle-end/57073
12354 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
12357 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12359 PR rtl-optimization/57268
12360 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12361 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12363 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12365 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
12366 unordered comparison operators when -fno-trapping-math is in effect
12368 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
12369 and implement unordered comparison operators properly on the e500.
12371 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12373 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
12374 for constant scalar integers.
12375 (simplify_relational_operation_1): Likewise.
12377 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
12379 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
12380 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
12383 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12384 Igor Zamyatin <igor.zamyatin@intel.com>
12386 Silvermont (SLM) architecture performance tuning.
12387 * config/i386/i386.h (enum ix86_tune_indices): Add
12388 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
12389 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
12391 * config/i386/i386.c (initial_ix86_tune_features)
12392 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
12393 (ix86_lea_outperforms): Handle Silvermont tuning.
12394 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
12396 (ix86_use_lea_for_mov): Likewise.
12397 (ix86_avoid_lea_for_addr): Likewise.
12398 (ix86_lea_for_add_ok): Likewise.
12399 (exact_dependency_1): New function.
12400 (exact_store_load_dependency): Likewise.
12401 (ix86_adjust_cost): Handle Silvermont tuning.
12402 (do_reoder_for_imul): Likewise.
12403 (swap_top_of_ready_list): New function.
12404 (ix86_sched_reorder): Changed to handle Silvermont tuning.
12406 * config/i386/i386.md (peepholes that split memory operand in fp
12409 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12411 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
12412 Remove un-necessary braces.
12414 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12416 * config/aarch64/aarch64.c (aarch64_classify_symbol):
12417 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
12419 2013-05-31 Tobias Burnus <burnus@net-b.de>
12421 PR middle-end/57073
12422 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
12424 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12427 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
12428 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
12429 * config/arm/neon.md (iordi3_neon): Remove.
12430 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
12431 * config/arm/predicates.md (imm_for_neon_logic_operand):
12432 Move to earlier in the file.
12433 (neon_logic_op2): Likewise.
12434 (arm_iordi_operand_neon): New predicate.
12436 2013-05-31 Richard Biener <rguenther@suse.de>
12438 PR tree-optimization/57478
12439 PR tree-optimization/57453
12440 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
12443 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12445 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
12446 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
12448 2013-05-30 Tobias Burnus <burnus@net-b.de>
12449 Thomas Koenig <tkoenig@gcc.gnu.org>
12451 PR middle-end/57073
12452 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
12453 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
12455 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
12457 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
12459 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
12461 * target.def (register_usage_leveling_p): New hook.
12462 * targhooks.c (default_register_usage_leveling_p): New.
12463 * targhooks.h (default_register_usage_leveling_p): New prototype.
12464 * lra-assigns.c (register_usage_leveling_p): Use the hook.
12465 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
12466 * doc/tm.texi: Update.
12467 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
12469 2013-05-30 Ian Bolton <ian.bolton@arm.com>
12471 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
12472 (*insv_reg<mode>): New define_insn.
12474 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
12476 PR rtl-optimization/57439
12477 * postreload.c (move2add_valid_value_p): Check that we have
12478 a zero subreg_regno_offset when accessing the register in
12479 the requested mode.
12481 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12482 Igor Zamyatin <igor.zamyatin@intel.com>
12484 Silvermont (SLM) architecture pipeline model, tuning and
12486 * config.gcc: Add slm config options and target.
12488 * config/i386/slm.md: New.
12490 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
12492 * config/i386/i386-c.c (ix86_target_macros_internal): New case
12494 (ix86_target_macros_internal): Likewise.
12496 * config/i386/i386.c (slm_cost): New cost.
12497 (m_SLM): New macro flag.
12498 (initial_ix86_tune_features): Set m_SLM.
12499 (x86_accumulate_outgoing_args): Likewise.
12500 (x86_arch_always_fancy_math_387): Likewise.
12501 (processor_target_table): Add slm cost.
12502 (cpu_names): Add slm cpu name.
12503 (x86_option_override_internal): Set SLM ISA.
12504 (ix86_issue_rate): New case PROCESSOR_SLM.
12505 (ia32_multipass_dfa_lookahead): Likewise.
12506 (fold_builtin_cpu): Add slm.
12508 * config/i386/i386.h (TARGET_SLM): New target macro.
12509 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
12510 (processor_type): Add PROCESSOR_SLM.
12512 * config/i386/i386.md (cpu): Add new value "slm".
12513 (slm.md): Include slm.md.
12515 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
12516 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12518 * config/arm/arm-protos.h: Add and update function protos.
12519 * config/arm/arm.c (use_simple_return_p): New added.
12520 (thumb2_expand_return): Check simple_return flag.
12521 * config/arm/arm.md: Add simple_return and conditional simple_return.
12522 * config/arm/iterators.md: Add iterator for return and simple_return.
12524 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12526 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
12527 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
12528 (arm_emit_vfp_multi_reg_pop): Likewise.
12529 (thumb2_emit_ldrd_pop): Likewise.
12530 (arm_expand_epilogue): Add misc REG_CFA notes.
12531 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
12533 2013-05-29 Lawrence Crowl <crowl@google.com>
12535 * config/arm/t-arm: Update for below.
12537 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
12538 Change type to hash_table. Update dependent calls and types.
12540 * config/i386/t-cygming: Update for below.
12542 * config/i386/t-interix: Update for below.
12544 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12545 Change type to hash_table. Update dependent calls and types.
12546 (i386_find_on_wrapper_list::wrappers): Likewise.
12548 * config/ia64/t-ia64: Update for below.
12550 * config/ia64/ia64.c (bundle_state_table):
12551 Change type to hash_table. Update dependent calls and types.
12553 * config/mips/mips.c (mips_reorg_process_insns::htab):
12554 Change type to hash_table. Update dependent calls and types.
12556 * config/sol2.c (solaris_comdat_htab):
12557 Change type to hash_table. Update dependent calls and types.
12559 * config/t-sol2: Update for above.
12561 2013-05-29 Teresa Johnson <tejohnson@google.com>
12563 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12564 functions are not yet marked as defined.
12566 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12567 Pat Haugen <pthaugen@us.ibm.com>
12568 Peter Bergner <bergner@vnet.ibm.com>
12570 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12574 (vrotl<mode>3): Likewise.
12575 (vashl<mode>3): Likewise.
12576 (vlshr<mode>3): Likewise.
12577 (vashr<mode>3): Likewise.
12579 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12580 support for power8 V2DI builtins.
12582 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12583 power8 V2DI builtins.
12584 (vupkhsw): Likewise.
12585 (vupklsw): Likewise.
12586 (vaddudm): Likewise.
12587 (vminsd): Likewise.
12588 (vmaxsd): Likewise.
12589 (vminud): Likewise.
12590 (vmaxud): Likewise.
12591 (vpkudum): Likewise.
12592 (vpksdss): Likewise.
12593 (vpkudus): Likewise.
12594 (vpksdus): Likewise.
12599 (vsubudm): Likewise.
12600 (vcmpequd): Likewise.
12601 (vcmpgtsd): Likewise.
12602 (vcmpgtud): Likewise.
12603 (vcmpequd_p): Likewise.
12604 (vcmpgtsd_p): Likewise.
12605 (vcmpgtud_p): Likewise.
12606 (vupkhsw): Likewise.
12607 (vupklsw): Likewise.
12608 (vaddudm): Likewise.
12609 (vmaxsd): Likewise.
12610 (vmaxud): Likewise.
12611 (vminsd): Likewise.
12612 (vminud): Likewise.
12613 (vpksdss): Likewise.
12614 (vpksdus): Likewise.
12615 (vpkudum): Likewise.
12616 (vpkudus): Likewise.
12621 (vsubudm): Likewise.
12623 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12624 support for power8 V2DI instructions.
12626 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12627 power8 V2DI instructions. Combine pack and unpack insns to use an
12628 iterator for each mode. Check whether a particular mode supports
12629 Altivec instructions instead of just checking TARGET_ALTIVEC.
12630 (UNSPEC_VPKUWUM): Likewise.
12631 (UNSPEC_VPKSHSS): Likewise.
12632 (UNSPEC_VPKSWSS): Likewise.
12633 (UNSPEC_VPKUHUS): Likewise.
12634 (UNSPEC_VPKSHUS): Likewise.
12635 (UNSPEC_VPKUWUS): Likewise.
12636 (UNSPEC_VPKSWUS): Likewise.
12637 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12638 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12639 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12640 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12641 (UNSPEC_VUPKHSB): Likewise.
12642 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12643 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12644 (UNSPEC_VUPKHSH): Likewise.
12645 (UNSPEC_VUPKLSB): Likewise.
12646 (UNSPEC_VUPKLSH): Likewise.
12648 (VI_char): Likewise.
12649 (VI_scalar): Likewise.
12650 (VI_unit): Likewise.
12652 (VP_small): Likewise.
12653 (VP_small_lc): Likewise.
12654 (VU_char): Likewise.
12655 (add<mode>3): Likewise.
12656 (altivec_vaddcuw): Likewise.
12657 (altivec_vaddu<VI_char>s): Likewise.
12658 (altivec_vadds<VI_char>s): Likewise.
12659 (sub<mode>3): Likewise.
12660 (altivec_vsubcuw): Likewise.
12661 (altivec_vsubu<VI_char>s): Likewise.
12662 (altivec_vsubs<VI_char>s): Likewise.
12663 (altivec_vavgs<VI_char>): Likewise.
12664 (altivec_vcmpbfp): Likewise.
12665 (altivec_eq<mode>): Likewise.
12666 (altivec_gt<mode>): Likewise.
12667 (altivec_gtu<mode>): Likewise.
12668 (umax<mode>3): Likewise.
12669 (smax<mode>3): Likewise.
12670 (umin<mode>3): Likewise.
12671 (smin<mode>3): Likewise.
12672 (altivec_vpkuhum): Likewise.
12673 (altivec_vpkuwum): Likewise.
12674 (altivec_vpkshss): Likewise.
12675 (altivec_vpkswss): Likewise.
12676 (altivec_vpkuhus): Likewise.
12677 (altivec_vpkshus): Likewise.
12678 (altivec_vpkuwus): Likewise.
12679 (altivec_vpkswus): Likewise.
12680 (altivec_vpks<VI_char>ss): Likewise.
12681 (altivec_vpks<VI_char>us): Likewise.
12682 (altivec_vpku<VI_char>us): Likewise.
12683 (altivec_vpku<VI_char>um): Likewise.
12684 (altivec_vrl<VI_char>): Likewise.
12685 (altivec_vsl<VI_char>): Likewise.
12686 (altivec_vsr<VI_char>): Likewise.
12687 (altivec_vsra<VI_char>): Likewise.
12688 (altivec_vsldoi_<mode>): Likewise.
12689 (altivec_vupkhsb): Likewise.
12690 (altivec_vupkhs<VU_char>): Likewise.
12691 (altivec_vupkls<VU_char>): Likewise.
12692 (altivec_vupkhsh): Likewise.
12693 (altivec_vupklsb): Likewise.
12694 (altivec_vupklsh): Likewise.
12695 (altivec_vcmpequ<VI_char>_p): Likewise.
12696 (altivec_vcmpgts<VI_char>_p): Likewise.
12697 (altivec_vcmpgtu<VI_char>_p): Likewise.
12698 (abs<mode>2): Likewise.
12699 (vec_unpacks_hi_v16qi): Likewise.
12700 (vec_unpacks_hi_v8hi): Likewise.
12701 (vec_unpacks_lo_v16qi): Likewise.
12702 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12703 (vec_unpacks_lo_v8hi): Likewise.
12704 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12705 (vec_pack_trunc_v8h): Likewise.
12706 (vec_pack_trunc_v4si): Likewise.
12707 (vec_pack_trunc_<mode>): Likewise.
12709 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12711 (vec_vmaxsd): Likewise.
12712 (vec_vmaxud): Likewise.
12713 (vec_vminsd): Likewise.
12714 (vec_vminud): Likewise.
12715 (vec_vpksdss): Likewise.
12716 (vec_vpksdus): Likewise.
12717 (vec_vpkudum): Likewise.
12718 (vec_vpkudus): Likewise.
12719 (vec_vrld): Likewise.
12720 (vec_vsld): Likewise.
12721 (vec_vsrad): Likewise.
12722 (vec_vsrd): Likewise.
12723 (vec_vsubudm): Likewise.
12724 (vec_vupkhsw): Likewise.
12725 (vec_vupklsw): Likewise.
12727 2013-05-29 Jan Hubicka <jh@suse.cz>
12729 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12730 flags; reorder rest of fields in more consistent way.
12731 (varpool_node): Remove analyzed, finalized and alias.
12732 (cgraph_ndoe): Likewise.
12733 (symtab_alias_ultimate_target): New function.
12734 (cgraph_function_node): Move offline.
12735 (cgraph_reset_node): Declare.
12736 (cgraph_comdat_can_be_unshared_p): Remove.
12737 (varpool_remove_initializer): Declare.
12738 (varpool_first_defined_variable, varpool_next_defined_variable
12739 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12740 (cgraph_function_with_gimple_body_p): Update.
12741 (varpool_all_refs_explicit_p): Update.
12742 (symtab_alias_target): New function.
12743 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12744 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12745 (cgraph_function_or_thunk_node): Simplify using
12746 symtab_alias_ultimate_target.
12747 (varpool_variable_node): Likewise.
12748 * cgraph.c (cgraph_create_function_alias): Update.
12749 (cgraph_add_thunk): Update.
12750 (cgraph_remove_node): Update.
12751 (dump_cgraph_node): Do not dump removed flags.
12752 (cgraph_function_body_availability): Update.
12753 (cgraph_propagate_frequency): Update.
12754 (verify_cgraph_node): Check sanity of local flag.
12755 (cgraph_function_node): Move here from cgraph.h; revamp for
12756 cgraph_function_or_thunk_node.
12757 * lto-symtab.c (lto_varpool_replace_node): Update.
12758 (lto_symtab_resolve_can_prevail_p): Update.
12759 (lto_symtab_merge_cgraph_nodes): Update.
12760 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12761 propagate_constants_accross_call, devirtualization_time_bonus,
12762 ipcp_propagate_stage): Update.
12763 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12764 * ipa-inline-transform.c (clone_inlined_nodes,
12765 preserve_function_body_p): Update.
12766 * ipa-reference.c (propagate): Update.
12767 (write_node_summary_p): Update.
12768 * toplev.c (wrapup_global_declaration_2): Update.
12769 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12770 (analyze_function) ... this one.
12771 (cgraph_process_new_functions): Update.
12772 (cgraph_reset_node): Export.
12773 (cgraph_finalize_function): Update.
12774 (cgraph_add_new_function): Update.
12775 (process_function_and_variable_attributes): Update.
12776 (varpool_finalize_decl): Update.
12777 (symbol_finalized): Remove.
12778 (symbol_finalized_and_needed): Rename to ...
12779 (symbol_defined_and_needed): ... update.
12780 (cgraph_analyze_functions): Update.
12781 (handle_alias_pairs): Update.
12782 (mark_functions_to_output): Update.
12783 (assemble_thunk): Update.
12784 (output_in_order): Update.
12785 (output_weakrefs): Update.
12786 (finalize_compilation_unit): Update.
12787 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12788 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12789 input_node, input_varpool_node): Update.
12790 * dbxout.c (dbxout_expand_expr): Update.
12791 * cgraphclones.c (cgraph_clone_node): Update.
12792 (cgraph_copy_node_for_versioning): Update.
12793 (cgraph_materialize_clone): Update.
12794 (cgraph_materialize_all_clones): Update.
12795 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12796 propagate_pure_const, propagate_nothrow): Update.
12797 * lto-streamer-out.c (lto_output, write_symbol): Update.
12798 * ipa-utils.c (ipa_reverse_postorder): Update.
12799 * ipa-inline.c (can_inline_edge_p): Update.
12800 (update_caller_keys, ipa_inline): Update.
12801 * dwarf2out.c (reference_to_unused,
12802 premark_types_used_by_global_vars_helper): Update.
12803 * tree-eh.c (tree_could_trap_p): Update.
12804 * ipa-split.c (consider_split, execute_split_functions): Update.
12805 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12806 has_addr_references_p): Update; move ahead in file for better
12808 (process_references): Simplify.
12809 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12810 bodies are removed.
12811 (cgraph_comdat_can_be_unshared_p): Make static.
12812 (cgraph_externally_visible_p): Update.
12813 (varpool_externally_visible_p): Update.
12814 (function_and_variable_visibility): Update.
12815 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12816 ipa_tm_mark_force_output_node): Update.
12817 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12818 estimate_edge_devirt_benefit, inline_generate_summary,
12819 inline_write_summary): Update.
12820 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12821 * ipa-prop.c (ipa_compute_jump_functions): Update.
12822 (ipa_print_node_params, ipa_prop_read_section,
12823 ipa_update_after_lto_read, read_replacements_section): Update.
12824 * varasm.c (mark_decl_referenced): Update.
12825 (assemble_alias, dump_tm_clone_pairs): Update.
12826 * tree-inline.c (copy_bb): Update.
12827 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12829 * symtab.c (dump_symtab_base): Print new flags.
12830 (verify_symtab_base): Verify new flags.
12831 (symtab_alias_ultimate_target): New function.
12832 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12833 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12835 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12837 * i386.c (ix86_get_function_versions_dispatcher,
12838 ix86_generate_version_dispatcher_body): Update.
12839 (fold_builtin_cpu): Use varpool_add_new_variable.
12840 * varpool.c (varpool_remove_initializer): Break out from ...
12841 (varpool_remove_node): ... this one.
12842 (dump_varpool_node, varpool_node_for_asm,
12843 cgraph_variable_initializer_availability, varpool_analyze_node,
12844 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12845 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12848 2013-05-29 Jan Hubicka <jh@suse.cz>
12850 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12852 2013-05-29 Easwaran Raman <eraman@google.com>
12854 PR tree-optimization/57442
12855 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12856 when control exits the main loop.
12858 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12860 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12862 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12863 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12864 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12867 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12869 PR tree-optimization/57441
12870 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12871 Don't limit size of incr_vec to number of candidates.
12873 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12875 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12876 and mips16 directories.
12877 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12878 (MULTILIB_DIRNAMES): Ditto.
12879 (MULTILIB_EXCEPTIONS): Add new exceptions.
12880 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12881 (MULTILIB_DIRNAMES): Ditto.
12882 (MULTILIB_EXCEPTIONS): Add new exceptions.
12884 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12885 Marcus Shawcroft <marcus.shawcroft@arm.com>
12887 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12888 SYMBOL_TINY_ABSOLUTE.
12889 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12890 SYMBOL_TINY_ABSOLUTE.
12891 (aarch64_expand_mov_immediate): Likewise.
12892 (aarch64_classify_symbol): Likewise.
12893 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12894 Permit SYMBOL_TINY_ABSOLUTE.
12895 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12897 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12898 Marcus Shawcroft <marcus.shawcroft@arm.com>
12900 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12901 Refactor if/switch. Replace gcc_assert with if.
12903 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12905 * config/i386/i386.c (initial_ix86_tune_features): Enable
12906 FP Reassociation for AMD bdver1 and bdver2.
12908 2013-05-29 Martin Jambor <mjambor@suse.cz>
12910 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12911 and IMAGPART_EXPR do not occur within other handled_components.
12913 2013-05-29 Richard Biener <rguenther@suse.de>
12915 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12916 access on whether the use is in the BB we currently try to
12918 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12919 vectorize to vect_bb_slp_scalar_cost.
12921 2013-05-29 Richard Biener <rguenther@suse.de>
12923 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12924 computing scalar cost offsetted by stmts that are kept live
12926 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12927 for computation of scalar cost.
12929 2013-05-28 Steve Ellcey <sellcey@mips.com>
12931 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12933 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12935 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12936 array notation built-in reduction functions.
12937 * doc/passes.texi (Passes): Added documentation about changes done
12939 * doc/invoke.texi (C Dialect Options): Added documentation about
12940 the -fcilkplus flag.
12941 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12942 (BUILTINS_DEF): Depend on cilkplus.def.
12943 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12944 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12945 * cilkplus.def: New file.
12947 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12949 PR rtl-optimization/57439
12950 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12952 2013-05-28 Easwaran Raman <eraman@google.com>
12954 PR tree-optimization/57337
12955 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12956 (find_insert_point): Correctly identify the insertion point
12957 when two statements with the same UID is compared.
12959 2013-05-28 Richard Biener <rguenther@suse.de>
12961 PR tree-optimization/56787
12962 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12963 from the list of data references.
12964 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12966 (vect_analyze_loop_operations): Likewise.
12967 (vect_transform_loop): Remove clobbers.
12969 2013-05-28 Martin Jambor <mjambor@suse.cz>
12971 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12972 and REALPART_EXPRs have scalar type.
12974 2013-05-28 Richard Biener <rguenther@suse.de>
12976 PR tree-optimization/57411
12977 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12979 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12980 virtual operand propagation.
12982 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12984 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12985 destination register for bmasksi_vis.
12986 (vector_init_bshuffle): Likewise.
12987 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12989 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12991 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12992 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12993 mode if the instruction isn't available in the original mode.
12994 * config/sparc/sparc.opt (mfix-ut699): New option.
12995 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12996 (divdf3): Turn into expander.
12997 (divdf3_nofix): New insn.
12998 (divdf3_fix): Likewise.
12999 (divsf3): Disable if -mfix-ut699.
13000 (sqrtdf2): Turn into expander.
13001 (sqrtdf2_nofix): New insn.
13002 (sqrtdf2_fix): Likewise.
13003 (sqrtsf2): Disable if -mfix-ut699.
13005 2013-05-27 Richard Biener <rguenther@suse.de>
13007 PR middle-end/57412
13008 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
13009 block for the new loop.
13011 2013-05-27 Richard Biener <rguenther@suse.de>
13013 PR tree-optimization/57343
13014 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13015 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13016 (number_of_iterations_cond): Do not build the folded tree.
13018 2013-05-27 Richard Biener <rguenther@suse.de>
13021 PR middle-end/57381
13022 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13023 OEP_CONSTANT_ADDRESS_OF retained.
13025 PR tree-optimization/57417
13026 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13027 for unchanged base.
13028 (set_ssa_val_to): Compare addresses using
13029 get_addr_base_and_unit_offset.
13031 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13033 PR rtl-optimization/56833
13034 * postreload.c (move2add_record_mode): New function.
13035 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13036 (move2add_use_add2_insn): Use move2add_record_sym_value.
13037 (move2add_use_add3_insn): Likewise.
13038 (reload_cse_move2add): Use move2add_valid_value_p and
13039 move2add_record_mode. Invalidate call-clobbered and REG_INC
13040 affected regs by setting reg_mode to VOIDmode.
13041 (move2add_note_store): Don't pretend the inside of a SUBREG is
13042 the actual destination. Invalidate single/leading registers by
13043 setting reg_mode to VOIDmode.
13044 Use move2add_record_sym_value, move2add_valid_value_p and
13045 move2add_record_mode.
13047 2013-05-27 Richard Biener <rguenther@suse.de>
13049 PR tree-optimization/57396
13050 * tree-affine.c (double_int_constant_multiple_p): Properly
13051 return false for val == 0 and div != 0.
13053 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13055 * config/mips/mips.h: Use #elif in preprocessor conditions.
13057 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13060 * config/mips/constraints.md (kl): New constraint.
13061 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13062 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13063 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13064 from LO for MIPS16.
13065 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13067 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13070 * config/mips/mips.c (mips_can_inline_p): New function.
13071 (TARGET_CAN_INLINE_P): Define.
13073 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13075 * sched-int.h (ds_t, dw_t): Make unsigned int.
13076 Fix documentation that describes how all the ds_t bits are used.
13077 Reserve the last bit for delayed-branch scheduling.
13078 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13079 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13080 (gen_dep_weak_1): Remove prototype.
13081 * sched-deps.c (get_dep_weak_1): Make static.
13082 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13083 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13084 * doc/tm.texi: Regenerate.
13085 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13087 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13090 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13092 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13093 Sandra Loosemore <sandra@codesourcery.com>
13095 * config.gcc (powerpc-*): Allow native for with-cpu.
13097 2013-05-24 Jeff Law <law@redhat.com>
13099 PR tree-optimization/57124
13100 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13101 conversion feeding a condition if the range has an overflow
13102 if -fstrict-overflow. Add warnings for when we do make the
13105 2013-05-24 Dehao Chen <dehao@google.com>
13107 * tree-cfg.c (locus_discrim_map): Fix the typo.
13108 (locus_discrim_hasher): Likewise.
13109 (locus_discrim_hasher::hash): Likewise.
13110 (locus_discrim_hasher::equal): Likewise.
13112 2013-05-24 Martin Jambor <mjambor@suse.cz>
13114 PR tree-optimization/57294
13115 * cgraph.h (ipa_record_stmt_references): Declare.
13116 * cgraphbuild.c (ipa_record_stmt_references): New function.
13117 (build_cgraph_edges): Use ipa_record_stmt_references.
13118 (rebuild_cgraph_edges): Likewise.
13119 (cgraph_rebuild_references): Likewise.
13120 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13121 associated with the old statement and build references from the
13122 newly built statements.
13123 * ipa-ref.c (ipa_remove_stmt_references): New function.
13124 * ipa-ref.h (ipa_remove_stmt_references): Declare.
13126 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
13128 * lra-constraints.c (emit_spill_move): Use smaller mode for
13130 (check_and_process_move): Consider mem-reg moves for secondary
13132 (curr_insn_transform): Don't lose insns emitted before for
13133 secondary memory moves.
13134 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
13135 reg set up in the current insn.
13137 2013-05-24 Dehao Chen <dehao@google.com>
13139 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
13141 (locus_descrim_hasher::equal): Likewise.
13142 (build_gimple_cfg): New discriminator assignment algorithm.
13143 (make_edges): Likewise.
13144 (next_discriminator_for_locus): Likewise.
13145 (same_line_p): Likewise.
13146 (assign_discriminators): Likewise.
13147 (make_cond_expr_edges): Likewise.
13148 (make_gimple_switch_edges): Likewise.
13149 (make_goto_expr_edges): Likewise.
13150 (make_gimple_asm_edges): Likewise.
13152 2013-05-24 Ian Bolton <ian.bolton@arm.com>
13154 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
13155 X format specifier to only display bottom 16 bits.
13156 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
13157 immediate to match for operand 2, since it will be masked.
13159 2013-05-24 Richard Biener <rguenther@suse.de>
13161 PR tree-optimization/57287
13162 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
13163 all SSA names that occur in abnormal PHIs.
13165 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
13168 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
13169 that index is not negative.
13171 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
13173 PR rtl-optimization/55177
13174 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
13175 (simplify_byte_swapping_operation): New.
13176 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
13177 (simplify_relational_operation_1): Deal with BSWAP.
13179 2013-05-23 Richard Henderson <rth@redhat.com>
13182 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
13183 (ix86_reorg): Call it.
13185 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
13188 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
13189 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
13190 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
13192 2013-05-23 Christian Bruel <christian.bruel@st.com>
13195 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
13197 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13198 Marcus Shawcroft <marcus.shawcroft@arm.com>
13200 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
13201 * config/aarch64/constraints.md (Usa): Remove.
13202 * doc/md.texi (AArch64 Usa): Remove.
13204 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13205 Marcus Shawcroft <marcus.shawcroft@arm.com>
13207 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
13208 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
13209 * config/aarch64/predicates.md (aarch64_const_address): Remove.
13210 (aarch64_mov_operand): Use aarch64_mov_operand_p.
13212 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
13214 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
13215 instruction (AdvSIMD).
13216 * config/aarch64/aarch64-builtins.c
13217 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
13218 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
13220 2013-05-23 Martin Jambor <mjambor@suse.cz>
13222 PR middle-end/57347
13223 * tree.h (contains_bitfld_component_ref_p): Declare.
13224 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
13225 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
13227 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
13228 not access a bit-field. Assert all final offsets are byte-aligned.
13230 2013-05-23 Richard Biener <rguenther@suse.de>
13232 PR tree-optimization/57380
13233 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
13234 least one invariant or re-used load.
13235 * passes.c (init_optimization_passes): Move pass_phiprop before
13238 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
13240 * config/aarch64/aarch64-simd.md
13241 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
13243 2013-05-23 Richard Biener <rguenther@suse.de>
13245 PR middle-end/57381
13246 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13247 OEP_CONSTANT_ADDRESS_OF retained.
13249 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13251 PR middle-end/57344
13252 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13253 don't lower unit. Handle unit not being always BITS_PER_WORD.
13255 2013-05-23 Richard Biener <rguenther@suse.de>
13257 PR rtl-optimization/57341
13258 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13259 instead of true_dependence.
13261 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13263 * bb-reorder.c (branch_threshold): Make const.
13264 (exec_threshold): Ditto.
13266 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13267 Pat Haugen <pthaugen@us.ibm.com>
13268 Peter Bergner <bergner@vnet.ibm.com>
13270 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
13271 documentation for the power8 crypto builtins.
13273 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
13275 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
13276 macros for defining power8 builtin functions.
13277 (BU_P8V_AV_2): Likewise.
13278 (BU_P8V_AV_P): Likewise.
13279 (BU_P8V_VSX_1): Likewise.
13280 (BU_P8V_OVERLOAD_1): Likewise.
13281 (BU_P8V_OVERLOAD_2): Likewise.
13282 (BU_CRYPTO_1): Likewise.
13283 (BU_CRYPTO_2): Likewise.
13284 (BU_CRYPTO_3): Likewise.
13285 (BU_CRYPTO_OVERLOAD_1): Likewise.
13286 (BU_CRYPTO_OVERLOAD_2): Likewise.
13287 (XSCVSPDP): Fix typo, point to the correct instruction.
13288 (VCIPHER): Add power8 crypto builtins.
13289 (VCIPHERLAST): Likewise.
13290 (VNCIPHER): Likewise.
13291 (VNCIPHERLAST): Likewise.
13292 (VPMSUMB): Likewise.
13293 (VPMSUMH): Likewise.
13294 (VPMSUMW): Likewise.
13295 (VPERMXOR_V2DI): Likewise.
13296 (VPERMXOR_V4SI: Likewise.
13297 (VPERMXOR_V8HI: Likewise.
13298 (VPERMXOR_V16QI: Likewise.
13299 (VSHASIGMAW): Likewise.
13300 (VSHASIGMAD): Likewise.
13301 (VPMSUM): Likewise.
13302 (VPERMXOR): Likewise.
13303 (VSHASIGMA): Likewise.
13305 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13306 __CRYPTO__ if the crypto instructions are available.
13307 (altivec_overloaded_builtins): Add support for overloaded power8
13310 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
13311 support for power8 crypto builtins.
13312 (builtin_function_type): Likewise.
13313 (altivec_init_builtins): Add support for builtins that take vector
13314 long long (V2DI) arguments.
13316 * config/rs6000/crypto.md: New file, define power8 crypto
13319 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13320 Pat Haugen <pthaugen@us.ibm.com>
13321 Peter Bergner <bergner@vnet.ibm.com>
13323 * doc/invoke.texi (Option Summary): Add power8 options.
13324 (RS/6000 and PowerPC Options): Likewise.
13326 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
13327 constraints.md instead of rs6000.h. Reorder w* constraints. Add
13328 wm, wn, wr documentation.
13330 * config/rs6000/constraints.md (wm): New constraint for VSX
13331 registers if direct move instructions are enabled.
13332 (wn): New constraint for no registers.
13333 (wq): New constraint for quad word even GPR registers.
13334 (wr): New constraint if 64-bit instructions are enabled.
13335 (wv): New constraint if power8 vector instructions are enabled.
13336 (wQ): New constraint for quad word memory locations.
13338 * config/rs6000/predicates.md (const_0_to_15_operand): New
13339 constraint for 0..15 for crypto instructions.
13340 (gpc_reg_operand): If VSX allow registers in VSX registers as well
13341 as GPR and floating point registers.
13342 (int_reg_operand): New predicate to match only GPR registers.
13343 (base_reg_operand): New predicate to match base registers.
13344 (quad_int_reg_operand): New predicate to match even GPR registers
13345 for quad memory operations.
13346 (vsx_reg_or_cint_operand): New predicate to allow vector logical
13347 operations in both GPR and VSX registers.
13348 (quad_memory_operand): New predicate for quad memory operations.
13349 (reg_or_indexed_operand): New predicate for direct move support.
13351 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
13352 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
13353 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
13354 (POWERPC_MASKS): Add power8 options.
13355 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
13358 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
13359 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
13361 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
13362 (-mpower8-fusion): New power8 options.
13363 (-mpower8-fusion-sign): Likewise.
13364 (-mpower8-vector): Likewise.
13365 (-mcrypto): Likewise.
13366 (-mdirect-move): Likewise.
13367 (-mquad-memory): Likewise.
13369 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
13370 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
13372 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
13373 (rs6000_debug_vector_unit): Add p8_vector.
13374 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
13375 definitions. Also print fusion state.
13376 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
13377 (rs6000_builtin_mask_calculate): Add power8 builtin support.
13378 (rs6000_option_override_internal): Add support for power8.
13379 (rs6000_common_init_builtins): Add debugging for skipped builtins
13380 if -mdebug=builtin.
13381 (rs6000_adjust_cost): Add power8 support.
13382 (rs6000_issue_rate): Likewise.
13383 (insn_must_be_first_in_group): Likewise.
13384 (insn_must_be_last_in_group): Likewise.
13385 (force_new_group): Likewise.
13386 (rs6000_register_move_cost): Likewise.
13387 (rs6000_opt_masks): Likewise.
13389 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
13390 power8 capable assembler, default to power7 options.
13391 (TARGET_DIRECT_MOVE): Likewise.
13392 (TARGET_CRYPTO): Likewise.
13393 (TARGET_P8_VECTOR): Likewise.
13394 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
13395 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
13396 (VECTOR_MEM_P8_VECTOR_P): Likewise.
13397 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
13398 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
13399 (TARGET_XSCVDPSPN): Likewise.
13400 (TARGET_XSCVSPDPN): Likewsie.
13401 (TARGET_SYNC_HI_QI): Likewise.
13402 (TARGET_SYNC_TI): Likewise.
13403 (MASK_CRYPTO): Likewise.
13404 (MASK_DIRECT_MOVE): Likewise.
13405 (MASK_P8_FUSION): Likewise.
13406 (MASK_P8_VECTOR): Likewise.
13407 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
13408 temporary used by some of the direct move instructions to get two FP
13409 temporary registers does not force creation of a stack frame.
13410 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
13411 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
13412 that any VSX registers are tieable, even if they are also an
13413 Altivec vector mode.
13414 (r6000_reg_class_enum): Add wm, wr, wv constraints.
13415 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
13416 (RS6000_BTM_CRYPTO): Likewise.
13417 (RS6000_BTM_COMMON): Likewise.
13419 * config/rs6000/rs6000.md (cpu attribute): Add power8.
13420 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
13421 (enum rs6000_vector): Add power8 vector support.
13423 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13427 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
13428 (any_sibcall_could_use_r3): this and handle indirect calls.
13429 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
13431 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13433 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
13435 2013-05-22 Richard Biener <rguenther@suse.de>
13437 PR middle-end/57349
13438 * profile.c (branch_prob): Do not split blocks that are
13439 abnormally receiving from ECF_RETURNS_TWICE functions.
13441 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13443 * recog.c (offsettable_address_addr_space_p): Fix calculation of
13444 address mode. Move pointer mode initialization to the same place.
13446 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13448 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
13449 while it has any effect.
13451 2013-05-21 Easwaran Raman <eraman@google.com>
13453 PR tree-optimization/57322
13454 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
13455 UID of the statement added to the BB to be 1.
13457 2013-05-21 Jakub Jelinek <jakub@redhat.com>
13459 PR tree-optimization/57331
13460 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
13461 of conversion from pointer type to integral type with integer.
13463 2013-05-21 Martin Jambor <mjambor@suse.cz>
13466 * ipa-prop.c (ipa_read_node_info): Process param_used and
13467 controlled_uses in the same order as when writing.
13469 2013-05-21 Magnus Granberg <baldrick@free.fr>
13472 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
13474 2013-05-21 Richard Biener <rguenther@suse.de>
13476 PR tree-optimization/57318
13477 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
13478 estimate stmts with side-effects as likely eliminated.
13480 2013-05-21 Richard Biener <rguenther@suse.de>
13482 PR tree-optimization/57330
13483 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
13484 preserve the call stmts fntype.
13486 2013-05-21 Richard Biener <rguenther@suse.de>
13488 PR tree-optimization/57303
13489 * tree-ssa-sink.c (statement_sink_location): Improve killing
13490 stmt detection and properly handle self-assignments.
13492 2013-05-21 Christian Bruel <christian.bruel@st.com>
13494 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
13495 spanning registers. LEAF_REG_REMAP is supported only for contiguous
13496 registers. Set register size out of the PARALLEL loop.
13498 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
13501 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
13502 (*fmasf4, *fmasf4_media): New insns.
13504 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13506 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
13507 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
13508 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
13509 (mips_idiv_insns): Update the comments to say that the returned
13510 instruction counts are in units of BASE_INSN_LENGTH.
13511 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
13512 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
13513 using 2 rather than 4 as the length of indirect MIPS16 and
13514 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
13515 length of a NOP. Don't divide MIPS16 lengths by 2.
13516 (mips16_split_long_branches): Assume a branch is long if the
13517 length is greater than 4 rather than 8.
13518 * config/mips/mips.md (length): Give MIPS16 lengths directly,
13519 rather than multiplying them by 2. Multiply instruction counts
13520 by BASE_INSN_LENGTH rather than 4.
13521 (*jump_mips16, tls_get_tp_mips16_<mode>)
13522 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
13524 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13526 * config/mips/mips.md (extended_mips16): Remove branch case.
13527 (length): Remove duplicated extended_mips16 test.
13529 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13531 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
13533 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
13535 * recog.h (Recog_data): Rename to...
13536 (recog_data_d): ...this.
13537 (recog_data): Update accordingly.
13538 * recog.c (recog_data): Likewise.
13539 * reload.c (save_recog_data): Likewise.
13540 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13541 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13543 2013-05-17 Julian Brown <julian@codesourcery.com>
13545 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13546 found in a REG_EQUAL note, invalidate it.
13548 2013-05-17 Easwaran Raman <eraman@google.com>
13550 * tree-ssa-reassoc.c (find_insert_point): New function.
13551 (insert_stmt_after): Likewise.
13552 (get_def_stmt): Likewise.
13553 (ensure_ops_are_available): Likewise.
13554 (not_dominated_by): Likewise.
13555 (rewrite_expr_tree): Do not move statements beyond what is
13556 necessary. Remove call to swap_ops_for_binary_stmt...
13557 (reassociate_bb): ... and move it here.
13558 (build_and_add_sum): Assign UIDs for new statements.
13559 (linearize_expr): Likewise.
13560 (do_reassoc): Renumber gimple statement UIDs.
13562 2013-05-17 Jan Hubicka <jh@suse.cz>
13564 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13566 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13567 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13569 (output_weakrefs): Update.
13571 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13572 Martin Jambor <mjambor@suse.cz>
13574 PR middle-end/57276
13575 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13576 value that corresponds to the given aggval is found in values vector.
13578 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13580 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13581 sse, sse2, sse3, ssse3 and sse4a flags to options.
13583 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13585 * gengtype-state.c: (s_expr_writer): New class, to handle
13586 prettifying of output layout of s-expressions.
13587 (state_writer): New class, to write out gtype.state.
13588 (state_written_type_count): Move this variable into member data of
13590 (s_expr_writer::s_expr_writer): New code: constructor for new class
13591 (state_writer::state_writer(): ditto
13592 (s_expr_writer::write_new_line): New function
13593 (s_expr_writer::write_any_indent): ditto
13594 (s_expr_writer::begin_s_expr): ditto
13595 (s_expr_writer::end_s_expr): ditto
13596 (write_state_fileloc): convert to method of state_writer...
13597 (state_writer:: write_state_fileloc): ...and use methods of
13598 s_expr_writer to write indentation into the gtype.state output file
13599 to visually represent the hierarchical structure of the list
13601 (write_state_fields): ditto, renaming to...
13602 (state_writer::write_state_fields)
13603 (write_state_a_string): ditto, renaming to...
13604 (state_writer::write_state_a_string)
13605 (write_state_string_option): ditto, renaming to...
13606 (state_writer::write_state_string_option)
13607 (write_state_type_option): ditto, renaming to...
13608 (state_writer::write_state_type_option)
13609 (write_state_nested_option): ditto, renaming to...
13610 (state_writer::write_state_nested_option)
13611 (write_state_option): ditto, renaming to...
13612 (state_writer::write_state_option)
13613 (write_state_options): ditto, renaming to...
13614 (state_writer::write_state_options)
13615 (write_state_lang_bitmap): ditto, renaming to...
13616 (state_writer::write_state_lang_bitmap)
13617 (write_state_version): ditto, renaming to...
13618 (state_writer::write_state_version)
13619 (write_state_scalar_type): ditto, renaming to...
13620 (state_writer::write_state_scalar_type)
13621 (write_state_string_type): ditto, renaming to...
13622 (state_writer::write_state_string_type)
13623 (write_state_undefined_type): ditto, renaming to...
13624 (state_writer::write_state_undefined_type)
13625 (write_state_struct_union_type): ditto, renaming to...
13626 (state_writer::write_state_struct_union_type)
13627 (write_state_struct_type): ditto, renaming to...
13628 (state_writer::write_state_struct_type)
13629 (write_state_user_struct_type): ditto, renaming to...
13630 (state_writer::write_state_user_struct_type)
13631 (write_state_lang_struct_type): ditto, renaming to...
13632 (state_writer::write_state_lang_struct_type)
13633 (write_state_param_struct_type): ditto, renaming to...
13634 (state_writer::write_state_param_struct_type)
13635 (write_state_pointer_type): ditto, renaming to...
13636 (state_writer::write_state_pointer_type)
13637 (write_state_array_type): ditto, renaming to...
13638 (state_writer::write_state_array_type)
13639 (write_state_gc_used): ditto, renaming to...
13640 (state_writer::write_state_gc_used)
13641 (write_state_common_type_content): ditto, renaming to...
13642 (state_writer::write_state_common_type_content)
13643 (write_state_type): ditto, renaming to...
13644 (state_writer::write_state_type)
13645 (write_state_pair_list): ditto, renaming to...
13646 (state_writer::write_state_pair_list)
13647 (write_state_pair): ditto, renaming to...
13648 (state_writer::write_state_pair)
13649 (write_state_typedefs): ditto, renaming to...
13650 (state_writer::write_state_typedefs)
13651 (write_state_structures): ditto, renaming to...
13652 (state_writer::write_state_structures)
13653 (write_state_param_structs): ditto, renaming to...
13654 (state_writer::write_state_param_structs)
13655 (write_state_variables): ditto, renaming to...
13656 (state_writer::write_state_variables)
13657 (write_state_srcdir): ditto, renaming to...
13658 (state_writer::write_state_srcdir)
13659 (write_state_files_list): ditto, renaming to...
13660 (state_writer::write_state_files_list)
13661 (write_state_languages): ditto, renaming to...
13662 (state_writer::write_state_languages)
13663 (write_state): create a state_writer instance and use it when
13664 writing out the state file
13666 2013-05-17 Mike Stump <mikestump@comcast.net>
13668 PR rtl-optimization/57304
13669 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13670 accessing DF_REF_REAL_LOC.
13672 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13674 PR rtl-optimization/57281
13675 PR rtl-optimization/57300
13676 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13677 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13678 what the other splitter did if the registers are dead.
13680 2013-05-17 Richard Biener <rguenther@suse.de>
13682 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13685 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13687 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13690 2013-05-17 Marek Polacek <polacek@redhat.com>
13692 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13693 length when doing non-zero store of storing '\0' to '\0'.
13695 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13697 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13698 vect_external_def oprnd1 with loop_vinfo, try to emit
13699 optional cast, negation and and stmts on the loop preheader
13700 edge instead of into the pattern def seq.
13702 PR tree-optimization/57051
13703 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13704 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13706 2013-05-16 Nick Clifton <nickc@redhat.com>
13708 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13709 (rl78_is_naked_func): New function.
13710 (rl78_expand_prologue): Skip prologue generation for naked functions.
13711 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13712 * doc/extend.texi (naked): Add RL78 to the list of processors
13713 that supports this attribute.
13715 2013-05-16 Jeff Law <law@redhat.com>
13717 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13719 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13721 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13722 cache parameters using detect_caches_amd also for CYRIX,
13723 NSC and TM2 signatures.
13725 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13726 Dzianis Kahanovich <mahatma@eu.by>
13730 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13731 VIA/Centaur processors and determine their cache parameters
13732 using detect_caches_amd.
13734 2013-05-16 Teresa Johnson <tejohnson@google.com>
13736 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13737 (rtl_verify_edges): New function.
13738 (rtl_verify_bb_insns): Ditto.
13739 (rtl_verify_bb_pointers): Ditto.
13740 (rtl_verify_bb_insn_chain): Ditto.
13741 (rtl_verify_fallthru): Ditto.
13742 (rtl_verify_bb_layout): Ditto.
13743 (rtl_verify_flow_info_1): Outline checks into new functions.
13744 (rtl_verify_flow_info): Ditto.
13746 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13748 * cfghooks.c (copy_bbs): Add update_dominance argument.
13749 * cfghooks.h (copy_bbs): Update prototype.
13750 * tree-cfg.c (gimple_duplicate_sese_region):
13751 Add update_dominance argument.
13752 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13753 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13754 gimple_duplicate_sese_region call.
13755 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13756 Update copy_bbs call.
13757 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13758 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13760 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13762 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13763 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13764 vect_recog_rotate_pattern.
13765 (vect_recog_rotate_pattern): New function.
13767 2013-05-16 Jason Merrill <jason@redhat.com>
13769 * Makefile.in (LLINKER): New variable.
13770 (mostlyclean): Remove link mutex.
13771 * configure.ac: Handle --enable-link-mutex.
13772 * lock-and-run.sh: New script.
13774 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13777 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13780 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13782 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13784 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13786 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13787 * config/arm/arm.c (next_consecutive_mem): New function.
13788 (gen_movmem_ldrd_strd): Likewise.
13789 * config/arm/arm.md (movmemqi): Update condition and code.
13790 (unaligned_loaddi, unaligned_storedi): New patterns.
13792 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13794 * config.gcc: Obsolete *-*-solaris2.9*.
13795 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13797 2013-05-16 Richard Biener <rguenther@suse.de>
13799 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13800 earlier, after GRAPHITE transforms and IV canonicalization.
13802 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13804 * omp-low.c (extract_omp_for_data): For collapsed loops,
13805 if at least one of the loops is known at compile time to
13806 iterate zero times, set count to 0.
13807 (expand_omp_regimplify_p): New function.
13808 (expand_omp_for_generic): For collapsed loops, if at least
13809 one of the loops isn't known to iterate at least once,
13810 add runtime check with setting count to 0.
13811 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13812 For unsigned types if it isn't known at compile time that
13813 the loop will iterate at least once, add runtime check to bypass
13814 the whole loop if initial condition isn't true.
13816 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13818 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13820 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13822 PR middle-end/57286
13823 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13824 transformations to avoid an infinite loop.
13826 2013-05-16 Marek Polacek <polacek@redhat.com>
13828 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13830 2013-05-15 Leif Ekblad <leif@rdos.net>
13832 * config/i386/i386.c (ix86_decompose_address): Use
13833 DEFAULT_TLS_SEG_REG to access TLS segment register.
13834 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13835 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13836 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13838 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13841 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13842 sibling calls to functions that would normally be lazily bound,
13843 unless $gp is call-clobbered.
13845 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13847 * config/i386/i386.c (ix86_option_override_internal): Update
13848 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13849 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13850 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13851 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13853 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13855 2013-05-15 Andreas Schwab <schwab@suse.de>
13857 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13858 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13861 2013-05-15 Teresa Johnson <tejohnson@google.com>
13863 * loop-unroll.c (report_unroll_peel): Check decision before
13864 emitting unroll/peel message.
13866 2013-05-15 Teresa Johnson <tejohnson@google.com>
13868 * function.h (has_bb_partition): New rtl_data flag.
13869 (bb_reorder_complete): Ditto.
13870 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13871 instead of flag_reorder_blocks_and_partition.
13872 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13873 with some enhancements.
13874 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13875 * bb-reorder.c (connect_traces): Check for has_bb_partition
13876 instead of flag_reorder_blocks_and_partition.
13877 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13878 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13879 verify_hot_cold_block_grouping.
13880 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13882 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13885 * config/arm/predicates.md (call_insn_operand): New predicate.
13886 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13887 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13888 if insn is not a tail call.
13889 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13891 * config/arm/arm.h (enum reg_class): New caller save register class.
13892 (REG_CLASS_NAMES): Likewise.
13893 (REG_CLASS_CONTENTS): Likewise.
13894 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13897 2013-05-15 Richard Biener <rguenther@suse.de>
13899 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13900 of MSG_OPTIMIZED_LOCATIONS.
13901 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13902 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13904 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13905 of MSG_OPTIMIZED_LOCATIONS.
13906 (execute_vect_slp): Likewise.
13907 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13908 (vect_create_cond_for_alias_checks): Likewise.
13909 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13910 (vect_recog_widen_mult_pattern): Likewise.
13911 (vect_recog_widen_sum_pattern): Likewise.
13912 (vect_recog_over_widening_pattern): Likewise.
13913 (vect_recog_widen_shift_pattern): Likewise.
13914 (vect_recog_vector_vector_shift_pattern): Likewise.
13915 (vect_recog_divmod_pattern): Likewise.
13916 (vect_recog_mixed_size_cond_pattern): Likewise.
13917 (vect_recog_bool_pattern): Likewise.
13918 (vect_pattern_recog_1): Likewise.
13920 2013-05-15 Martin Jambor <mjambor@suse.cz>
13922 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13923 non-functions to builtin_unreachable.
13924 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13925 correct when new direct edges were discovered.
13927 2013-05-15 Martin Jambor <mjambor@suse.cz>
13929 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13930 header, print symbol order instead of node uid, print more information
13931 about indirect edge targets.
13932 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13934 (ipa_make_edge_direct_to_target): Likewise.
13935 (remove_described_reference): Likewise.
13936 (propagate_controlled_uses): Likewise.
13937 (ipa_print_node_params): Also print symbol order.
13938 (ipcp_transform_function): Print symbol order instead of node uids.
13939 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13940 (cgraph_get_create_real_symbol_node): Likewise.
13941 * ipa-cp.c (print_lattice): Likewise.
13942 (print_all_lattices): Likewise.
13943 (determine_versionability): Likewise.
13944 (initialize_node_lattices): Likewise.
13945 (estimate_local_effects): Likewise.
13946 (update_profiling_info): Likewise.
13947 (create_specialized_node): Likewise.
13948 (perhaps_add_new_callers): Likewise.
13949 (decide_about_value): Likewise.
13950 (decide_whether_version_node): Likewise.
13951 (identify_dead_nodes): Likewise.
13952 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13953 (dump_inline_summary): Likewise.
13954 (estimate_node_size_and_time): Likewise.
13955 (inline_analyze_function): Likewise.
13956 * ipa-inline.c (report_inline_failed_reason): Likewise.
13957 (want_early_inline_function_p): Likewise.
13958 (edge_badness): Likewise.
13959 (update_edge_key): Likewise.
13960 (inline_small_functions): Likewise. Add dumping of order to two other
13962 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13963 instead of node uids.
13964 (propagate_pure_const): Likewise.
13965 (propagate_pure_const): Likewise.
13966 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13967 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13968 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13970 * tree-pretty-print.c (dump_function_header): Likewise.
13971 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13972 Print symbol order instead of node uids.
13974 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13976 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13977 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13979 2013-05-15 Richard Biener <rguenther@suse.de>
13981 PR tree-optimization/57275
13982 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13983 return value for fail to do runtime alias checks for gather loads.
13985 2013-05-15 Jan Hubicka <jh@suse.cz>
13989 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13990 weakrefs are not external.
13991 (lto_symtab_merge_decls): Fix thinko when dealing with
13992 non-lto_symtab decls.
13993 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13994 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13995 * varpool.c (dump_varpool_node): Dump more flags.
13997 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13999 * config/i386/i386.c (processor_alias_table): Add instruction
14000 FSGSBASE for AMD bdver3 architecture.
14002 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14004 * tree.c (warn_deprecated_use): Print file:line using locus color.
14005 * diagnostic.c (diagnostic_report_current_module): Print file:line
14006 and file:line:column using locus color.
14008 2013-05-14 Mike Stump <mikestump@comcast.net>
14010 * gdbinit.in: Add __null.
14012 2013-05-14 Mike Stump <mikestump@comcast.net>
14014 * recog.h: Rename struct recog_data to Recog_data.
14015 * recog.c: Likewise.
14016 * reload.c (can_reload_into): Likewise.
14017 * config/picochip/picochip.c: Likewise.
14019 2013-05-14 Mike Stump <mikestump@comcast.net>
14021 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14023 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14025 * resource.h (struct resources): Remove unch_memory member.
14026 (CLEAR_RESOURCE): Don't clear unch_memory.
14027 * resource.c (mark_referenced_resources): Don't set it.
14028 (mark_set_resources): Likewise.
14029 (mark_target_live_regs): Don't clear it.
14030 (init_resource_info): Likewise.
14031 * reorg.c (resource_conflicts_p): Don't compare it.
14032 (redundant_insn): Don't set it.
14034 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14036 * emit-rtl.c (next_label): Remove unused function.
14037 (skip_consecutive_labels, link_cc0_insns): Move to ...
14038 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14039 only place where these functions are used, and make them static.
14041 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14043 * fold-const.c (fold_negate_expr): Handle vectors.
14044 (fold_truth_not_expr): Make it static.
14045 (fold_invert_truthvalue): New static function.
14046 (invert_truthvalue_loc): Handle vectors. Do not call
14047 fold_truth_not_expr directly.
14048 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14049 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14050 (fold_comparison): Handle vector constants.
14051 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14052 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14053 * tree.h (fold_truth_not_expr): Remove declaration.
14055 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14057 * config/aarch64/aarch64-simd.md
14058 (aarch64_vcond_internal<mode>): Rename to...
14059 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14060 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14061 float modes. Clarify all iterator modes.
14062 (vcond<mode><mode>): Use new name for vcond expanders.
14063 (vcond<v_cmp_result><mode>): Likewise.
14064 (vcondu<mode><mode>: Likewise.
14065 * config/aarch64/iterators.md (VDQF_COND): New.
14067 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14070 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14071 variable for the shift amount. Check that we shift by non-negative
14074 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14077 * config/arm/arm.h (EPILOGUE_USES): Only return true
14078 for LR_REGNUM after epilogue_completed.
14080 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14082 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14083 is error_mark_node.
14085 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14088 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14089 and Solaris 11+/x86 with gld.
14090 * configure: Regenerate.
14092 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14094 * expmed.c (expand_shift_1): Canonicalize rotates by
14095 constant bitsize / 2 to bitsize - 1.
14096 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14097 case ROTATERT>: Likewise.
14100 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14102 * config/i386/i386.md (rotateinv): New code attr.
14103 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14104 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14105 roll $31, %eax, etc.
14107 2013-05-14 Richard Biener <rguenther@suse.de>
14109 PR middle-end/57235
14110 * tree-eh.c (sink_clobbers): Give up for successors with
14111 multiple predecessors and no virtual uses.
14113 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14115 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14116 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14118 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14120 PR middle-end/57251
14121 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14122 the case when both op0 and op1 have VOIDmode.
14124 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14126 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
14127 in multiply-accumulate mode.
14129 2013-05-13 Guozhi Wei <carrot@google.com>
14131 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
14133 2013-05-13 Kai Tietz <ktietz@redhat.com>
14136 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
14137 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
14138 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
14139 * config/i386/i386.c (ix86_option_override_internal): Likewise.
14140 (ix86_expand_prologue): Likewise.
14141 (ix86_expand_split_stack_prologue): Likewise.
14142 (legitimate_pic_address_disp_p): Likewise.
14143 (legitimize_pic_address): Likewise.
14144 (legitimize_tls_address): Likewise.
14145 (legitimize_pe_coff_symbol): Likewise.
14146 (output_pic_addr_const): Likewise.
14147 (construct_plt_address): Likewise.
14148 (ix86_expand_call): Likewise.
14149 (x86_output_mi_thunk): Likewise.
14150 (x86_function_profiler): Likewise.
14152 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
14154 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
14155 similar switch cases.
14156 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
14157 (aarch64_simd_mov_to_<mode>low): Delete.
14158 (aarch64_simd_mov_to_<mode>high): Delete.
14159 (move_lo_quad_<mode>): Add w<-r alternative.
14160 (aarch64_simd_move_hi_quad_<mode>): Likewise.
14161 (aarch64_simd_mov_from_*): Update type attribute.
14162 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
14165 2013-05-13 Jan Hubicka <jh@suse.cz>
14167 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
14168 * config/i386/i386.c (ix86_compute_frame_layout,
14169 ix86_expand_epilogue, emit_i387_cw_initialization,
14170 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
14171 ix86_local_alignment): Fix use of size/speed predicates.
14173 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14175 PR tree-optimization/45216
14176 PR tree-optimization/57157
14177 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
14178 the (-Y) & (B - 1) variant if OP is |.
14179 * expmed.c (expand_shift_1): For rotations by const0_rtx just
14180 return shifted. Use (-op1) & (prec - 1) as other_amount
14181 instead of prec - op1.
14183 2013-05-13 Martin Jambor <mjambor@suse.cz>
14185 PR middle-end/42371
14186 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
14187 (ipa_constant_data): New type.
14188 (ipa_jump_func): Use ipa_constant_data to hold information about
14189 constant jump functions.
14190 (ipa_get_jf_constant): Adjust to jump function type changes.
14191 (ipa_get_jf_constant_rdesc): New function.
14192 (ipa_param_descriptor): New field controlled_uses.
14193 (ipa_get_controlled_uses): New function.
14194 (ipa_set_controlled_uses): Likewise.
14195 * ipa-ref.h (ipa_find_reference): Declare.
14196 * ipa-prop.c (ipa_cst_ref_desc): New type.
14197 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
14199 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
14200 New parameter cs. Adjust all callers.
14201 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
14202 (remove_described_reference): New function.
14203 (jfunc_rdesc_usable): Likewise.
14204 (try_make_edge_direct_simple_call): Decrement controlled use count,
14205 attempt to remove reference if it hits zero.
14206 (combine_controlled_uses_counters): New function.
14207 (propagate_controlled_uses): Likewise.
14208 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
14209 (ipa_edge_duplication_hook): Duplicate reference descriptions.
14210 (ipa_print_node_params): Print described use counter.
14211 (ipa_write_jump_function): Adjust to jump function type changes.
14212 (ipa_read_jump_function): New parameter CS, pass it to
14213 ipa_set_jf_constant. Adjust caller.
14214 (ipa_write_node_info): Stream controlled use count
14215 (ipa_read_node_info): Likewise.
14216 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
14218 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
14219 count. Remove cloning-added reference if it reaches zero.
14220 * ipa-ref.c (ipa_find_reference): New function.
14222 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14224 * config/i386/i386.c (processor_target_table): Modified default
14225 alignment values for AMD BD and BT architectures.
14227 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
14229 * tree-vect-generic.c (uniform_vector_p): Move ...
14230 * tree.c (uniform_vector_p): ... here.
14231 * tree.h (uniform_vector_p): Declare it.
14232 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
14235 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14237 PR tree-optimization/57230
14238 * tree-ssa-strlen.c (handle_char_store): Record length for
14239 array store from STRING_CST.
14241 PR tree-optimization/57230
14242 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
14245 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
14247 * config/epiphany/epiphany.c (epiphany_init): Check size of
14248 NUM_MODES_FOR_MODE_SWITCHING.
14249 (epiphany_expand_prologue):
14250 Remove CONFIG_REGNUM initial value handling code.
14251 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14252 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14253 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14254 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14255 Don't return 1 for FP_MODE_NONE.
14256 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14257 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14258 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14259 * config/epiphany/epiphany.md (save_config): New pattern.
14261 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14263 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14265 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14267 * config/i386/i386.md (memory): Handle sseishft1.
14268 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
14269 (*vec_extractv2di_1): Ditto.
14271 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
14273 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
14276 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
14278 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
14279 Add mthumb/march=armv7-a multilib.
14280 Add mthumb/march=armv7-r multilib.
14281 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
14283 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
14285 * config/v850/t-rtems: Add more multilibs.
14287 2013-05-10 Richard Biener <rguenther@suse.de>
14289 PR tree-optimization/57214
14290 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
14291 not propagate from SSA names that occur in abnormal PHI nodes.
14293 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
14295 * stor-layout.c (element_precision): New function.
14296 * machmode.h (element_precision): Declare it.
14297 * tree.c (build_minus_one_cst): New function.
14298 (element_precision): Likewise.
14299 * tree.h (build_minus_one_cst): Declare new function.
14300 (element_precision): Likewise.
14301 * fold-const.c (operand_equal_p): Use element_precision.
14302 (fold_binary_loc): Handle vector types.
14303 * convert.c (convert_to_integer): Use element_precision.
14304 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
14307 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
14309 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
14310 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
14311 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
14312 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14313 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
14314 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
14315 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
14316 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14317 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
14318 (Uuw8): New constraints.
14319 (Usb4): Move into alphabetical order.
14320 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
14321 (sd8_operand, ub8_operand, uw8_operand): New predicates.
14322 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
14323 previously unnamed patterns.
14324 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
14325 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
14326 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
14327 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
14328 of set_attr_alternative/if_then_else. Use extended_mips16 instead
14329 of specific lengths.
14331 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14333 * config/i386/i386.md (rotateinv): New code attr.
14334 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14335 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14336 roll $31, %eax, etc.
14338 PR tree-optimization/45216
14339 PR tree-optimization/57157
14340 * tree-ssa-forwprop.c (simplify_rotate): New function.
14341 (ssa_forward_propagate_and_combine): Call it.
14343 2013-05-10 Richard Biener <rguenther@suse.de>
14345 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
14346 disable peeling when we version for aliasing.
14347 (vector_alignment_reachable_p): Honor explicit user alignment.
14348 (vect_supportable_dr_alignment): Likewise.
14349 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
14350 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
14351 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
14352 then peeling to arrange for the cost-model check to come first.
14354 2013-05-10 Alan Modra <amodra@gmail.com>
14356 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
14357 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
14358 * configure: Regenerate.
14360 2013-05-10 Alan Modra <amodra@gmail.com>
14363 * varasm.c (default_elf_select_section): Move !DECL_P check..
14364 (get_named_section): ..to here before calling get_section_name.
14366 (default_section_type_flags): Add DECL_P check.
14367 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
14368 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
14370 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
14372 * config/epiphany/epiphany.c (epiphany_expand_prologue):
14373 When using gen_stack_adjust_str with a register offset, add a
14374 REG_FRAME_RELATED_EXPR note.
14376 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
14378 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
14379 (*vec_extractv4si_zext_mem): Ditto.
14380 (*vec_extractv2di): Add 0->x and x->x alternatives.
14381 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
14382 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
14384 2013-05-09 Jason Merrill <jason@redhat.com>
14386 N3639 C++1y VLA support
14387 * gimplify.c (gimplify_vla_decl): Don't touch an existing
14390 * tree.c (build_constructor_va): New.
14391 * tree.h: Declare it.
14393 2013-05-09 Martin Jambor <mjambor@suse.cz>
14396 * gimple-fold.c (canonicalize_constructor_val): Call
14397 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
14399 2013-05-09 Jan Hubicka <jh@suse.cz>
14400 Richard Biener <rguenther@suse.de>
14403 * symtab.c (symtab_make_decl_local): Do not add private names.
14405 2013-05-09 Jan Hubicka <jh@suse.cz>
14408 * symtab.c (insert_to_assembler_name_hash): Handle clones.
14409 (unlink_from_assembler_name_hash): Likewise.
14410 (symtab_prevail_in_asm_name_hash, symtab_register_node,
14411 symtab_unregister_node, symtab_initialize_asm_name_hash,
14412 change_decl_assembler_name): Update.
14414 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
14416 * config/aarch64/aarch64.md: New movtf split.
14417 (*movtf_aarch64): Update.
14418 (aarch64_movdi_tilow): Handle TF modes and rename to
14419 aarch64_movdi_<mode>low.
14420 (aarch64_movdi_tihigh): Handle TF modes and rename to
14421 aarch64_movdi_<mode>high
14422 (aarch64_movtihigh_di): Handle TF modes and rename to
14423 aarch64_mov<mode>high_di
14424 (aarch64_movtilow_di): Handle TF modes and rename to
14425 aarch64_mov<mode>low_di
14426 (aarch64_movtilow_tilow): Remove spurious whitespace.
14427 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
14429 (aarch64_print_operand): Update.
14431 2013-05-09 Alan Modra <amodra@gmail.com>
14433 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
14435 * configure: Regenerate.
14437 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14439 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
14440 splitter preparation statements.
14441 * config/i386/sse.md (*vec_extract* splitters): Ditto.
14442 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
14445 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14447 * gimple-ssa-strength-reduction.c (count_candidates): Change
14448 return value to int.
14449 (analyze_candidates_and_replace): Change type of length to int.
14451 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14453 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
14454 (*vec_extract<mode>): Use VI12_128 mode iterator.
14455 (*vec_extract<mode>_mem): Ditto.
14456 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
14459 2013-05-08 Diego Novillo <dnovillo@google.com>
14464 2012-08-17 Diego Novillo <dnovillo@google.com>
14467 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
14468 * config.in: Regenerate.
14469 * configure: Regenerate.
14470 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
14472 2013-05-08 Jan Hubicka <jh@suse.cz>
14475 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
14476 * cgraph.h (symtab_node_base): Add unique_name.
14477 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14478 input_overwrite_node, input_varpool_node): Stream unique_name.
14479 * cgraphclones.c (cgraph_create_virtual_clone,
14480 cgraph_function_versioning): Set unique_name.
14481 * ipa.c (function_and_variable_visibility): Set unique_name.
14483 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14485 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
14486 (alloc_cand_and_find_basis): Restrict conditional candidate
14487 processing to CAND_MULTs.
14489 2013-05-08 Jan Hubicka <jh@suse.cz>
14492 lto-symtab.c (lto_symtab_symbol_p): New function.
14493 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
14494 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
14495 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
14496 Skip static symbols.
14498 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
14500 PR tree-optimization/57200
14501 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
14502 Only call inform if the preceding warning_at returns true.
14504 2013-05-07 Han Shen <shenhan@google.com>
14506 * cfgexpand.c (record_or_union_type_has_array_p): New function.
14507 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
14508 * common.opt (fstack-protector-strong): New option.
14509 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
14510 * doc/invoke.texi (Optimization Options): Document
14511 "-fstack-protector-strong".
14512 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
14514 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14516 * config/mips/mips.c (mips_machine_reorg2): Return 0.
14518 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
14520 * ira.c (update_equiv_regs): Add insn having equiv memory even if
14521 it is not lhs of the insn.
14522 (setup_reg_equiv): Remove insn having equiv memory which it is not
14524 * lra-constraints.c (process_address): Try to improve generation
14525 code for address base + disp.
14526 (lra_constraints): Make correct the code for checking insn setting
14527 up backward equivalence. Remove insn only if it is in the init
14529 * lra-eliminations.c (update_reg_eliminate): Change return value.
14530 (lra_eliminate): Use the result.
14532 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
14534 * config/i386/sse.md (ssescalarnummask): New mode attribute.
14535 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
14536 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
14537 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
14538 register target operands.
14539 (*vec_extractv8hi_sse2): New pattern.
14540 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14541 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14542 (*vec_extract<mode>_mem): New insn and split pattern.
14544 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14546 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14547 (TARGET_ASAN_SHADOW_OFFSET): Define.
14548 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14549 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14551 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14553 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14554 (incr_vec_index): Return -1 if increment not found.
14555 (create_add_on_incoming_edge): Assert if increment not found.
14556 (record_increment): Limit number of increments recorded.
14557 (all_phi_incrs_profitable): Return false if an increment not found.
14558 (replace_profitable_candidates): Don't process increments that were
14560 (analyze_candidates_and_replace): Limit size of incr_vec.
14562 2013-05-07 Richard Biener <rguenther@suse.de>
14564 * calls.c (special_function_p): setjmp-like functions are leaf.
14565 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14566 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14568 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14570 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14571 (aarch64_simd_mov<mode>): New expander.
14572 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14573 (aarch64_simd_mov_to_<mode>high): Likewise.
14574 (aarch64_simd_mov_from_<mode>low): Likewise.
14575 (aarch64_simd_mov_from_<mode>high): Likewise.
14576 (aarch64_dup_lane<mode>): Update.
14577 (aarch64_dup_lanedi): New instruction pattern.
14578 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14579 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14581 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14583 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14584 (replace_mult_candidate): Remove unnecessary argument; remove
14585 unnecessary parameter from call to introduce_cast_before_cand.
14586 (replace_unconditional_candidate): Remove unnecessary parameter
14587 from call to replace_mult_candidate.
14588 (replace_conditional_candidate): Likewise.
14589 (insert_initializers): Use make_temp_ssa_name.
14590 (introduce_cast_before_cand): Remove unnecessary argument; use
14591 make_temp_ssa_name.
14592 (replace_one_candidate): Remove unnecessary argument; remove
14593 unnecessary parameter from calls to introduce_cast_before_cand.
14594 (replace_profitable_candidates): Remove unnecessary parameters
14595 from calls to replace_one_candidate.
14597 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14599 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14600 phi def as possibly hiding a basis for a CAND_ADD whose operands
14601 have been commuted in the analysis.
14602 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14604 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14606 * config/aarch64/aarch64.md
14607 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14608 shift value between 0-4.
14610 2013-05-07 Richard Biener <rguenther@suse.de>
14612 * double-int.h (rshift): New overload.
14613 * double-int.c (rshift): New function.
14614 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14615 (create_reference_ops_from_ref): Remove.
14616 (vn_reference_insert): Use shared ops for constructing the
14617 reference and copy it.
14619 2013-05-07 Richard Biener <rguenther@suse.de>
14621 PR middle-end/57190
14622 * tree-eh.c (sink_clobbers): Properly propagate
14623 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14625 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14627 PR tree-optimization/57149
14628 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14629 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14630 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14631 uninit_undefined_value_p instead of ssa_undefined_value_p.
14634 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14635 for modifier == EXPAND_INITIALIZER.
14637 2013-05-07 Anton Blanchard <anton@samba.org>
14639 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14640 for powerpc64 little endian.
14641 * configure: Regenerate.
14643 2013-05-06 Graham Stott <grahams@btinternet.com>
14645 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14646 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14647 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14648 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14650 2013-05-06 Graham Stott <grahams@btinternet.com>
14652 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14653 codes which allow non-lvalues.
14655 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14657 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14658 components are all 1s.
14659 (integer_minus_onep): New function.
14660 * tree.h (integer_minus_onep): Declare it.
14661 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14662 integer_minus_onep instead of integer_all_onesp.
14664 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14667 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14668 variations of these patterns.
14670 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14672 * config/i386/i386.md (isa): Add x64_sse4 member.
14673 (enabled): Handle x64_sse4.
14674 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14675 instruction for 64bit SSE4_1 targets. Update insn attributes.
14676 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14677 instruction for SSE4_1 targets. Update insn attributes.
14678 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14679 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14680 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14682 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14683 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14684 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14686 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14689 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14691 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14693 * final.c (do_assembler_dialects): Don't handle curly braces and
14694 vertical bar escaped by % as dialect delimiters.
14695 (output_asm_insn): Print curly braces and vertical bar if escaped
14696 by % and ASSEMBLER_DIALECT defined.
14697 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14698 * doc/tm.texi: Regenerated.
14700 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14702 * config/mips/mips.c: Include tree-pass.h.
14703 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14704 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14705 (pass_mips_machine_reorg2): New machine specific pass.
14706 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14707 (mips_option_override): Register the new pass.
14708 * rtl.h (cleanup_barriers): Remove prototype.
14709 (dbr_schedule): Likewise.
14710 * jump.c (cleanup_barriers): Make static.
14711 * reorg.c (dbr_schedule): Likewise.
14713 2013-05-06 Richard Biener <rguenther@suse.de>
14715 PR tree-optimization/57185
14716 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14717 SSA names properly.
14719 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14722 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14723 between "first_arg" and "insn", not "last" and "insn".
14725 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14727 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14728 (find_candidates_in_block): Re-enable slsr_process_phi.
14729 (create_phi_basis): Fix double counting of candidate adjustment.
14731 2013-05-06 Richard Biener <rguenther@suse.de>
14733 PR middle-end/57147
14734 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14735 the edge is also fallthru, preserve it and just clear the
14737 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14738 also complex, preserve that and just clear the fallthru flag.
14739 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14740 update virtual operands.
14742 2013-05-06 Alan Modra <amodra@gmail.com>
14744 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14745 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14746 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14747 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14748 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14749 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14751 2013-05-06 Alan Modra <amodra@gmail.com>
14753 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14754 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14755 (DEFAULT_ASM_ENDIAN): Define.
14756 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14757 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14758 Update -K PIC clause from sysv4.h.
14759 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14760 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14762 2013-05-06 Alan Modra <amodra@gmail.com>
14764 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14765 twice for little-endian.
14766 (ashrdi3_no_power, ashrdi3): Support little-endian.
14768 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14771 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14772 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14774 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14775 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14776 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14779 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14780 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14782 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14783 * config/pa/pa.opt: Make mbig-switch a no-op.
14784 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14785 (CASE_VECTOR_MODE): Always return SImode.
14786 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14787 for the !TARGET_BIG_SWITCH case.
14788 * config/pa/pa-linux.h: Likewise.
14789 * config/pa/pa-openbsd.h: Likewise.
14790 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14791 * config/pa/pa.md (short_jump): Remove define_insn.
14792 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14793 (casesi0): Remove define_insn.
14794 (type): Remove btable_branch.
14795 (pa_combine_type): Likewise.
14796 (in_nullified_branch_delay): Likewise.
14797 (in_call_delay): Likewise.
14798 (define_delay): Likewise.
14799 (define_insn_reservation "Z3"): Likewise.
14800 (define_insn_reservation "Z4"): Likewise.
14801 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14802 (pa_adjust_insn_length): Remove adjustment for btable branches.
14803 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14806 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14808 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14809 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14810 Add m->r,x alternatives.
14811 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14812 splitters using SWI48x mode iterator.
14813 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14814 TARGET_64BIT. Add m->x alternative.
14815 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14816 Add o->x alternative. Enable for TARGET_SSE.
14817 (sse_storeq): Remove expander.
14818 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14819 with memory input operand.
14820 (*vec_extractv2di_1 splitter): New.
14821 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14822 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14824 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14826 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14827 (INT_LOWPART): Delete.
14828 (extract_MB): Adjust.
14829 (extract_ME): Adjust.
14830 (print_operand): Adjust.
14832 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14834 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14835 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14836 (reg_or_logical_cint_operand, easy_fp_constant,
14837 logical_const_operand): Delete "CONST_DOUBLE" case.
14838 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14839 "HOST_BITS_PER_WIDE_INT == 64" test.
14840 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14841 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14842 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14843 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14845 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14846 CONST_DOUBLE DImode/VOIDmode case.
14847 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14848 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14849 CONST_DOUBLE VOIDmode case.
14850 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14851 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14852 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14853 Delete CONST_DOUBLE case.
14854 (splitters for mov FMOVE64 const_double): Delete
14855 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14856 "HOST_BITS_PER_WIDE_INT >= 64" test.
14857 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14859 (mov DI const_double): Delete.
14861 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14863 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14864 on op shows all bits zero in mode of a lowpart subreg, return zero.
14866 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14869 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14870 to save TFmode registers and DImode to save TImode registers for
14871 caller save operations.
14872 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14873 mark being partially clobbered since they only use the first
14876 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14877 and TDmode only use the upper 64-bits of each VSX register.
14879 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14881 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14882 (find_candidates_in_block): Disable slsr_process_phi.
14884 2013-05-03 Guozhi Wei <carrot@google.com>
14886 * coverage.c (coverage_obj_init): Move the construction of gcov
14888 (build_init_ctor): ... here.
14890 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14892 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14893 (slsr_cand_d): Redefine def_phi.
14894 (stride_status, phi_adjust_status, count_phis_status): New enums.
14895 (find_phi_def): New.
14896 (find_basis_for_base_expr): New.
14897 (find_basis_for_candidate): Handle hidden bases.
14898 (alloc_cand_and_find_basis): Handle phi candidates.
14899 (slsr_process_phi): New.
14900 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14901 (create_mul_imm_cand): Likewise.
14902 (create_add_ssa_cand): Exclude phi base candidates.
14903 (create_add_imm_cand): Likewise.
14904 (slsr_process_cast): Likewise.
14905 (slsr_process_copy): Likewise.
14906 (find_candidates_in_block): Handle phi candidates.
14907 (dump_candidate): Likewise.
14908 (unconditional_cands): Delete.
14909 (unconditional_cands_with_known_stride_p): Delete.
14910 (phi_dependent_cand_p): New.
14911 (cand_increment): Handle phi-dependent candidates.
14912 (replace_dependent): Delete.
14913 (replace_mult_candidate): New.
14914 (replace_unconditional_candidate): New.
14915 (incr_vec_index): Move to avoid forward reference.
14916 (create_add_on_incoming_edge): New.
14917 (create_phi_basis): New.
14918 (replace_dependents): Delete.
14919 (replace_conditional_candidate): New.
14920 (phi_add_costs): New.
14921 (replace_uncond_cands_and_profitable_phis): New.
14922 (record_increment): Handle phi adjustments.
14923 (record_phi_increments): New.
14924 (record_increments): Handle phi adjustments.
14925 (phi_incr_cost): New.
14926 (lowest_cost_path): Handle phis.
14927 (total_savings): Likewise.
14928 (analyze_increments): Likewise.
14929 (ncd_with_phi): New.
14930 (ncd_of_cand_and_phis): New.
14931 (nearest_common_dominator_for_cands): Handle phi increments.
14932 (all_phi_incrs_profitable): New.
14933 (replace_profitable_candidates): Handle phi-dependent candidates.
14934 (analyze_candidates_and_replace): Likewise.
14936 2013-05-03 Teresa Johnson <tejohnson@google.com>
14939 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14940 do not exceed REG_BR_PROB_BASE.
14942 2013-05-03 Jeff Law <law@redhat.com>
14944 PR tree-optimization/57144
14945 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14946 operand of the condition will bit into the new type when eliminating
14947 a cast feeding a condition.
14949 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14951 PR rtl-optimization/57130
14952 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14953 of COMPARE as in_code to the recursive call if needed.
14955 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14957 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14958 (enabled): Handle new members.
14959 * config/i386/sse.md (*vec_concatv2si): Merge from
14960 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14961 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14963 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14965 PR tree-optimization/57027
14966 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14967 for fnms opportunity, check we got the prerequisite kind
14968 of tree / gimple before using accessor functions.
14970 2013-05-03 Richard Biener <rguenther@suse.de>
14972 * double-int.h (lshift): New overload without precision
14973 and arith argument.
14974 (operator *=, operator +=, operator -=): Move ...
14975 * double-int.c (operator *=, operator +=, operator -=): ... here
14976 and implement more efficiently.
14977 (mul_double_with_sign): Remove.
14978 (lshift_double): Adjust to take unsinged shift argument, push
14979 dispatching code to callers.
14980 (mul_double_wide_with_sign): Add early out for callers that
14981 are not interested in high parts or overflow.
14982 (lshift): New function.
14983 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14984 dispatch code here.
14985 (lrotate, rrotate): Use logical shifts.
14986 * expr.c (get_inner_reference): Use lshift.
14987 * fixed-value.c (do_fixed_divide): Likewise.
14988 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14989 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14990 (indirect_refs_may_alias_p): Likewise.
14991 (stmt_kills_ref_p_1): Likewise.
14993 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14995 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14997 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14999 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
15000 scalar form of FABD instruction.
15002 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
15004 * lra-constraints.c (process_alt_operands): Add checking alt
15005 number to choose the best alternative.
15007 2013-05-02 Richard Biener <rguenther@suse.de>
15009 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
15010 bitmap and its handling.
15011 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
15013 2013-05-02 Richard Biener <rguenther@suse.de>
15015 PR middle-end/57140
15016 * tree-inline.c (copy_loops): Properly handle removed loops.
15017 (copy_cfg_body): Mark destination loops for fixup if source
15018 loops needed fixup.
15020 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15023 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15024 generating simple_return for naked functions.
15026 2013-05-02 Martin Jambor <mjambor@suse.cz>
15028 PR middle-end/56988
15029 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15030 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15032 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15033 ipa_agg_replacement_value structures.
15034 (known_aggs_to_agg_replacement_list): Likewise.
15035 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15036 (read_agg_replacement_chain): Likewise.
15037 (ipcp_transform_function): Also check that by_ref flags match.
15039 2013-05-02 Richard Biener <rguenther@suse.de>
15041 * graphds.h (struct graph): Add obstack member.
15042 * graphds.c (new_graph): Initialize obstack and allocate
15044 (add_edge): Allocate edge from the obstack.
15045 (free_graph): Free the obstack instead of all edges and vertices.
15047 2013-05-02 Teresa Johnson <tejohnson@google.com>
15049 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15051 * cfg.c (update_bb_profile_for_threading): Ditto.
15052 * tree-inline.c (copy_bb): Ditto.
15053 (copy_edges_for_bb): Ditto.
15054 (initialize_cfun): Ditto.
15055 (copy_cfg_body): Ditto.
15056 (expand_call_inline): Ditto.
15057 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15058 (estimate_node_size_and_time): Ditto.
15059 (inline_merge_summary): Ditto.
15060 * cgraphclones.c (cgraph_clone_edge): Ditto.
15061 (cgraph_clone_node): Ditto.
15062 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15063 (compute_trg_info): Ditto.
15065 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15067 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15068 S reg when fp attribute set.
15069 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15071 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15073 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15075 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15076 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15077 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15079 2013-05-02 Richard Biener <rguenther@suse.de>
15081 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15082 (struct instantiate_cache_entry): New type.
15083 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15084 (struct instantiate_cache_type): New type.
15085 (set_instantiated_value, get_instantiated_value): Remove.
15086 (get_instantiated_value_entry): New function.
15087 (instantiate_scev_name): Use the new cache and adjust.
15088 (instantiate_scev_poly): Adjust.
15089 (instantiate_scev_binary): Likewise.
15090 (instantiate_array_ref): Likewise.
15091 (instantiate_scev_convert): Likewise.
15092 (instantiate_scev_not): Likewise.
15093 (instantiate_scev_3): Likewise.
15094 (instantiate_scev_2): Likewise.
15095 (instantiate_scev_r): Likewise.
15096 (instantiate_scev): Likewise.
15097 (resolve_mixers): Likewise.
15099 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15102 * lra-constraints.c (best_small_class_operands_num): Remove.
15103 (process_alt_operands): Remove small_class_operands_num. Take
15104 small classes operands into losers and only if the operand is not
15105 matched. Modify debugging output.
15106 (curr_insn_transform): Remove best_small_class_operands_num.
15109 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15111 * config/aarch64/aarch64-builtins.c
15112 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15113 * config/aarch64/aarch64-simd-builtins.def
15114 (reduc_splus_): Add new modes.
15115 (reduc_uplus_): New.
15116 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15117 (reduc_uplus_v4sf): Likewise.
15118 (reduc_splus_v4sf): Likewise.
15119 (aarch64_addv<mode>): Likewise.
15120 (reduc_uplus_<mode>): Likewise.
15121 (reduc_splus_<mode>): Likewise.
15122 (aarch64_addvv2di): Likewise.
15123 (reduc_uplus_v2di): Likewise.
15124 (reduc_splus_v2di): Likewise.
15125 (aarch64_addvv2si): Likewise.
15126 (reduc_uplus_v2si): Likewise.
15127 (reduc_splus_v2si): Likewise.
15128 (reduc_<sur>plus_<mode>): New.
15129 (reduc_<sur>plus_v2di): Likewise.
15130 (reduc_<sur>plus_v2si): Likewise.
15131 (reduc_<sur>plus_v4sf): Likewise.
15132 (aarch64_addpv4sf): Likewise.
15133 * config/aarch64/arm_neon.h
15134 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
15135 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
15136 add UNSPEC_SADDV, UNSPEC_UADDV.
15138 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
15140 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15142 * config/aarch64/arm_neon.h
15143 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
15145 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15147 * config/aarch64/aarch64-builtins
15148 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
15150 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15152 * config/aarch64/aarch64-simd-builtins.def
15153 (reduc_smax_): New.
15154 (reduc_smin_): Likewise.
15155 (reduc_umax_): Likewise.
15156 (reduc_umin_): Likewise.
15157 (reduc_smax_nan_): Likewise.
15158 (reduc_smin_nan_): Likewise.
15161 (smax): Update for V2SF, V4SF and V2DF modes.
15164 (smin_nan): Likewise.
15165 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
15166 (<su><maxmin><mode>3): ...This, refactor.
15167 (s<maxmin><mode>3): New.
15168 (<maxmin_uns><mode>3): Likewise.
15169 (reduc_<maxmin_uns>_<mode>): Refactor.
15170 (reduc_<maxmin_uns>_v4sf): Likewise.
15171 (reduc_<maxmin_uns>_v2si): Likewise.
15172 (aarch64_<fmaxmin><mode>: Remove.
15173 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
15175 (vmin<q>_f<32,64>): Likewise.
15176 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15178 (su): Add mappings for smax, smin, umax, umin.
15180 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15181 (FMAXMIN): Rename as...
15182 (FMAXMIN_UNS): ...This.
15184 (fmaxminv): Likewise.
15185 (fmaxmin): Likewise.
15187 (maxmin_uns_op): Likewise.
15189 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15191 * config/aarch64/arm_neon.h
15192 (vac<ge, gt><sd>_f<32, 64>): Rename to...
15193 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
15194 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
15196 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15198 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
15199 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
15201 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15203 * config/aarch64/aarch64-simd.md
15204 (vcond<mode>_internal): Handle special cases for constant masks.
15205 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
15206 (vcondu<mode><mode>): Likewise.
15207 (vcond<v_cmp_result><mode>): New.
15209 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15211 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
15212 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
15213 * config/aarch64/aarch64-simd-builtins.def
15214 (cmeq): Update to BUILTIN_VALLDI.
15219 * config/aarch64/arm_neon.h
15220 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
15221 to builtins or C as appropriate.
15223 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15225 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
15227 (cmhi): Rename to...
15229 * config/aarch64/aarch64-simd.md
15230 (simd_mode): Add SF.
15231 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
15232 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
15233 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
15234 (cstore<mode>_neg): ...This.
15235 * config/aarch64/iterators.md
15237 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
15238 (COMPARISONS): New.
15239 (UCOMPARISONS): Likewise.
15240 (optab): Add missing comparisons.
15246 (VCMP_S): Likewise.
15247 (VCMP_U): Likewise.
15248 (V_cmp_result): Add DF, SF modes.
15249 (v_cmp_result): Likewise.
15251 (vmtype): Likewise.
15252 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15254 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15256 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15257 define_insn to define_insn_and_split.
15258 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15259 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15260 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15261 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15262 (thumb2_negscc): Likewise.
15264 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15266 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15268 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15270 * config/arm/thumb2.md: Remove trailing whitespaces.
15272 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15274 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
15275 Use gen_int_mode rather than GEN_INT.
15277 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
15279 * value-prof.c (stream_in_histogram_value): Remove the strayed
15282 2013-04-30 Richard Biener <rguenther@suse.de>
15284 PR middle-end/57122
15285 * cfghooks.c (split_edge): Properly check for the loop latch edge.
15287 2013-04-30 Richard Biener <rguenther@suse.de>
15289 PR middle-end/57107
15290 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
15292 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
15294 PR rtl-optimization/56957
15295 PR rtl-optimization/57105
15296 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
15297 variable. Use just INSN_UID for determining whether an insn
15298 should be only disconnected from the insn stream.
15299 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
15301 2013-04-30 Jakub Jelinek <jakub@redhat.com>
15303 PR tree-optimization/57104
15304 * tsan.c (instrument_expr): Don't instrument accesses to
15305 DECL_HARD_REGISTER VAR_DECLs.
15307 2013-04-30 Richard Biener <rguenther@suse.de>
15309 * function.h (loops_for_fn): New inline function.
15310 (set_loops_for_fn): Likewise.
15311 * cfgloop.h (place_new_loop): Add struct function parameter.
15312 (get_loop): Likewise.
15313 (get_loops): Likewise.
15314 (number_of_loops): Likewise.
15315 (fel_next): Adjust.
15316 (fel_init): Likewise.
15317 * cfg.c (get_loop_copy): Adjust.
15318 * cfgloop.c (flow_loops_dump): Likewise.
15319 (record_loop_exits): Likewise.
15320 (verify_loop_structure): Likewise.
15321 * cfgloopanal.c (mark_irreducible_loops): Likewise.
15322 (estimate_reg_pressure_cost): Likewise.
15323 (mark_loop_exit_edges): Likewise.
15324 * cfgloopmanip.c (place_new_loop): Likewise.
15325 (add_loop): Likewise.
15326 (duplicate_loop): Likewise.
15327 * graph.c (draw_cfg_nodes): Likewise.
15328 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
15329 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
15330 (extract_affine_chrec): Likewise.
15331 (build_scop_iteration_domain): Likewise.
15332 * graphite.c (graphite_initialize): Likewise.
15333 * ira-build.c (create_loop_tree_nodes): Likewise.
15334 (more_one_region_p): Likewise.
15335 (rebuild_regno_allocno_maps): Likewise.
15336 (mark_loops_for_removal): Likewise.
15337 (mark_all_loops_for_removal): Likewise.
15338 (remove_unnecessary_regions): Likewise.
15339 (ira_build): Likewise.
15340 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15341 * loop-init.c (fix_loop_structure): Likewise.
15342 (gate_rtl_move_loop_invariants): Likewise.
15343 (gate_rtl_unswitch): Likewise.
15344 (gate_rtl_unroll_and_peel_loops): Likewise.
15345 (rtl_doloop): Likewise.
15346 * lto-streamer-in.c (input_cfg): Likewise.
15347 * lto-streamer-out.c (output_cfg): Likewise.
15348 * modulo-sched.c (sms_schedule): Likewise.
15349 * predict.c (tree_estimate_probability): Likewise.
15350 (tree_estimate_probability_driver): Likewise.
15351 (estimate_loops): Likewise.
15352 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
15353 (move_sese_region_to_fn): Likewise.
15354 (debug_loop_num): Likewise.
15355 * tree-chrec.c (chrec_evaluate): Likewise.
15356 (hide_evolution_in_other_loops_than_loop): Likewise.
15357 (chrec_component_in_loop_num): Likewise.
15358 (reset_evolution_in_loop): Likewise.
15359 (evolution_function_is_invariant_rec_p): Likewise.
15360 * tree-if-conv.c (main_tree_if_conversion): Likewise.
15361 * tree-inline.c (copy_loops): Likewise.
15362 (copy_cfg_body): Likewise.
15363 (tree_function_versioning): Likewise.
15364 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
15365 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
15367 (add_to_evolution_1): Likewise.
15368 (scev_const_prop): Likewise.
15369 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
15370 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
15371 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
15372 (tree_ssa_lim_initialize): Likewise.
15373 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
15374 (verify_loop_closed_ssa): Likewise.
15375 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
15376 (tree_ssa_loop_im): Likewise.
15377 (tree_ssa_loop_unswitch): Likewise.
15378 (tree_vectorize): Likewise.
15379 (check_data_deps): Likewise.
15380 (tree_ssa_loop_ivcanon): Likewise.
15381 (tree_ssa_loop_bounds): Likewise.
15382 (tree_complete_unroll): Likewise.
15383 (tree_complete_unroll_inner): Likewise.
15384 (tree_parallelize_loops): Likewise.
15385 (tree_ssa_loop_prefetch): Likewise.
15386 (tree_ssa_loop_ivopts): Likewise.
15387 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
15388 * tree-vectorizer.c (vectorize_loops): Likewise.
15390 2013-04-29 Mike Frysinger <vapier@gentoo.org>
15392 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
15393 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
15394 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
15395 with EABI_LINK_SPEC.
15397 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15400 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
15403 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
15406 * lra-constraints.c (process_alt_operands): Discourage a bit more
15407 using memory for pseudos. Print cost dump for alternatives.
15408 Modify cost values for conflicts with early clobbers.
15409 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
15411 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15414 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
15416 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15418 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
15419 from/to S register.
15420 (movdi_aarch64): Support LDR/STR from/to D register.
15422 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15424 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
15425 or higher by default.
15427 2013-04-29 Richard Biener <rguenther@suse.de>
15429 PR middle-end/57075
15430 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
15431 even if not adding abnormal edges for calls that can make
15434 2013-04-29 Richard Biener <rguenther@suse.de>
15436 PR middle-end/57103
15437 * tree-cfg.c (move_stmt_op): Fix condition under which to update
15439 (move_stmt_r): Remove redundant checking.
15441 2013-04-29 Teresa Johnson <tejohnson@google.com>
15444 * basic-block.h (apply_scale): New function.
15445 (apply_probability): Use apply_scale.
15446 * gimple-streamer-in.c (input_bb): Ditto.
15447 * lto-streamer-in.c (input_cfg): Ditto.
15448 * lto-cgraph.c (merge_profile_summaries): Ditto.
15449 * tree-optimize.c (execute_fixup_cfg): Ditto.
15450 * tree-inline.c (copy_bb): Update comment to use apply_scale.
15451 (copy_edges_for_bb): Ditto.
15452 (copy_cfg_body): Ditto.
15454 2013-04-29 Tom de Vries <tom@codesourcery.com>
15456 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
15457 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
15458 (tail_merge_optimize): Handle current_loops == NULL.
15460 2013-04-26 Jeff Law <law@redhat.com>
15462 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
15463 (simplify_cond_using_ranges): Generalize code to simplify
15464 COND_EXPRs where one argument is a constant and the other
15465 is an SSA_NAME created by an integral type conversion.
15467 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15469 * config/arm/arm.md (store_minmaxsi): Use only when
15470 optimize_insn_for_size_p.
15472 2013-04-29 Christian Bruel <christian.bruel@st.com>
15475 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
15477 2013-04-29 Richard Biener <rguenther@suse.de>
15479 PR middle-end/57089
15480 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
15481 loop tree make sure to schedule a fixup for the child as well.
15482 (expand_omp_for_generic): Properly add loops.
15483 (expand_omp_for_static_nochunk): Likewise.
15484 (expand_omp_for_static_chunk): Likewise.
15485 (expand_omp_for): For the degenerate case fixup loops.
15486 (expand_omp_sections): Fix default bb placement in loops.
15487 (expand_omp_atomic_pipeline): Properly add loops.
15489 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15491 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
15493 2013-04-29 Tom de Vries <tom@codesourcery.com>
15495 * tree-ssa-tail-merge.c: Update header comment.
15497 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15499 * config/aarch64/arm_neon.h
15500 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
15501 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
15502 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
15503 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
15504 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
15505 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
15506 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
15507 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
15509 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15511 * config/aarch64/aarch64-simd.md
15512 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
15513 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
15514 fix_trunc, fixuns_trunc.
15515 (ftrunc<VDQF:mode>2): New.
15516 * config/aarch64/iterators.md (optab): Add fix, fixuns.
15517 (fix_trunc_optab): New.
15519 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15521 * config/aarch64/aarch64-builtins.c
15522 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
15523 iceilf, lround, iroundf.
15525 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15528 * config/i386/i386.h (enum ix86_tune_indices)
15529 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
15530 New, split from X86_TUNE_INTER_UNIT_MOVES.
15531 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
15532 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
15533 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
15534 (TARGET_INTER_UNIT_MOVES): Remove.
15535 * config/i386/i386.c (initial_ix86_tune_features): Update.
15536 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
15537 (ix86_expand_convert_uns_didf_sse): Use
15538 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15539 (ix86_expand_vector_init_one_nonzero): Ditto.
15540 (ix86_expand_vector_init_interleave): Ditto.
15541 (inline_secondary_memory_needed): Return true for moves from SSE class
15542 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15543 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15544 * config/i386/constraints.md (Yi, Ym): Depend on
15545 TARGET_INTER_UNIT_MOVES_TO_VEC.
15546 (Yj, Yn): New constraints.
15547 * config/i386/i386.md (*movdi_internal): Change constraints of
15548 operand 1 from Yi to Yj and from Ym to Yn.
15549 (*movsi_internal): Ditto.
15550 (*movdf_internal): Ditto.
15551 (*movsf_internal): Ditto.
15552 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15553 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15554 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15555 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15556 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15557 * config/i386/sse.md (movdi_to_sse): Ditto.
15558 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15559 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15560 TARGET_INTER_UNIT_MOVES.
15561 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15562 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15563 instead of TARGET_INTER_UNIT_MOVES.
15564 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15565 operand 1 from Yi to Yj and from Ym to Yn.
15567 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15569 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15570 (float_truncate_hi_): Likewise.
15571 (float_extend_lo_): Likewise.
15572 (float_truncate_lo_): Likewise.
15573 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15574 (aarch64_float_extend_lo_v2df): Likewise.
15575 (vec_unpacks_hi_v4sf): Likewise.
15576 (aarch64_float_truncate_lo_v2sf): Likewise.
15577 (aarch64_float_truncate_hi_v4sf): Likewise.
15578 (vec_pack_trunc_v2df): Likewise.
15579 (vec_pack_trunc_df): Likewise.
15581 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15583 * config/aarch64/aarch64-builtins.c
15584 (aarch64_fold_builtin): Fold float conversions.
15585 * config/aarch64/aarch64-simd-builtins.def
15586 (floatv2si, floatv4si, floatv2di): New.
15587 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15588 * config/aarch64/aarch64-simd.md
15589 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15590 * config/aarch64/iterators.md (FLOATUORS): New.
15591 (optab): Add float, floatuns.
15592 (su_optab): Likewise.
15594 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15596 * config/aarch64/aarch64-builtins.c
15597 (aarch64_builtin_vectorized_function): Use new names for
15599 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15600 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15601 (fcvtzu): Split as...
15602 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15603 (fcvtas): Split as...
15604 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15605 (fcvtau): Split as...
15606 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15607 (fcvtps): Split as...
15608 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15609 (fcvtpu): Split as...
15610 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15611 (fcvtms): Split as...
15612 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15613 (fcvtmu): Split as...
15614 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15615 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15616 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15617 (lfrintnusf, lfrintnudf): Likewise.
15618 * config/aarch64/aarch64-simd.md
15619 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15621 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15622 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15623 (fcvt_pattern): Likewise.
15625 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15627 * config/aarch64/aarch64-simd.md
15628 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15629 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15631 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15633 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15634 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15635 (vrnd<a,m,n,p>_f32): Implement using builtins.
15636 (vrnd<i,x><q>_f<32, 64>): New.
15638 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15640 * config/aarch64/aarch64-builtins.c
15641 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15642 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15643 (frintz): Rename to...
15645 (frintp): Rename to...
15647 (frintm): Rename to...
15649 (frinti): Rename to...
15650 (nearbyint): ...this.
15651 (frintx): Rename to...
15653 (frinta): Rename to...
15655 * config/aarch64/aarch64-simd.md
15656 (aarch64_frint<frint_suffix><mode>): Delete.
15657 (<frint_pattern><mode>2): Convert to insn.
15658 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15659 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15660 (frint_pattern): Likewise.
15661 (frint_suffix): Likewise.
15663 2013-04-29 Richard Biener <rguenther@suse.de>
15665 PR tree-optimization/57081
15666 * loop-init.c: Include tree-flow.h.
15667 (loop_optimizer_finalize): Free number of iteration estimates.
15668 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15670 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15672 PR tree-optimization/57083
15673 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15674 non-singleton shift count range, zero extend low_bound for uns case.
15676 * config/i386/predicates.md (general_vector_operand): New predicate.
15677 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15678 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15679 if they aren't nonimmediate operands. If their original values
15680 satisfy const_vector_equal_evenodd_p, don't shift them.
15681 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15682 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15683 if not nonimmediate_operand.
15684 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15685 instead of register_operand.
15686 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15688 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15690 * stor-layout.c (finalize_size_functions): Allocate a structure and
15691 reset cfun before dumping the functions.
15693 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15695 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15698 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15699 use xop_pmacsdqh if uns_p.
15700 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15701 the immediate rotate count.
15703 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15705 * rtl.h (struct rtx_def): Add comment for field jump.
15706 (LRA_SUBREG_P): New macro.
15707 * recog.c (register_operand): Check LRA_SUBREG_P.
15708 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15710 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15711 (lra_final_code_change): Skip subreg change for operators.
15712 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15713 if there are no operand changes.
15714 * lra-constraints.c (curr_insn_set): New.
15715 (match_reload): Set LRA_SUBREG_P.
15716 (emit_spill_move): Ditto.
15717 (check_and_process_move): Use curr_insn_set. Process only single
15718 set insns. Don't initialize sec_mem_p and change_p.
15719 (simplify_operand_subreg): Use LRA_SUBREG_P.
15720 (reg_in_class_p): New function.
15721 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15722 of #ifdef. Add code to remove cycling.
15723 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15724 non-null disp. Reload inner instead of disp when base and index
15725 are null. Try to put lo_sum into register.
15726 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15727 (check_and_process_move): Move code for move cost check to
15728 simple_move_p. Remove equiv_substitution.
15729 (simple_move_p): New function.
15730 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15731 curr_insn_set. Call check_and_process_move only for single set
15732 insns. Use the new function. Move call of check_and_process_move
15733 after operand equiv substitution and address process.
15735 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15738 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15739 with nonlocal goto receivers or returns twice calls, ignore
15740 unininitialized values from abnormal edges to nl goto receiver
15741 or returns twice call.
15743 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15745 PR tree-optimization/57051
15746 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15747 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15750 2013-04-26 Richard Biener <rguenther@suse.de>
15752 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15753 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15754 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15755 (fixup_loop_arrays_after_move): New function.
15756 (move_sese_region_to_fn): Properly outline the loop tree parts
15757 of the SESE region.
15759 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15761 * config/i386/i386.md (type, unit): Fix long lines.
15763 2013-04-26 Richard Biener <rguenther@suse.de>
15765 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15766 (lto-streamer-out.o): Likewise.
15767 * cfgloop.c (init_loops_structure): Export, add struct function
15768 argument and adjust.
15769 (flow_loops_find): Adjust.
15770 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15771 (init_loops_structure): Declare.
15772 * lto-streamer-in.c: Include cfgloop.h.
15773 (input_cfg): Input the loop tree.
15774 * lto-streamer-out.c: Include cfgloop.h.
15775 (output_cfg): Output the loop tree.
15776 (output_struct_function_base): Do not drop PROP_loops.
15778 2013-03-26 Richard Biener <rguenther@suse.de>
15780 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15781 (pass_build_cfg): Provide PROP_loops.
15782 (move_sese_region_to_fn): Remove loops that are outlined into fn
15784 * tree-inline.c: Include cfgloop.h.
15785 (initialize_cfun): Do not drop PROP_loops.
15786 (copy_loops): New function.
15787 (copy_cfg_body): Copy loop structure.
15788 (tree_function_versioning): Initialize destination loop tree.
15789 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15790 (pass_parallelize_loops): Do IL verification.
15791 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15792 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15793 the CFG make sure we fixup loops as well.
15794 * tree-ssa-tail-merge.c: Include cfgloop.h.
15795 (replace_block_by): When merging loop latches mark loops for fixup.
15796 * lto-streamer-out.c (output_struct_function_base): Drop
15797 PROP_loops for now.
15798 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15799 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15800 * ipa-split.c: Include cfgloop.h.
15801 (split_function): Add the new return block to the loop tree root.
15802 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15803 whether we have removed the forwarder block.
15804 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15805 * cfgloop.h (place_new_loop): Declare.
15806 * cfgloopmanip.c (place_new_loop): Export.
15807 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15808 (tree-switch-conversion.o): Likewise.
15809 (tree-complex.o): Likewise.
15810 (tree-inline.o): Likewise.
15811 (tree-ssa-tailmerge.o): Likewise.
15812 (ipa-split.o): Likewise.
15813 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15814 (tree-ssa-copy.o): Likewise.
15815 * tree-switch-conversion.c: Include cfgloop.h
15816 (process_switch): If we emit a bit-test cascade, schedule loops
15818 * tree-complex.c: Include cfgloop.h.
15819 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15820 * asan.c: Include cfgloop.h.
15821 (create_cond_insert_point): Properly add new basic-blocks to
15822 loops, schedule loop fixup.
15823 * cfgloop.c (verify_loop_structure): Check that looks are not
15825 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15827 (expand_omp_for_generic): Likewise.
15828 (expand_omp_sections): Likewise.
15829 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15830 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15831 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15832 is initialized, not when loops are present.
15833 * tree-parloops.c (parallelize_loops): Remove checking here.
15834 * passes.c (init_optimization_passes): Schedule a copy-propagation
15835 pass before complete unrolling of inner loops.
15837 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15839 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15840 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15841 (colorize_init): Add argument to _WIN32 version.
15842 * toplev.c: Include diagnostic-color.h.
15843 (process_options): Default to -fdiagnostics-color=auto if
15844 GCC_COLORS env var is in the environment.
15845 * common.opt (fdiagnostics-color=): Add Var and Init.
15846 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15847 env var is in the environment, the default is auto rather than never.
15849 * diagnostic.h (file_name_as_prefix): Add context argument.
15850 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15851 the string as locus.
15852 * langhooks.c (lhd_print_error_function): Adjust caller.
15854 2013-04-25 Lawrence Crowl <crowl@google.com>
15856 * var-tracking.c (shared_hash_def::htab):
15857 Change type to hash_table. Update dependent calls and types.
15859 2013-04-25 Lawrence Crowl <crowl@google.com>
15861 * Makefile.in: Update as needed below.
15863 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15864 Move declaration to after the type's method definitons.
15866 * attribs.c (htab_t scoped_attributes::attribute_hash):
15867 Change type to hash_table. Update dependent calls and types.
15869 * bitmap.c (htab_t bitmap_desc_hash):
15870 Change type to hash_table. Update dependent calls and types.
15872 * cselib.c (htab_t cselib_hash_table):
15873 Change type to hash_table. Update dependent calls and types.
15875 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15876 (hash_string_slot_node): Move implementation into lto-streamer.h
15877 struct string_slot_hasher.
15878 (eq_string_slot_node): Likewise.
15880 * data-streamer-out.c: Update output_block::string_hash_table
15881 dependent calls and types.
15883 * dwarf2cfi.c (htab_t trace_index):
15884 Change type to hash_table. Update dependent calls and types.
15886 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15887 Change type to hash_table. Update dependent calls and types.
15888 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15889 (htab_t optimize_external_refs::map): Likewise.
15890 (htab_t output_comp_unit::extern_map): Likewise.
15891 (htab_t output_comdat_type_unit::extern_map): Likewise.
15892 (htab_t output_macinfo::macinfo_htab): Likewise.
15893 (htab_t optimize_location_lists::htab): Likewise.
15894 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15896 * except.c (htab_t ehspec_hash_type):
15897 Change type to hash_table. Update dependent calls and types.
15898 (assign_filter_values::ttypes): Likewise.
15899 (assign_filter_values::ehspec): Likewise.
15900 (sjlj_assign_call_site_values::ar_hash): Likewise.
15901 (convert_to_eh_region_ranges::ar_hash): Likewise.
15903 * gcse.c (htab_t pre_ldst_table):
15904 Change type to hash_table. Update dependent calls and types.
15906 * ggc-common.c (htab_t saving_htab):
15907 Change type to hash_table. Update dependent calls and types.
15908 (htab_t loc_hash): Likewise.
15909 (htab_t ptr_hash): Likewise.
15910 (call_count): Rename ggc_call_count.
15911 (call_alloc): Rename ggc_call_alloc.
15912 (loc_descriptor): Rename make_loc_descriptor.
15913 (add_statistics): Rename ggc_add_statistics.
15915 * ggc-common.c (saving_htab):
15916 Change type to hash_table. Update dependent calls and types.
15918 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15919 (push_gimplify_context): Likewise.
15920 (pop_gimplify_context): Likewise.
15921 (struct gimple_temp_hash_elt): Added.
15922 (struct gimplify_hasher): Likewise.
15923 (struct gimplify_ctx.temp_htab):
15924 Change type to hash_table. Update dependent calls and types.
15926 * gimple-fold.c: Include gimplify-ctx.h.
15928 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15929 Change type to hash_table. Update dependent calls and types.
15930 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15931 avoid potential global name collision.
15933 * gimplify.c: Include gimplify-ctx.h.
15934 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15935 (htab_t gimplify_ctx::temp_htab):
15936 Update dependent calls and types for new type hash_table.
15937 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15938 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15940 * gimplify-ctx.h: New.
15941 (struct gimple_temp_hash_elt): Move from gimplify.c.
15942 (class gimplify_hasher): New.
15943 (struct gimplify_ctx): Move from gimple.h.
15944 (htab_t gimplify_ctx::temp_htab):
15945 Change type to hash_table. Update dependent calls and types.
15947 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15948 (htab_t ivs_params::newivs_index):
15949 Change type to hash_table. Update dependent calls and types.
15950 (htab_t ivs_params::params_index): Likewise.
15951 (htab_t print_generated_program::params_index): Likewise.
15952 (htab_t gloog::newivs_index): Likewise.
15953 (htab_t gloog::params_index): Likewise.
15955 * graphite.c: Include graphite-htab.h.
15956 4htab_t graphite_transform_loops::bb_pbb_mapping):
15957 Change type to hash_table. Update dependent calls and types.
15959 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15960 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15961 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15963 * graphite-dependences.c: Include graphite-htab.h.
15964 (loop_is_parallel_p): Change hash table type of parameter.
15966 * graphite-htab.h: New.
15967 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15968 (extern find_pbb_via_hash): Move from graphite-poly.h.
15969 (extern loop_is_parallel_p): Move from graphite-poly.h.
15970 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15972 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15973 (extern loop_is_parallel_p): Move to graphite-htab.h.
15974 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15976 * haifa-sched.c (htab_t delay_htab):
15977 Change type to hash_table. Update dependent calls and types.
15978 (htab_t delay_htab_i2): Likewise.
15980 * ira-color.c (htab_t allocno_hard_regs_htab):
15981 Change type to hash_table. Update dependent calls and types.
15983 * ira-costs.c (htab_t cost_classes_htab):
15984 Change type to hash_table. Update dependent calls and types.
15986 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15987 Change type to hash_table. Update dependent calls and types.
15989 * loop-iv.c (htab_t bivs):
15990 Change type to hash_table. Update dependent calls and types.
15992 * loop-unroll.c (htab_t opt_info::insns_to_split):
15993 Change type to hash_table. Update dependent calls and types.
15994 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15996 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15997 (struct string_slot_hasher): New.
15998 (htab_t output_block::string_hash_table):
15999 Change type to hash_table. Update dependent calls and types.
16001 * lto-streamer-in.c (freeing_string_slot_hasher): New.
16002 (htab_t file_name_hash_table):
16003 Change type to hash_table. Update dependent calls and types.
16005 * lto-streamer-out.c: Update output_block::string_hash_table dependent
16008 * lto-streamer.c (htab_t tree_htab):
16009 Change type to hash_table. Update dependent calls and types.
16011 * omp-low.c: Include gimplify-ctx.h.
16013 * passes.c (htab_t name_to_pass_map):
16014 Change type to hash_table. Update dependent calls and types.
16015 (pass_traverse): Rename to passes_pass_traverse.
16017 * plugin.c (htab_t event_tab):
16018 Change type to hash_table. Update dependent calls and types.
16020 * postreload-gcse.c (htab_t expr_table):
16021 Change type to hash_table. Update dependent calls and types.
16022 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16024 * sese.c (debug_rename_map_1): Make extern.
16025 (htab_t copy_bb_and_scalar_dependences::rename_map):
16026 Change type to hash_table. Update dependent calls and types.
16028 * sese.h (extern debug_rename_map): Move to .c file.
16030 * store-motion.c (htab_t store_motion_mems_table):
16031 Change type to hash_table. Update dependent calls and types.
16033 * trans-mem.c (htab_t tm_new_mem_hash):
16034 Change type to hash_table. Update dependent calls and types.
16036 * tree-browser.c (htab_t TB_up_ht):
16037 Change type to hash_table. Update dependent calls and types.
16039 * tree-cfg.c (htab_t discriminator_per_locus):
16040 Change type to hash_table. Update dependent calls and types.
16042 * tree-complex.c: Include tree-hasher.h
16043 (htab_t complex_variable_components):
16044 Change type to hash_table. Update dependent calls and types.
16046 * tree-eh.c (htab_t finally_tree):
16047 Change type to hash_table. Update dependent calls and types.
16049 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16050 struct int_tree_hasher.
16051 (extern int_tree_map_eq): Likewise.
16052 (uid_decl_map_hash): Removed.
16053 (extern decl_tree_map_eq): Likewise.
16055 * tree-hasher.h: New.
16056 (struct int_tree_hasher): New.
16057 (typedef int_tree_htab_type): New.
16059 * tree-inline.c: Include gimplify-ctx.h.
16061 * tree-mudflap.c: Include gimplify-ctx.h.
16063 * tree-parloops.c: Include tree-hasher.h.
16064 (htab_t eliminate_local_variables_stmt::decl_address):
16065 Change type to hash_table. Update dependent calls and types.
16066 (htab_t separate_decls_in_region::decl_copies): Likewise.
16068 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16069 Change type to hash_table. Update dependent calls and types.
16071 * tree-sra.c (candidates):
16072 Change type to hash_table. Update dependent calls and types.
16074 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16076 (int_tree_map_hash): Likewise.
16078 * tree-ssa-dom.c (htab_t avail_exprs):
16079 Change type to hash_table. Update dependent calls and types.
16081 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16082 Change type to hash_table. Update dependent calls and types.
16084 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16085 Change type to hash_table. Update dependent calls and types.
16087 * tree-ssa-phiopt.c (seen_ssa_names):
16088 Change type to hash_table. Update dependent calls and types.
16090 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16091 Change type to hash_table. Update dependent calls and types.
16093 * tree-ssa-uncprop.c (equiv):
16094 Change type to hash_table. Update dependent calls and types.
16096 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16098 PR rtl-optimization/57003
16099 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16100 call note_stores with kill_clobbered_value callback again after
16101 killing regs_invalidated_by_call.
16103 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16105 * config/aarch64/aarch64-simd.md
16106 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16107 (aarch64_simd_bsl<mode>): Likewise.
16108 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16110 2013-04-25 Marek Polacek <polacek@redhat.com>
16112 PR tree-optimization/57066
16113 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16115 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16117 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16119 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16121 * config/aarch64/aarch64-builtins.c
16122 (aarch64_fold_builtin): New.
16123 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
16124 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
16125 * config/aarch64/aarch64-simd-builtins.def (abs): New.
16126 * config/aarch64/arm_neon.h
16127 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
16129 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16130 Tejas Belagod <tejas.belagod@arm.com>
16132 * config/aarch64/aarch64-builtins.c
16133 (aarch64_gimple_fold_builtin): New.
16134 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
16135 * config/aarch64/aarch64-simd-builtins.def (addv): New.
16136 * config/aarch64/aarch64-simd.md (addpv4sf): New.
16137 (addvv4sf): Update.
16138 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
16140 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16142 * config/aarch64/aarch64.md
16143 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
16145 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16147 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
16148 (*ngcsi_uxtw): New pattern.
16150 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16151 Julian Brown <julian@codesourcery.com>
16153 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
16154 (TB_DREG): Add T_V4HF.
16155 (v4hf_UP): New macro.
16156 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16157 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16158 Handle initialisation of V4HF. Adjust initialisation of reinterpret
16160 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16161 (arm_vector_mode_supported_p): Handle V4HF.
16162 (arm_mangle_map): Handle V4HFmode.
16163 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
16164 * config/arm/arm_neon_builtins.def: Add entries for
16165 vcvtv4hfv4sf, vcvtv4sfv4hf.
16166 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
16167 (neon_vcvtv4hfv4sf): Likewise.
16168 * config/arm/neon-gen.ml: Handle half-precision floating point
16170 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
16171 * config/arm/arm_neon.h: Regenerate.
16172 * config/arm/neon.ml (type elts): Add F16.
16173 (type vectype): Add T_float16x4, T_floatHF.
16174 (type vecmode): Add V4HF.
16175 (type features): Add Requires_FP_bit feature.
16176 (elt_width): Handle F16.
16177 (elt_class): Likewise.
16178 (elt_of_class_width): Likewise.
16179 (mode_of_elt): Refactor.
16180 (type_for_elt): Handle F16, fix error messages.
16181 (vectype_size): Handle T_float16x4.
16182 (vcvt_sh): New function.
16183 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
16184 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
16185 (string_of_mode): Handle V4HF.
16186 * doc/arm-neon-intrinsics.texi: Regenerate.
16188 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16190 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
16191 format specifier in 'X' case.
16193 2013-04-25 Alan Modra <amodra@gmail.com>
16196 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
16197 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
16198 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
16199 Repeat for many other rotate/shift and mask patterns using subregs.
16200 Name lshiftrt insns.
16201 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
16202 on WORDS_BIG_ENDIAN.
16204 2013-04-25 Alan Modra <amodra@gmail.com>
16206 * config.gcc: Support little-endian powerpc-linux targets.
16207 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
16208 (LINK_OS_LINUX_SPEC): Define.
16209 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
16210 Preserve MASK_LITTLE_ENDIAN.
16211 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
16212 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
16213 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
16214 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
16215 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
16216 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
16217 Correct fp word order for little-endian. Don't shift toc entries
16218 smaller than a word for little-endian.
16219 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
16220 (bswapdi2 splits): Correct low-part subreg for little-endian.
16221 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
16222 low/high where such is correct only for be.
16223 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
16224 little-endian for -mcall-aixdesc.
16226 2013-04-25 Alan Modra <amodra@gmail.com>
16228 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
16229 replace_equiv_address_nv.
16231 2013-04-25 Alan Modra <amodra@gmail.com>
16233 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
16235 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16238 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16239 * rtl.h (struct rtx_def): ...
16241 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16243 PR rtl-optimizations/57046
16244 * lra-constraints (split_reg): Set up lra_risky_transformations_p
16245 for multi-reg splits.
16247 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
16249 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16251 2013-04-24 Sterling Augustine <saugustine@google.com>
16253 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16254 (comp_dir_string, debug_str_dwo_section): New.
16255 (DEBUG_STR_DWO_SECTION): Rename to ...
16256 (DEBUG_DWO_STR_SECTION): ... this.
16257 (DEBUG_NORM_STR_SECTION): Delete.
16258 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16259 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16260 (find_AT_string): Move most logic to ...
16261 (find_AT_string_in_table): ... here. New.
16262 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16263 add_skeleton_AT_string. Delete logic.
16264 (output_skeleton_debug_sections): Remove call to
16265 add_top_level_skeleton_die_attrs.
16266 (add_comp_dir_attribute): Move logic to comp_dir_string.
16267 (dwarf2out_init): Initialize debug_str_dwo_section.
16268 (output_indirect_string): Call find_string_form.
16269 (output_indirect_strings): Rewrite.
16270 (prune_unused_types): Empty skeleton_debug_str_hash.
16271 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
16272 (dwarf2out_finish): Call output_indirect_strings.
16274 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16276 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
16278 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16280 * rtl.h (struct rtx_def): Add comment for field jump.
16281 (LRA_SUBREG_P): New macro.
16282 * recog.c (register_operand): Check LRA_SUBREG_P.
16283 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16285 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16286 (lra_final_code_change): Skip subreg change for operators.
16287 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16288 if there are no operand changes.
16289 * lra-constraints.c (curr_insn_set): New.
16290 (match_reload): Set LRA_SUBREG_P.
16291 (emit_spill_move): Ditto.
16292 (check_and_process_move): Use curr_insn_set. Process only single
16293 set insns. Don't initialize sec_mem_p and change_p.
16294 (simplify_operand_subreg): Use LRA_SUBREG_P.
16295 (reg_in_class_p): New function.
16296 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16297 of #ifdef. Add code to remove cycling.
16298 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16299 non-null disp. Reload inner instead of disp when base and index
16300 are null. Try to put lo_sum into register.
16301 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16302 (check_and_process_move): Move code for move cost check to
16303 simple_move_p. Remove equiv_substitution.
16304 (simple_move_p): New function.
16305 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16306 curr_insn_set. Call check_and_process_move only for single set
16307 insns. Use the new function. Move call of check_and_process_move
16308 after operand equiv substitution and address process.
16310 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
16312 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
16313 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
16314 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
16316 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16318 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
16320 2013-04-24 Marek Polacek <polacek@redhat.com>
16322 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
16323 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
16324 (select_loops_exit_conditions): Likewise.
16325 (number_of_iterations_for_all_loops): Likewise.
16326 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
16327 (scev_analysis): Likewise.
16329 2013-04-02 Catherine Moore <clm@codesourcery.com>
16330 Chao-ying Fu <fu@mips.com>
16332 * config/mips/micromips.md (jraddiusp): New pattern.
16333 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
16334 instruction if possible.
16336 2013-04-24 Alan Modra <amodra@gmail.com>
16338 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
16340 2013-04-24 Julian Brown <julian@codesourcery.com>
16341 Chung-Lin Tang <cltang@codesourcery.com>
16343 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
16344 dependency behavior in enumeration type DIE generation. Add TODO note
16345 to comments about future DW_FORM_sdata/udata re-work of related code.
16347 2013-04-23 Lawrence Crowl <crowl@google.com>
16349 * Makefile.in: Update as needed below.
16351 * hash-table.h (class hash_table):
16352 Correct many methods with parameter types compare_type to the correct
16353 value_type. (Correct code was unlikely to notice the change.)
16354 (hash_table::elements_with_deleted) New.
16355 (class hashtable::iterator): New.
16356 (hashtable::begin()): New.
16357 (hashtable::end()): New.
16358 (FOR_EACH_HASH_TABLE_ELEMENT): New.
16360 * statistics.c (statistics_hashes):
16361 Change type to hash_table. Update dependent calls and types.
16363 * tree-into-ssa.c (var_infos):
16364 Change type to hash_table. Update dependent calls and types.
16366 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
16367 Change type to hash_table. Update dependent calls and types.
16369 * tree-ssa-loop-im.c (struct mem_ref.refs):
16370 Change type to hash_table. Update dependent calls and types.
16372 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
16373 Change type to hash_table. Update dependent calls and types.
16375 * tree-ssa-sccvn.c (vn_tables_s::nary):
16376 Change type to hash_table. Update dependent calls and types.
16377 (vn_tables_s::phis): Likewise.
16378 (vn_tables_s::references): Likewise.
16380 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
16381 (vn_reference_eq): Update parameter and return types.
16383 * tree-ssa-structalias.c (pointer_equiv_class_table):
16384 Change type to hash_table. Update dependent calls and types.
16385 (location_equiv_class_table): Likewise.
16387 * tree-vect-data-refs.c: Consequential changes for making
16388 peeling a hash_table.
16390 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
16391 (destroy_loop_vec_info): Dependent hash_table update.
16393 * tree-vectorizer.h (peeling_htab):
16394 Change type to hash_table. Update dependent calls and types.
16396 2013-04-23 Shiva Chen <shiva0217@gmail.com>
16398 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
16399 to check the register content is equal or not.
16400 * lra-constraints.c (match_reload): Use lra_assign_reg_val
16401 to assign register content record.
16402 * lra-eliminations.c (update_reg_eliminate): Use
16403 lra_update_reg_val_offset to update register content offset.
16404 * lra-int.h (struct lra_reg): Add offset member.
16405 (lra_reg_val_equal_p): New static inline function.
16406 (lra_update_reg_val_offset): New static inline function.
16407 (lra_assign_reg_val): New static inline function.
16408 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
16409 to assign register content record.
16410 (initialize_lra_reg_info_element): Initial offset to zero.
16412 2013-04-23 Catherine Moore <clm@codesourcery.com>
16414 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
16415 operands. Record compression.
16417 2013-04-23 Xinliang David Li <davidxl@google.com>
16419 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
16421 2013-04-23 Richard Biener <rguenther@suse.de>
16423 PR middle-end/57036
16424 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
16425 parameter, only add abnormal goto edges from the copied body
16426 if the call could perform abnormal gotos.
16427 (copy_cfg_body): Adjust.
16429 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
16431 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
16433 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
16435 * coretypes.h (gimple_stmt_iterator): Add struct to make
16438 2013-04-23 Richard Biener <rguenther@suse.de>
16440 PR tree-optimization/57026
16441 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
16442 from SSA names occuring in abnormal PHI nodes.
16444 2013-04-22 Andi Kleen <ak@linux.intel.com>
16446 * lto/lto.c (print_lto_report_1): Fix LTO report names.
16448 2013-04-22 Andi Kleen <ak@linux.intel.com>
16450 * lto/lto.c (print_lto_report_1): Declare early.
16451 (read_cgraph_and_symbols): Call print_lto_report_1 early.
16453 2013-04-22 Andi Kleen <ak@linux.intel.com>
16455 * common.opt (-flto-report-wpa): Add.
16456 * doc/invoke.texi (-flto-report-wpa): Add.
16457 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
16460 2013-04-22 Xinliang David Li <davidxl@google.com>
16462 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
16463 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
16464 * Makefile.in: New dependency
16466 David Daney <ddaney.cavm@gmail.com>
16468 * configure.ac (gcc_cv_as_micromips_support): Use the
16469 --fatal-warnings option.
16470 * configure: Regenerate.
16472 2013-04-22 Marek Polacek <polacek@redhat.com>
16475 * tsan.c (instrument_expr): Don't instrument expression
16476 in case its size is zero.
16478 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
16482 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16484 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16486 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16488 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
16489 (gimple_stmt_iterator): New typedef.
16490 * gimple.h (gimple_stmt_iterator): Rename to...
16491 (gimple_stmt_iterator_d): ... This.
16492 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
16493 trees be valid for GIMPLE and GENERIC.
16494 (TARGET_GIMPLE_FOLD_BUILTIN): New.
16495 * gimple-fold.c (gimple_fold_call): Call target hook
16496 gimple_fold_builtin.
16497 * hooks.c (hook_bool_gsiptr_false): New.
16498 * hooks.h (hook_bool_gsiptr_false): New.
16499 * target.def (fold_stmt): New.
16500 * doc/tm.texi: Regenerate.
16502 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
16505 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
16506 a set sp if no stack realignment.
16508 2013-04-22 Nick Clifton <nickc@redhat.com>
16510 * config.gcc (tilegx-linux): Extend extra_objs rather than
16512 (tilepro-linux): Likewise.
16514 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16516 * config/aarch64/aarch64-builtins.c
16518 (CF0, CF1, CF2, CF3, CF4, CF10): New.
16519 (VAR<1-12>): Add MAP parameter.
16520 (BUILTIN_*): Likewise.
16521 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
16522 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
16523 (aarch64_ushl_n<mode>): Likewise.
16524 (aarch64_sshr_n<mode>): Likewise.
16525 (aarch64_ushr_n<mode>): Likewise.
16526 (aarch64_<maxmin><mode>): Likewise.
16527 (aarch64_sqrt<mode>): Likewise.
16528 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
16529 (vshr<q>_n_*): Likewise.
16531 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16533 * config/aarch64/aarch64-builtins.c
16534 (aarch64_simd_builtin_type_mode): Handle SF types.
16536 (BUILTIN_GPF): Define.
16537 (aarch64_init_simd_builtins): Handle SF types.
16538 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16539 (frecps): Likewise.
16540 (frecpx): Likewise.
16541 * config/aarch64/aarch64-simd.md
16542 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16543 (aarch64_frecpe<mode>): New.
16544 (aarch64_frecps<mode>): Likewise.
16545 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16546 (v8type): Add frecp<esx>.
16547 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16548 (aarch64_frecps<mode>): Likewise.
16549 * config/aarch64/iterators.md (FRECP): New.
16550 (frecp_suffix): Likewise.
16551 * config/aarch64/arm_neon.h
16552 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16554 2013-04-22 Christian Bruel <christian.bruel@st.com>
16557 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16558 (REG_CLASS_NAMES): Idem.
16559 (REG_CLASS_CONTENTS): Idem.
16560 (REGCLASS_HAS_FP_REG): Idem.
16561 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16562 (sh_conditional_register_usage): Idem.
16564 2013-04-21 Jeff Law <law@redhat.com>
16566 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16567 (ssa_forward_propagate_and_combine): Use it.
16569 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16571 * lra.c: Update the flow chart diagram.
16573 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16575 PR rtl-optimization/56847
16576 * lra-constraints.c (process_alt_operands): Discourage alternative
16577 with non-matche doffsettable memory constraint fro memory with
16580 2013-04-19 Richard Biener <rguenther@suse.de>
16582 PR tree-optimization/56982
16583 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16585 * gimplify.c (gimplify_call_expr): Notice special calls.
16586 (gimplify_modify_expr): Likewise.
16587 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16588 abnormal control flow receivers.
16589 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16590 in the same way as cfun->has_nonlocal_labels.
16591 (gimple_purge_dead_abnormal_call_edges): Likewise.
16592 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16593 receivers start a basic-block.
16595 2013-04-19 Richard Biener <rguenther@suse.de>
16597 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16599 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16600 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16601 (SLP_TREE_LOAD_PERMUTATION): Add.
16602 (vect_transform_slp_perm_load): Adjust prototype.
16603 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16604 (vect_free_slp_instance): Likewise.
16605 (vect_create_new_slp_node): Likewise.
16606 (vect_supported_slp_permutation_p): Remove.
16607 (vect_slp_rearrange_stmts): Adjust.
16608 (vect_supported_load_permutation_p): Likewise. Inline
16609 vect_supported_slp_permutation_p here.
16610 (vect_analyze_slp_instance): Compute load permutations per
16611 slp node instead of per instance.
16612 (vect_get_slp_defs): Adjust.
16613 (vect_transform_slp_perm_load): Likewise.
16614 (vect_schedule_slp_instance): Remove redundant code.
16615 (vect_schedule_slp): Remove hack for PR56270, add it ...
16616 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16617 CSE loads for SLP. Adjust.
16619 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16621 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16622 spelling in two comments.
16624 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16627 * config/arm/arm.c (load_multiple_sequence): Require SP
16628 as base register for loads if SP is in the register list.
16630 2013-04-19 Martin Jambor <mjambor@suse.cz>
16632 PR tree-optimization/56718
16633 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16634 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16635 and made public. Adjusted all callers.
16636 (ipa_intraprocedural_devirtualization): New function.
16637 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16638 (ipa_intraprocedural_devirtualization): Likewise.
16639 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16641 2013-04-19 Richard Biener <rguenther@suse.de>
16643 PR tree-optimization/57000
16644 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16646 2013-04-19 Terry Guo <terry.guo@arm.com>
16648 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16650 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16651 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16652 (cortex_m4_fmacs): Use new reservations.
16653 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16655 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16657 PR rtl-optimization/56999
16658 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16660 (lra_coalesce): Remove split_origin_bitmap and related code.
16661 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16662 ranges if necessary.
16664 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16666 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16668 (ix86_expand_call): Remove clobbered_registers array and use
16669 x86_64_ms_sysv_extra_clobbered_registers instead.
16670 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16672 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16674 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16675 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16676 (*call_value_rex64_ms_sysv): Ditto.
16678 2013-04-18 Cary Coutant <ccoutant@google.com>
16680 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16681 parent instead of die_mark.
16683 2013-04-18 Diego Novillo <dnovillo@google.com>
16685 * gimple.c (create_gimple_tmp): New.
16686 (get_expr_type): New.
16687 (build_assign): New.
16688 (build_type_cast): New.
16689 * gimple.h (enum ssa_mode): Define.
16690 (gimple_seq_set_location): New.
16691 * asan.c (build_check_stmt): Change some gimple_build_* calls
16692 to use build_assign and build_type_cast.
16694 2013-04-18 Richard Biener <rguenther@suse.de>
16696 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16697 handle negative step. Remove redundant checks.
16698 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16699 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16700 for negative step and grouped loads fail to vectorize.
16702 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16704 * emit-rtl.c (reset_insn_used_flags): New function.
16705 (reset_all_used_flags): Use it.
16706 (verify_insn_sharing): New function.
16707 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16709 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16711 PR tree-optimization/56984
16712 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16713 and (x >> M) >= N don't register any assertion if N << M is the
16716 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16718 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16719 do not delete_insn insns that have not yet been emitted, only
16720 unlink them with remove_insn.
16721 * df-scan.c (df_insn_delete): Revert r197492.
16723 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16725 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16726 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16728 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16730 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16731 define_insn_and_split.
16732 (and_scc,ior_scc,negscc): Likewise.
16733 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16735 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16737 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16738 can benefit from using a sequence of LDRD instructions in epilogue
16739 instead of a single LDM instruction.
16741 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16744 * doc/extend.texi: Fix typo.
16746 2013-04-17 Richard Biener <rguenther@suse.de>
16748 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16749 (vect_build_slp_tree): ... here.
16750 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16751 match. Remove special-casing of mismatched complex loads.
16752 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16753 re-try the match with swapped commutative operands.
16754 (vect_supported_load_permutation_p): Remove special-casing of
16755 mismatched complex loads.
16756 (vect_analyze_slp_instance): Adjust.
16758 2013-04-17 Richard Biener <rguenther@suse.de>
16760 PR rtl-optimization/56921
16761 * cfgloop.h (struct loop): Add simple_loop_desc member.
16762 (struct niter_desc): Mark with GTY(()).
16763 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16764 * loop-iv.c (get_simple_loop_desc): Likewise.
16765 (free_simple_loop_desc): Likewise.
16768 2013-04-16 Richard Biener <rguenther@suse.de>
16770 PR rtl-optimization/56921
16771 * loop-init.c (pass_rtl_move_loop_invariants): Add
16772 TODO_do_not_ggc_collect to todo_flags_finish.
16773 (pass_rtl_unswitch): Same.
16774 (pass_rtl_unroll_and_peel_loops): Same.
16775 (pass_rtl_doloop): Same.
16777 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16779 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16780 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16781 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16783 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16784 * tree-streamer.c (record_common_node): Adjust reference in comment.
16786 2013-04-17 Terry Guo <terry.guo@arm.com>
16788 * config/arm/cortex-m4.md: Add a new bypass.
16790 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16792 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16794 (*subs_<optab><mode>_multp2): New pattern.
16795 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16796 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16798 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16800 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16801 (*subs_mul_imm_<mode>): New pattern.
16803 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16806 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16807 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16808 (vsx_movti_32bit): Change j->wa to O->wa.
16810 2013-04-16 Richard Biener <rguenther@suse.de>
16812 PR rtl-optimization/56921
16813 * loop-init.c (pass_rtl_move_loop_invariants): Add
16814 TODO_do_not_ggc_collect to todo_flags_finish.
16815 (pass_rtl_unswitch): Same.
16816 (pass_rtl_unroll_and_peel_loops): Same.
16817 (pass_rtl_doloop): Same.
16819 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16821 * config/arm/arm.c (emit_multi_reg_push): New declaration
16822 for an existing function.
16823 (arm_emit_strd_push): New function.
16824 (arm_expand_prologue): Used here.
16825 (arm_emit_ldrd_pop): New function.
16826 (arm_expand_epilogue): Used here.
16827 (arm_get_frame_offsets): Update condition.
16828 (arm_emit_multi_reg_pop): Add a special case for load of a single
16829 register with writeback.
16831 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16833 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16836 2013-04-16 Richard Biener <rguenther@suse.de>
16838 PR tree-optimization/56756
16839 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16840 (first_mem_ref_loc): New.
16841 (execute_sm): Place the load temporarily before a previous
16842 access instead of in the latch edge to ensure its SSA dependencies
16843 are defined at points dominating the load.
16845 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16847 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16848 correct fix by moving header and footer insn to the footer of
16849 the merged basic block. Clear BB_END of the merged-away block.
16851 PR middle-end/43631
16852 * emit-rtl.c (make_note_raw): New function.
16853 (link_insn_into_chain): New static inline function.
16854 (add_insn): Use it.
16855 (add_insn_before, add_insn_after): Factor insn chain linking code...
16856 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16857 using link_insn_into_chain.
16858 (note_outside_basic_block_p): New helper function for emit_note_after
16859 and emit_note_before.
16860 (emit_note_after): Use nobb variant of add_insn_after if the note
16861 should not be contained in a basic block.
16862 (emit_note_before): Use nobb variant of add_insn_before if the note
16863 should not be contained in a basic block.
16864 (emit_note_copy): Use make_note_raw.
16865 (emit_note): Likewise.
16866 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16867 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16868 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16869 the moved barrier the tail of the basic block it follows.
16870 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16872 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16874 PR tree-optimization/56962
16875 * gimple-ssa-strength-reduction.c (record_increment): Only set
16876 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16877 either rhs1 or rhs2 is equal to c->base_expr.
16879 2013-04-15 Richard Biener <rguenther@suse.de>
16881 PR tree-optimization/56933
16882 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16884 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16885 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16886 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16887 dependence check ...
16888 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16890 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16891 GROUP_READ_WRITE_DEPENDENCE.
16893 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16895 * emit-rtl.c (reset_all_used_flags): New function.
16896 (verify_rtl_sharing): Call reset_all_used_flags before and after
16897 performing the checks.
16899 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16901 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16902 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16903 * config/arm/constraints.md (De): New constraint.
16904 * config/arm/neon.md (anddi3_neon): Delete.
16905 (neon_vand<mode>): Expand to standard anddi3 pattern.
16906 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16907 Move earlier in the file.
16908 (neon_inv_logic_op2): Likewise.
16909 (arm_anddi_operand_neon): New predicate.
16911 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16913 * configure.ac (gcc_cv_ld_as_needed): Set
16914 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16915 Use -z ignore, -z record on *-*-solaris2*.
16916 (HAVE_LD_AS_NEEDED): Update comment.
16917 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16918 * configure: Regenerate.
16919 * config.in: Regenerate.
16920 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16921 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16922 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16923 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16924 equivalents. Fix markup.
16925 * doc/tm.texi: Regenerate.
16927 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16929 * config/i386/i386.opt: New option mstack-protector-guard=.
16930 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16931 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16932 TARGET_SSP_TLS_GUARD.
16933 * config/i386/i386.c (ix86_option_override_internal): Set
16934 ix86_stack_protector_guard.
16935 * config/i386/i386.md (stack_protect_set): Enable for
16936 TARGET_SSP_TLS_GUARD only.
16937 (stack_protect_set_<mode>): Ditto.
16938 (stack_protect_test): Ditto.
16939 (stack_protect_test_<mode>): Ditto.
16940 * doc/invoke.texi (i386 Option): Document.
16942 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16945 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16946 (S_MODES): Set H_MODE bit.
16947 (SF_MODES): Set only S_MODE and SF_MODE bits.
16948 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16949 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16950 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16951 <MODE_FLOAT>: Likewise.
16953 2013-04-15 Joey Ye <joey.ye@arm.com>
16955 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16957 2013-04-15 Joey Ye <joey.ye@arm.com>
16959 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16961 (thumb_far_jump_used_p): Count instruction size and set
16964 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16966 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16967 * resource.c (mark_target_live_regs): Reformat conditional block.
16969 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16971 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16972 notes, they are emitted only just before final.
16973 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16975 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16977 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16978 * cfgrtl.c (delete_insn): Call it here instead.
16979 * lra-spills.c (lra_final_code_change): Use delete_insn.
16980 * haifa-sched.c (sched_remove_insn): Likewise.
16981 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16982 returning to the nop pool.
16983 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16984 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16986 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16988 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16989 * doc/tm.texi: Regenerated.
16991 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16993 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16996 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16998 * df-core.c (df_find_def): Compare register numbers.
16999 (df_find_use): Likewise.
17001 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
17004 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
17005 lra_in_progress for return.
17007 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17009 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
17010 define_insn into define_insn_and_split and emit movsicc patterns.
17012 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17014 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17016 2013-04-12 Richard Biener <rguenther@suse.de>
17018 * tree-pass.h (TODO_do_not_ggc_collect): New.
17019 * passes.c (execute_one_ipa_transform_pass): Honor
17020 TODO_do_not_ggc_collect.
17021 (execute_one_pass): Likewise.
17024 2013-04-10 Richard Biener <rguenther@suse.de>
17026 * passes.c (init_optimization_passes): Remove reload pass.
17027 * ira.c (do_reload): Merge into ...
17029 (rest_of_handle_reload): Remove.
17030 (pass_reload): Likewise.
17031 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17032 of reload for vzeroupper pass placement.
17034 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17036 PR tree-optimization/56918
17037 PR tree-optimization/56920
17038 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17039 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17040 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17041 use wide_mul_with_sign method.
17043 2013-04-12 Richard Biener <rguenther@suse.de>
17045 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17046 not be considered a gimple constant.
17048 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17050 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17051 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17054 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17055 Jakub Jelinek <jakub@redhat.com>
17057 * opts.c: Include diagnostic-color.h.
17058 (common_handle_option): Handle OPT_fdiagnostics_color_.
17059 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17060 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17061 (diagnostic-color.o): New.
17062 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17063 (diagnostic_color_rule): New enum.
17064 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17065 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17066 the location string.
17067 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17068 either NULL, or color kind.
17069 * diagnostic-color.c: New file.
17070 * diagnostic-color.h: New file.
17071 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17073 * doc/invoke.texi (-fdiagnostics-color): Document.
17074 * pretty-print.h (pp_show_color): Define.
17075 (struct pretty_print_info): Add show_color field.
17076 * diagnostic.c: Include diagnostic-color.h.
17077 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17078 macros. Colorize error:, warning: etc. strings and also the location
17080 (diagnostic_show_locus): Colorize the caret line.
17081 * pretty-print.c: Include diagnostic-color.h.
17082 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17083 inside of %< %> quotes or quoted through q format modifier.
17085 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17087 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17089 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17091 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17092 code in CC_NZ mode.
17093 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17096 2013-04-11 Marek Polacek <polacek@redhat.com>
17098 PR tree-optimization/48184
17099 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17101 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17103 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17104 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17105 (skip_simple_arithmetic): Tidy up.
17106 * tree.h (skip_simple_constant_arithmetic): Declare.
17108 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17110 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17112 2013-04-11 Richard Biener <rguenther@suse.de>
17114 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17115 generate vector constants.
17117 2013-04-11 Richard Biener <rguenther@suse.de>
17119 PR tree-optimization/56878
17120 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17121 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17123 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
17124 Prefer to align the DR with the most invariant base address.
17126 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17128 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
17131 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
17133 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
17134 floating-point vector comparisons against 0.
17136 2013-04-11 Jakub Jelinek <jakub@redhat.com>
17138 PR tree-optimization/56899
17139 * fold-const.c (extract_muldiv_1): Apply distributive law
17140 only if TYPE_OVERFLOW_WRAPS (ctype).
17142 2013-04-11 Bin Cheng <bin.cheng@arm.com>
17145 * ira-costs.c (scan_one_insn): Check whether the source rtx of
17146 loading has side effect.
17148 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
17150 * config/sparc/sparc.c: Include tree-pass.h.
17151 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
17152 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
17153 head of file. Change return type. Split off gate function.
17154 (sparc_gate_work_around_errata): New function.
17155 (pass_work_around_errata): New pass definition.
17156 (insert_pass_work_around_errata) New pass insert definition to
17157 insert pass_work_around_errata just after delayed-branch scheduling.
17158 (sparc_option_override): Insert the pass.
17159 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
17161 2013-04-10 David S. Miller <davem@davemloft.net>
17163 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
17164 or -mcpu=hypersparc.
17166 * target.def (cstore_mode): New hook.
17167 * target.h: Include insn-codes.h
17168 * targhooks.c: Likewise.
17169 (default_cstore_mode): New function.
17170 * targhooks.h: Declare it.
17171 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
17172 * doc/tm.texi: Rebuild.
17173 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
17174 target hook, rather than inspecting the insn_data.
17175 * config/sparc/sparc.c (sparc_cstore_mode): New function.
17176 (TARGET_CSTORE_MODE): Redefine.
17177 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
17179 * config/sparc/predicates.md (cstore_result_operand): New special
17181 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
17182 Use it for operand 0.
17183 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
17184 (*snesi_special): Likewise.
17185 (*snesi_zero): Likewise.
17186 (*seqsi_zero): Likewise.
17187 (*sltu_insn): Likewise.
17188 (*sgeu_insn): Likewise.
17189 (*seqdi_special): Make operand 0 and comparison operation be of
17191 (*snedi_special): Likewise.
17192 (*snedi_special_vis3): Likewise.
17193 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
17194 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
17195 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
17196 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
17197 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
17198 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
17199 (*sltu_extend_sp64): Likewise.
17200 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
17201 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
17202 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
17203 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
17204 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
17206 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
17208 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
17209 (aarch64_start_file): Use the new function.
17211 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17212 Jason Merrill <jason@redhat.com>
17214 * common.opt: Add -gdwarf.
17215 * opts.c (common_handle_option): Handle it.
17216 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
17218 2013-04-10 Richard Biener <rguenther@suse.de>
17220 * passes.c (execute_todo): Do not call ggc_collect conditional here.
17221 (execute_one_ipa_transform_pass): But unconditionally here.
17222 (execute_one_pass): And here.
17223 (init_optimization_passes): Remove reload pass.
17224 * tree-pass.h (TODO_ggc_collect): Remove.
17225 (pass_reload): Likewise.
17226 * ira.c (do_reload): Merge into ...
17228 (rest_of_handle_reload): Remove.
17229 (pass_reload): Likewise.
17230 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17231 of reload for vzeroupper pass placement.
17232 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
17233 and todo_flags_finish of all passes.
17235 2013-04-10 Richard Biener <rguenther@suse.de>
17237 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
17238 first_const_oprnd field, rename first_def_type to first_op_type.
17239 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
17240 (vect_get_and_check_slp_defs): Always use the type of the
17241 operand. Allow mixed vect_external_def, vect_constant_def types.
17242 (vect_get_constant_vectors): Handle mixed vect_external_def,
17243 vect_constant_def types.
17245 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
17247 PR tree-optimization/55524
17248 * tree-ssa-math-opts.c
17249 (convert_mult_to_fma): Don't use an fms construct
17250 when we don't have an fms operation, but fnma, and it looks
17251 likely that we'll be able to use the latter.
17253 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17255 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17257 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17258 inline fail caused by overwritable functions.
17260 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17262 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17263 unnecessary bits in the constant power of two case.
17265 2013-04-10 Richard Biener <rguenther@suse.de>
17267 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17268 broken code swapping operands.
17269 (vect_build_slp_tree): Do not compute load permutations here.
17270 (vect_analyze_slp_instance): Compute load permutations here,
17271 after building the SLP tree.
17273 2013-04-09 Christian Bruel <christian.bruel@st.com>
17275 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
17276 of next/prev_real_insn.
17278 2013-04-09 Jan Hubicka <jh@suse.cz>
17280 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17281 Drop aliased parameter.
17282 (function_and_variable_visibility): Do not handle alias pairs.
17283 * cgraph.c (varpool_externally_visible_p): Update prototype.
17284 * varpool.c (varpool_add_new_variable): Update.
17286 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17288 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
17290 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
17292 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
17294 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
17296 2013-04-09 Marek Polacek <polacek@redhat.com>
17298 PR tree-optimization/48762
17299 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
17301 2013-04-09 Richard Biener <rguenther@suse.de>
17303 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
17305 (vect_build_slp_tree): Likewise.
17306 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
17307 calculating the cost of a SLP instance.
17308 (vect_analyze_slp_instance): Use it from here, after building
17311 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17313 PR middle-end/56883
17314 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
17315 expand_omp_for_static_chunk): Use simple_p = true in
17316 force_gimple_operand_gsi calls when assigning to addressable decls.
17318 2013-04-09 Jeff Law <law@redhat.com>
17320 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
17321 when the boolean was created by converting a wider object which
17322 had a boolean range.
17324 2013-04-09 Richard Biener <rguenther@suse.de>
17326 * tree-vectorizer.h (slp_void_p): Remove.
17327 (slp_tree): Typedef before _slp_tree declaration.
17328 (struct _slp_tree): Use a vector of slp_tree as children.
17329 (vect_get_place_in_interleaving_chain): Remove.
17330 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
17332 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
17334 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
17335 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
17336 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
17337 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
17338 Use slp_node instead of slp_void_p and adjust.
17340 2013-04-09 Richard Biener <rguenther@suse.de>
17342 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
17343 work that is not necessary.
17345 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17347 PR tree-optimization/56854
17348 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
17349 forward into clobber stmts if it would change MEM_REF lhs into
17352 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
17354 * tree.c (type_hash_lookup, type_hash_add): Make static.
17355 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
17357 2013-04-09 Richard Biener <rguenther@suse.de>
17359 * tree.h (unsave_expr_now): Remove.
17360 * tree-inline.c (mark_local_for_remap_r): Remove.
17361 (unsave_expr_1): Likewise.
17362 (unsave_r): Likewise.
17363 (unsave_expr_now): Likewise.
17364 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
17365 (propagate_tree_value): Likewise.
17367 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
17369 * doc/rtl.texi (sequence): Rewrite documentation to match the
17370 current use of SEQUENCE rtl objects.
17371 * rtl.def (SEQUENCE): Likewise.
17373 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
17374 Update documentation.
17375 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
17376 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
17378 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
17380 2013-04-08 Teresa Johnson <tejohnson@google.com>
17382 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
17383 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
17385 (estimate_edge_size_and_time): Add comment to suggest using rounding
17387 (estimate_node_size_and_time): Ditto.
17388 (remap_edge_change_prob): Use helper rounding divide methods.
17389 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
17390 (gimple_mod_pow2_value_transform): Ditto.
17391 (gimple_mod_subtract_transform): Ditto.
17392 (gimple_ic_transform): Ditto.
17393 (gimple_stringops_transform): Ditto.
17394 * stmt.c (conditional_probability): Ditto.
17395 (emit_case_dispatch_table): Ditto.
17396 * lto-cgraph.c (merge_profile_summaries): Ditto.
17397 * tree-optimize.c (execute_fixup_cfg): Ditto.
17398 * cfgcleanup.c (try_forward_edges): Ditto.
17399 * cfgloopmanip.c (scale_loop_profile): Ditto.
17401 (duplicate_loop_to_header_edge): Ditto.
17402 (lv_adjust_loop_entry_edge): Ditto.
17403 * tree-vect-loop.c (vect_transform_loop): Ditto.
17404 * profile.c (compute_branch_probabilities): Ditto.
17405 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
17406 * lto-streamer-in.c (input_cfg): Ditto.
17407 * gimple-streamer-in.c (input_bb): Ditto.
17408 * ipa-cp.c (update_profiling_info): Ditto.
17409 (update_specialized_profile): Ditto.
17410 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
17411 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
17413 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17414 (compute_trg_info): Ditto.
17415 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
17416 (purge_dead_edges): Ditto.
17417 * loop-unswitch.c (unswitch_loop): Ditto.
17418 * cgraphclones.c (cgraph_clone_edge): Ditto.
17419 (cgraph_clone_node): Ditto.
17420 * tree-inline.c (copy_bb): Ditto.
17421 (copy_edges_for_bb): Ditto.
17422 (initialize_cfun): Ditto.
17423 (copy_cfg_body): Ditto.
17424 (expand_call_inline): Ditto.
17426 2013-04-08 Kai Tietz <ktietz@redhat.com>
17428 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
17429 TARGET_CYGWIN64 by TARGET_64BIT.
17431 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17433 * config/epiphany/epiphany.md (GPR_1): New constant.
17434 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
17435 * config/epiphany/epiphany.c (gen_compare_reg):
17436 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
17437 is already in place.
17438 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
17439 Don't require being called during rtl expansion; If y operlaps r0,
17441 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
17442 (epiphany_expand_epilogue): Likewise.
17444 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
17445 Don't use CC_FPmode for ORDERED / UNORDERED.
17446 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
17448 * config/epiphany/constraints.md (CnL): New constraint.
17449 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
17450 * config/epiphany/predicates.md (add_operand): Allow 1024.
17452 * config/epiphany/epiphany.md (logical_op): New code iterator.
17453 (op_mnc): New code attribute.
17454 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
17455 (mov_f+1, mov_f+2): New peephole2 patterns.
17457 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
17458 (cstoresi4): Also allow re-use of zero result when doing a NE
17459 comparison to a non-zero operand.
17460 Use (clobber (scratch)) for first insn if the gpr output is not needed.
17462 * config/epiphany/epiphany.md (<insn_opname>v2si3):
17463 Use gen_addsi3_i / gen_subsi3_i.
17465 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17469 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
17470 contain anything but clobbers, at most one __builtin_stack_restore,
17471 optionally debug stmts and final resx, and if it has at least one
17472 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
17473 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
17474 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
17475 which isn't defaut definition, remove them.
17476 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
17477 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
17478 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
17479 with MEM_REF LHS with SSA_NAME address.
17481 2013-04-08 Jeff Law <law@redhat.com>
17483 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
17485 2013-04-08 Richard Biener <rguenther@suse.de>
17487 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
17489 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
17490 determined vector type.
17491 (vect_analyze_data_refs): Likewise.
17492 (vect_get_new_vect_var): Adjust.
17493 (vect_create_destination_var): Preserve SSA name versions.
17494 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
17495 not dump anything here.
17497 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17499 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
17500 Add member lr_slot_known.
17501 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
17503 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17504 Remove code that sets lr_slot_offset according to what a previous
17505 version of epiphany_emit_save_restore used to do.
17506 (epiphany_emit_save_restore): When doing an lr save or restore,
17507 set/verify lr_slot_known and lr_slot_offset.
17509 2013-04-08 Xinyu Qi <xyqi@marvell.com>
17512 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
17515 2013-04-08 Richard Biener <rguenther@suse.de>
17517 * alias.c (find_base_term): Fix thinko in previous change.
17519 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17521 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
17522 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
17523 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
17524 if possible to compute val.
17525 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
17526 For QImode integers don't require anything about precision. Use
17527 const_with_all_bytes_same to find out if the constant doesn't have
17528 repeated bytes in it.
17530 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17532 * config/s390/s390.c (s390_expand_insv): Only accept insertions
17535 2013-04-08 Marek Polacek <polacek@redhat.com>
17537 PR rtl-optimization/48182
17538 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17541 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17544 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17545 nuses, make sure we have a label.
17547 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17550 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17551 (rs6000_emit_swdiv_low_precision): Remove.
17552 (rs6000_emit_swdiv): Rewrite to handle between one and four
17553 iterations of Newton-Raphson generally; modify required number of
17554 iterations for some cases.
17555 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17557 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17559 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17560 set-but-unused variable.
17562 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17563 basic blocks of released function bodies garbage-collectable.
17565 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17566 (struct rtl_opt_pass): Add TODO_df_finish.
17568 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17570 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17572 * config/arm/constraints.md (q): New constraint.
17573 * config/arm/ldrdstrd.md: New file.
17574 * config/arm/arm.md (ldrdstrd.md) New include.
17575 (arm_movdi): Use "q" instead of "r" constraint
17576 for double-word memory access.
17577 (movdf_soft_insn): Likewise.
17578 * config/arm/vfp.md (movdi_vfp): Likewise.
17579 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17580 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17581 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17582 (mem_ok_for_ldrd_strd): Likewise.
17583 (output_move_double): Update assertion.
17585 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17587 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17589 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17591 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17592 define_insn_and_split.
17593 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17595 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17597 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17598 define_insn_and_split.
17599 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17600 (shiftsi3_compare): New pattern.
17601 (rrx): New pattern.
17602 * config/arm/unspecs.md (UNSPEC_RRX): New.
17604 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17606 * config/arm/arm.md (negdi_extendsidi): New pattern.
17607 (negdi_zero_extendsidi): Likewise.
17609 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17611 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17612 define_insn_and_split.
17613 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17614 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17616 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17618 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17619 define_insn_and_split.
17620 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17621 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17623 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17625 * config/arm/arm.md (subsi3_carryin): New pattern.
17626 (subsi3_carryin_const): Likewise.
17627 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17628 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17630 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17632 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17634 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17636 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17637 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17639 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17641 * config/arm/arm.c (arm_expand_builtin): Change fcode
17642 type to unsigned int.
17644 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17646 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17648 2013-04-04 Ian Lance Taylor <iant@google.com>
17650 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17653 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17655 PR middle-end/56729
17656 * df-scan.c (df_insn_delete): Disable failing assert.
17658 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17660 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17661 New function prototype.
17662 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17663 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17664 (arm_builtin_vectorized_function): New function.
17666 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17668 * config/arm/arm_neon_builtins.def: New file.
17669 * config/arm/arm.c (neon_builtin_data): Move contents to
17670 arm_neon_builtins.def.
17671 (enum arm_builtins): Include neon builtin definitions.
17672 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17673 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17675 2013-04-04 Marek Polacek <polacek@redhat.com>
17677 PR tree-optimization/48186
17678 * predict.c (maybe_hot_frequency_p): Return false if
17679 HOT_BB_FREQUENCY_FRACTION is 0.
17680 (cgraph_maybe_hot_edge_p): Likewise.
17682 2013-04-04 Richard Biener <rguenther@suse.de>
17684 PR tree-optimization/56826
17685 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17688 2013-04-04 Richard Biener <rguenther@suse.de>
17690 PR tree-optimization/56213
17691 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17692 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17694 2013-04-04 Richard Biener <rguenther@suse.de>
17696 PR tree-optimization/56837
17697 * tree-loop-distribution.c (classify_partition): For non-zero
17698 values require that the value has the same precision as its
17699 mode to be useful as memset value.
17701 2013-04-03 Nick Clifton <nickc@redhat.com>
17703 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17704 (fmssf4): Use fmsf.s on E3V5 architectures.
17705 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17706 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17708 2013-04-03 Jeff Law <law@redhat.com>
17710 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17711 (lra-eliminations.o): Likewise.
17713 2013-04-03 Teresa Johnson <tejohnson@google.com>
17715 * gcov-io.c (compute_working_sets): Moved most of body of old
17716 compute_working_sets here from profile.c.
17717 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17718 (gcov_working_set_t): Moved typedef here from basic-block.h
17719 (compute_working_set): Declare.
17720 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17721 (get_working_sets): Renamed from compute_working_set,
17722 replace most of body with call to new compute_working_sets.
17723 (get_exec_counts): Replace call to compute_working_sets
17724 to get_working_sets.
17725 * profile.h (get_working_sets): Renamed from compute_working_set.
17726 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17727 to get_working_sets.
17728 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17729 * gcov-dump.c (dump_working_sets): New function.
17731 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17733 * hwint.c (sext_hwi, zext_hwi): New functions.
17734 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17735 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17736 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17737 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17738 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17739 (sext_hwi, zext_hwi): New functions.
17741 2013-04-03 Jeff Law <law@redhat.com>
17743 PR tree-optimization/56799
17744 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17745 back test for widening conversion erroneously dropped in prior change.
17747 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17750 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17751 instead of next_real_insn.
17753 2013-04-03 Marek Polacek <polacek@redhat.com>
17756 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17758 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17761 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17763 (thumb1_output_casesi): Likewise.
17764 (thumb2_output_casesi): Likewise.
17766 2013-04-03 Richard Biener <rguenther@suse.de>
17768 PR tree-optimization/56817
17769 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17771 (tree_unroll_loops_completely_1): ... new function to manually
17772 walk the loop tree, properly defering outer loops of unrolled
17773 loops to later iterations.
17775 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17777 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17778 (vectorizable_load): Likewise.
17779 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17780 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17782 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17784 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17787 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17789 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17791 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17793 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17795 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17797 PR tree-optimization/56790
17798 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17801 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17803 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17805 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17806 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17809 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17812 * tree.h (force_folding_builtin_constant_p): New decl.
17813 * builtins.c (force_folding_builtin_constant_p): New variable.
17814 (fold_builtin_constant_p): Fold immediately also if
17815 force_folding_builtin_constant_p.
17817 2013-04-03 Richard Biener <rguenther@suse.de>
17819 PR tree-optimization/56812
17820 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17821 DRs of the same interleaving chain are independent.
17823 2013-04-02 Jason Merrill <jason@redhat.com>
17825 * gdbinit.in (pbb): Use debug fn.
17827 2013-04-02 Lawrence Crowl <crowl@google.com>
17829 * sese.h (struct ivtype_map_elt_s): Remove unused.
17830 (extern debug_ivtype_map): Remove unused.
17831 (extern eq_ivtype_map_elts): Remove unused.
17832 * sese.c (debug_ivtype_map): Removed unused.
17833 (debug_ivtype_map_1): Removed unused.
17834 (debug_ivtype_elt): Remove unused.
17835 (eq_ivtype_map_elts): Remove unused.
17838 2013-04-02 Kai Tietz <ktietz@redhat.com>
17841 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17842 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17843 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17845 (legitimize_pe_coff_symbol): Likewise.
17846 (is_imported_p): New helper-function.
17847 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17848 for Windows x64 targets.
17849 (ix86_expand_prologue): Optimize for pe-coff targets.
17850 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17851 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17852 medium/large code-model.
17853 (legitimize_pic_address): Likewise.
17854 (legitimize_tls_address): Likewise.
17855 (ix86_expand_call): Likewise.
17856 (x86_output_mi_thunk): Likewise.
17857 (get_dllimport_decl): Add new beimport argument.
17858 (construct_plt_address): Don't assert for x64 pe-coff targets.
17859 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17861 (SYMBOL_FLAG_STUBVAR): New macro.
17862 (SYMBOL_REF_STUBVAR_P): Likewise.
17863 * config/i386/winnt.c (stub_list): New structure.
17864 (stub_head): New local variable.
17865 (i386_pe_record_stub): New function.
17866 (i386_pe_file_end): Emit refptr-stubs.
17868 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17870 PR rtl-optimization/56745
17871 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17872 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17875 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17876 and both of them are MEM_REFs, just compare first argument for
17877 equality and attempt to deal even with differing offsets.
17880 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17881 of gimple_clobber_p to be MEM_REF.
17882 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17883 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17884 after gimplification.
17885 * asan.c (get_mem_ref_of_assignment): Don't instrument
17886 gimple_clobber_p stmts.
17887 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17888 gimple_clobber_p stmt if they have MEM_REF lhs and
17889 are dead because of another gimple_clobber_p stmt.
17890 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17891 gimple_clobber_p stmts like debug stmts.
17892 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17893 that refer to unused VAR_DECLs or uninitialized values.
17894 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17895 gimple_clobber_p stmts if they refer to removed parameters.
17896 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17899 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17901 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17902 using SWI48 mode attribute.
17904 2013-04-02 Wei Mi <wmi@google.com>
17906 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17907 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17908 *<rotate_insn><mode>3_mask in i386.md.
17910 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17912 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17914 2013-04-02 Richard Biener <rguenther@suse.de>
17916 PR tree-optimization/56778
17917 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17918 Runtime alias tests are not supported for gather loads.
17919 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17920 stmts referenced from SSA operands before updating SSA form.
17922 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17923 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17925 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17926 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17927 * config/arm/cortex-a53.md: New file.
17928 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17929 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17930 * config/arm/arm.c (arm_issue_rate): Likewise.
17931 * config/arm/arm-tune.md: Regenerate
17932 * config/arm/arm-tables.opt: Regenerate.
17933 * config/arm/arm-cores.def: Add cortex-a53.
17935 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17937 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17940 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17942 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17943 scalar load/store operations using B/H registers.
17944 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17946 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17948 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17950 * config/aarch64/aarch64.c
17951 (aarch64_simd_scalar_immediate_valid_for_move): New.
17952 * config/aarch64/aarch64-protos.h
17953 (aarch64_simd_scalar_immediate_valid_for_move): New.
17954 * config/aarch64/constraints.md (Dh, Dq): New.
17955 * config/aarch64/iterators.md (hq): New.
17957 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17959 * reorg.c (get_branch_condition): Deal with conditional returns.
17960 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17962 2013-04-01 Wei Mi <wmi@google.com>
17964 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17965 Truncate operand 2 using %b asm operand modifier.
17966 (*<shift_insn><mode>3_mask): Ditto.
17967 (*<rotate_insn><mode>3_mask): Ditto.
17969 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17971 PR middle-end/56798
17972 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17974 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17976 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17978 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17980 2013-03-30 Lawrence Crowl <crowl@google.com>
17982 * dse.c (clear_alias_sets): Remove never set.
17983 (disqualified_clear_alias_sets): Remove never set.
17984 (clear_alias_mode_pool): Remove never set.
17985 (dse_step0): Remove condition that is never true.
17986 (canon_address): Remove condition that is never true.
17987 (dse_step7): Remove condition that is never true.
17988 (rest_of_handle_dse): Remove condition that is never true.
17989 (rest_of_handle_dse::did_global): Remove never read from above.
17990 (dse_step2_spill): Remove never called from above.
17991 (dse_step5_spill): Remove never called from above.
17993 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17995 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17996 JUMP_TABLE_DATA changes.
17997 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17998 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17999 (Insns) <jump_table_data>: New entry.
18000 * doc/tm.texi: Regenerate.
18002 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
18004 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
18005 for table jump at the end of a basic block using tablejump_p.
18006 * targhooks.c (default_invalid_within_doloop): Likewise.
18007 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
18008 target hook implementation that is identical to the default hook.
18009 (rs6000_invalid_within_doloop): Remove.
18011 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
18012 unused variable from tablejump_p call.
18014 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18015 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18016 (INSN_DELETED_P): Likewise.
18017 (emit_jump_table_data): New prototype.
18018 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18019 after 4th as unused.
18020 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18021 * sched-vis.c (print_insn): Likewise.
18022 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18023 insn for compatibility with back ends that use next_active_insn to
18024 identify jump table data.
18025 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18026 (remove_insn): Likewise.
18027 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18029 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18030 (emit_jump_table_data): New function.
18032 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18033 basic block, a JUMP_TABLE_DATA never is.
18034 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18035 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18036 off from code handling real insns.
18037 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18038 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18039 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18040 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18041 is not a NONDEBUG_INSN_P.
18042 * ira-costs.c (scan_one_insn): Likewise.
18043 * jump.c (mark_all_labels): Likewise.
18044 (mark_jump_label_1): Likewise.
18045 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18046 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18048 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18049 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18050 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18051 JUMP_TABLE_DATA_P insns.
18052 (calculate_elim_costs_all_insns): Likewise.
18053 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18054 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18055 (delete_output_reload): Code style fixups.
18056 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18057 insn flags on this non-insn.
18058 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18059 as scheduling barriers, for pre-change compatibility.
18060 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18061 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18063 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18064 redundant JUMP_TABLE_DATA_P test.
18065 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18066 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18067 (frv_for_each_packet): Likewise.
18068 * config/i386/i386.c (min_insn_size): Likewise.
18069 (ix86_avoid_jump_mispredicts): Likewise.
18070 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18071 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18072 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18073 (mips16_insn_length): Robustify.
18074 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18075 (mips16_split_long_branches): Likewise.
18076 * config/pa/pa.c (pa_combine_instructions): Likewise.
18077 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18078 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18079 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18080 as contributing to pool range lengths.
18081 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18082 Remove redundant JUMP_TABLE_DATA_P test.
18083 (sh_loop_align): Likewise.
18084 (split_branches): Likewise.
18085 (sh_insn_length_adjustment): Likewise.
18086 * config/spu/spu.c (get_branch_target): Likewise.
18088 2013-03-29 Jan Hubicka <jh@suse.cz>
18090 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18091 gcov streaming; stream hot bb threshold to ltrans.
18092 * predict.c (get_hot_bb_threshold): Break out from ....
18093 (maybe_hot_count_p): ... here.
18094 (set_hot_bb_threshold): New function.
18095 * lto-section-in.c (lto_section_name): Add profile.
18096 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18097 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18098 and data-streamer.h
18099 (histogram_entry): New structure.
18100 (histogram, histogram_pool): New global vars.
18101 (histogram_hash): New structure.
18102 (histogram_hash::hash): New method.
18103 (histogram_hash::equal): Likewise.
18104 (account_time_size): New function.
18105 (cmp_counts): New function.
18106 (dump_histogram): New function.
18107 (ipa_profile_generate_summary): New function.
18108 (ipa_profile_write_summary): New function.
18109 (ipa_profile_read_summary): New function.
18110 (ipa_profile): Decide on threshold.
18111 (pass_ipa_profile): Add ipa_profile_write_summary and
18112 ipa_profile_read_summary.
18113 * Makefile.in (ipa.o): Update dependencies.
18114 * lto-streamer.h (LTO_section_ipa_profile): New section.
18116 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18118 * tree.h (VAR_P): New.
18120 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18123 * doc/invoke.texi ([-fwhole-program]): Fix typo.
18125 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
18127 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
18128 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18129 (control_flow_insn_p): Likewise.
18130 * cfgrtl.c (duplicate_insn_chain): Likewise.
18131 * final.c (get_attr_length_1): Likewise.
18132 (shorten_branches): Likewise.
18133 (final_scan_insn): Likewise.
18134 * function.c (instantiate_virtual_regs): Likewise.
18135 * gcse.c (insert_insn_end_basic_block): Likewise.
18136 * ira-costs.c (scan_one_insn): Likewise.
18137 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18138 * lra.c (check_rtl): Likewise.
18139 * reload1.c (elimination_costs_in_insn): Likewise.
18140 * reorg.c (follow_jumps): Likewise.
18142 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
18143 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18144 (thumb_far_jump_used_p): Likewise.
18145 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
18146 (workaround_speculation): Likewise.
18147 (add_sched_insns_for_speculation): Likewise.
18148 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
18149 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18150 (frv_for_each_packet): Likewise.
18151 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
18152 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
18153 (final_emit_insn_group_barriers): Likewise.
18154 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18155 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18156 (mips16_insn_length): Likewise.
18157 * config/pa/pa.c (pa_reorg): Likewise.
18158 (pa_combine_instructions): Likewise.
18159 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
18160 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
18161 (sh_reorg): Likewise.
18162 (split_branches): Likewise.
18163 * config/spu/spu.c (get_branch_target): Likewise.
18165 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
18168 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
18170 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
18171 Fix declaration name.
18173 2013-03-28 Lawrence Crowl <crowl@google.com>
18175 * graphds.h (struct graph.indicies): Remove unused.
18176 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
18177 (SCOP_ORIGINAL_PDDRS): Remove unused.
18178 * sese.h (extern insert_loop_close_phis): Removed unused.
18179 (extern insert_guard_phis): Removed unused.
18180 (extern ivtype_map_elt_info): Removed unused.
18181 (new_ivtype_map_elt): Removed unused.
18182 * sese.c (ivtype_map_elt_info): Removed unused.
18184 2013-03-28 Lawrence Crowl <crowl@google.com>
18186 * Makefile.in: Add several missing include dependences.
18188 (test-dump.o): New. This object is not added to any executable,
18189 but is present for ad-hoc testing.
18191 (debug (const bitmap_head_def &)): New.
18192 (debug (const bitmap_head_def *)): New.
18194 (extern debug (const bitmap_head_def &)): New.
18195 (extern debug (const bitmap_head_def *)): New.
18197 (debug (edge_def &)): New.
18198 (debug (edge_def *)): New.
18200 (debug (basic_block_def &)): New.
18201 (debug (basic_block_def *)): New.
18203 (dump_node (const_tree, int, FILE *)): Correct source file.
18205 (debug (die_struct &)): New.
18206 (debug (die_struct *)): New.
18208 (extern debug (die_struct &)): New.
18209 (extern debug (die_struct *)): New.
18210 * gimple-pretty-print.c
18211 (debug (gimple_statement_d &)): New.
18212 (debug (gimple_statement_d *)): New.
18213 * gimple-pretty-print.h
18214 (extern debug (gimple_statement_d &)): New.
18215 (extern debug (gimple_statement_d *)): New.
18217 (debug (ira_allocno_copy &)): New.
18218 (debug (ira_allocno_copy *)): New.
18219 (debug (ira_allocno &)): New.
18220 (debug (ira_allocno *)): New.
18222 (extern debug (ira_allocno_copy &)): New.
18223 (extern debug (ira_allocno_copy *)): New.
18224 (extern debug (ira_allocno &)): New.
18225 (extern debug (ira_allocno *)): New.
18227 (debug (live_range &)): New.
18228 (debug (live_range *)): New.
18230 (debug (lra_live_range &)): New.
18231 (debug (lra_live_range *)): New.
18233 (debug (lra_live_range &)): New.
18234 (debug (lra_live_range *)): New.
18236 (debug (omega_pb_d &)): New.
18237 (debug (omega_pb_d *)): New.
18239 (extern debug (omega_pb_d &)): New.
18240 (extern debug (omega_pb_d *)): New.
18242 (debug (const rtx_def &)): New.
18243 (debug (const rtx_def *)): New.
18245 (debug_tree (tree): Move within file.
18246 (debug_raw (const tree_node &)): New.
18247 (debug_raw (const tree_node *)): New.
18248 (dump_tree_via_hooks (const tree_node *, int)): New.
18249 (debug (const tree_node &)): New.
18250 (debug (const tree_node *)): New.
18251 (debug_verbose (const tree_node &)): New.
18252 (debug_verbose (const tree_node *)): New.
18253 (debug_head (const tree_node &)): New.
18254 (debug_head (const tree_node *)): New.
18255 (debug_body (const tree_node &)): New.
18256 (debug_body (const tree_node *)): New.
18257 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18258 (debug (vec<tree, va_gc> &)): New.
18259 (debug (vec<tree, va_gc> *)): New.
18261 (extern debug (const rtx_def &)): New.
18262 (extern debug (const rtx_def *)): New.
18264 (debug_raw (simple_bitmap_def &)): New.
18265 (debug_raw (simple_bitmap_def *)): New.
18266 (debug (simple_bitmap_def &)): New.
18267 (debug (simple_bitmap_def *)): New.
18269 (extern debug (simple_bitmap_def &)): New.
18270 (extern debug (simple_bitmap_def *)): New.
18271 (extern debug_raw (simple_bitmap_def &)): New.
18272 (extern debug_raw (simple_bitmap_def *)): New.
18274 (debug (vinsn_def &)): New.
18275 (debug (vinsn_def *)): New.
18276 (debug_verbose (vinsn_def &)): New.
18277 (debug_verbose (vinsn_def *)): New.
18278 (debug (expr_def &)): New.
18279 (debug (expr_def *)): New.
18280 (debug_verbose (expr_def &)): New.
18281 (debug_verbose (expr_def *)): New.
18282 (debug (vec<rtx> &)): New.
18283 (debug (vec<rtx> *)): New.
18285 (extern debug (vinsn_def &)): New.
18286 (extern debug (vinsn_def *)): New.
18287 (extern debug_verbose (vinsn_def &)): New.
18288 (extern debug_verbose (vinsn_def *)): New.
18289 (extern debug (expr_def &)): New.
18290 (extern debug (expr_def *)): New.
18291 (extern debug_verbose (expr_def &)): New.
18292 (extern debug_verbose (expr_def *)): New.
18293 (extern debug (vec<rtx> &)): New.
18294 (extern debug (vec<rtx> *)): New.
18296 (_list_iter_cond_expr): Make inline instead of static.
18298 (debug (sreal &)): New.
18299 (debug (sreal *)): New.
18301 (extern debug (sreal &)): New.
18302 (extern debug (sreal *)): New.
18304 (extern debug_raw (const tree_node &)): New.
18305 (extern debug_raw (const tree_node *)): New.
18306 (extern debug (const tree_node &)): New.
18307 (extern debug (const tree_node *)): New.
18308 (extern debug_verbose (const tree_node &)): New.
18309 (extern debug_verbose (const tree_node *)): New.
18310 (extern debug_head (const tree_node &)): New.
18311 (extern debug_head (const tree_node *)): New.
18312 (extern debug_body (const tree_node &)): New.
18313 (extern debug_body (const tree_node *)): New.
18314 (extern debug (vec<tree, va_gc> &)): New.
18315 (extern debug (vec<tree, va_gc> *)): New.
18317 (debug (struct loop &)): New.
18318 (debug (struct loop *)): New.
18319 (debug_verbose (struct loop &)): New.
18320 (debug_verbose (struct loop *)): New.
18321 * tree-dump.c: Add header dependence.
18323 (extern debug (struct loop &)): New.
18324 (extern debug (struct loop *)): New.
18325 (extern debug_verbose (struct loop &)): New.
18326 (extern debug_verbose (struct loop *)): New.
18328 (debug (data_reference &)): New.
18329 (debug (data_reference *)): New.
18330 (debug (vec<data_reference_p> &)): New.
18331 (debug (vec<data_reference_p> *)): New.
18332 (debug (vec<ddr_p> &)): New.
18333 (debug (vec<ddr_p> *)): New.
18335 (extern debug (data_reference &)): New.
18336 (extern debug (data_reference *)): New.
18337 (extern debug (vec<data_reference_p> &)): New.
18338 (extern debug (vec<data_reference_p> *)): New.
18339 (extern debug (vec<ddr_p> &)): New.
18340 (extern debug (vec<ddr_p> *)): New.
18342 (debug (pt_solution &)): New.
18343 (debug (pt_solution *)): New.
18345 (extern debug (pt_solution &)): New.
18346 (extern debug (pt_solution *)): New.
18348 (debug (_var_map &)): New.
18349 (debug (_var_map *)): New.
18350 (debug (tree_live_info_d &)): New.
18351 (debug (tree_live_info_d *)): New.
18353 (extern debug (_var_map &)): New.
18354 (extern debug (_var_map *)): New.
18355 (extern debug (tree_live_info_d &)): New.
18356 (extern debug (tree_live_info_d *)): New.
18358 2013-03-28 Jan Hubicka <jh@suse.cz>
18360 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
18362 2013-03-28 Ian Bolton <ian.bolton@arm.com>
18364 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
18365 record only when desired or required.
18367 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
18369 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
18370 *vec_extractv2di_1_rex64. Use x64 isa attribute.
18372 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18374 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
18375 (*andsi3_compare0_uxtw): New pattern.
18376 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
18377 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
18379 2013-03-28 Jan Hubicka <jh@suse.cz>
18381 * data-streamer-in.c (streamer_read_gcov_count): New function.
18382 * gimple-streamer-out.c: Include value-prof.h.
18383 (output_gimple_stmt): Output histogram.
18384 (output_bb): Use streamer_write_gcov_count.
18385 * value-prof.c: Include data-streamer.h
18386 (dump_histogram_value): Add HIST_TYPE_MAX.
18387 (stream_out_histogram_value): New function.
18388 (stream_in_histogram_value): New function.
18389 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
18390 (stream_out_histogram_value, stream_in_histogram_value): Declare.
18391 * data-streamer-out.c (streamer_write_gcov_count): New function.
18392 (streamer_write_gcov_count_stream): New function.
18393 * lto-cgraph.c (lto_output_edge): Update counter streaming.
18394 (lto_output_node): Likewise.
18395 (input_node, input_edge): Likewise.
18396 * lto-streamer-out.c (output_cfg): Update streaming.
18397 * lto-streamer-in.c (input_cfg): Likewise.
18398 * data-streamer.h (streamer_write_gcov_count,
18399 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
18400 * gimple-streamer-in.c: Include value-prof.h
18401 (input_gimple_stmt): Input histograms.
18402 (input_bb): Update profile streaming.
18404 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18406 * genmodes.c (emit_max_int): New function.
18407 (emit_insn_modes_h): Added call to emit_max_function.
18408 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
18410 * machmode.def: Fixed comment.
18412 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18414 * combine.c (try_combine): Removed useless assert.
18415 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
18417 2013-03-28 Marek Polacek <polacek@redhat.com>
18418 Richard Biener <rguenther@suse.de>
18420 PR tree-optimization/56695
18421 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
18422 build signed result of a vector comparison.
18423 * tree-cfg.c (verify_gimple_comparison): Check that a result
18424 of a vector comparison has signed type.
18426 2013-03-28 Richard Biener <rguenther@suse.de>
18428 PR tree-optimization/37021
18429 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
18430 do not restrict gaps between groups.
18431 * tree-vect-stmts.c (vectorizable_load): Properly account for
18432 a gap between groups.
18434 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
18436 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
18437 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
18440 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
18442 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
18443 * doc/extend.texi (Named Address Spaces): Ditto.
18444 (Variable Attributes): Ditto.
18446 2013-03-27 Kai Tietz <ktietz@redhat.com>
18448 * config.build: Add support for cygwin x64 target.
18449 * config.gcc: Likewise.
18450 * config.host: Likewise.
18451 * configure.ac: Likewise
18452 * configure: Regenerated.
18454 2013-03-27 Kai Tietz <ktietz@redhat.com>
18456 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
18457 * config/i386/t-cygwin-w64: New file.
18458 * config/i386/cygwin-w64.h: New file.
18459 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
18460 and add support for x64-cygwin target.
18461 (CPP_SPEC): Likewise.
18462 (CXX_WRAP_SPEC_LIST): Undefine before define.
18463 (LIBGCJ_SONAME): Use 15 as version.
18465 2013-03-27 Richard Biener <rguenther@suse.de>
18467 PR tree-optimization/56716
18468 * tree-ssa-structalias.c (perform_var_substitution): Adjust
18469 dumping for ref nodes.
18471 2013-03-27 Martin Jambor <mjambor@suse.cz>
18473 PR tree-optimization/55334
18474 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
18475 restricted pointers to arrays.
18477 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
18479 * Makefile.in (.SUFFIXES): Add .cc.
18480 (.c.o): Apply same recipe for implicit rule .cc.o.
18482 2013-03-27 Richard Biener <rguenther@suse.de>
18484 PR tree-optimization/37021
18485 * tree-vect-data-refs.c (vect_check_strided_load): Allow
18486 REALPART/IMAGPART_EXPRs around the supported refs.
18487 * tree-ssa-structalias.c (find_func_aliases): Assume that
18488 floating-point values are not used to transfer pointers.
18490 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
18492 * target.def (TARGET_HAS_IFUNC_P): New target hook.
18493 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
18494 * doc/tm.texi: Regenerate.
18495 * targhooks.h (default_has_ifunc_p): New.
18496 * targhooks.c (default_has_ifunc_p): Ditto.
18497 * config/linux-protos.h: New file.
18498 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
18499 hook for linux which disables support of indirect functions in android.
18500 * config/linux-android.c: New file.
18501 * config/t-linux-android.c: Ditto.
18502 * config.gcc: Added new object file linux-android.o.
18503 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
18504 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
18505 * varasm.c (do_assemble_alias): Likewise.
18506 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
18507 doesn't support indirect functions.
18508 * configure: Regenerate.
18510 2013-03-27 Bin Cheng <bin.cheng@arm.com>
18513 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
18514 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
18517 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18519 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
18521 2013-03-27 Terry Guo <terry.guo@arm.com>
18523 * config/arm/arm-cores.def: Added core cortex-r7.
18524 * config/arm/arm-tune.md: Regenerated.
18525 * config/arm/arm-tables.opt: Regenerated.
18526 * doc/invoke.texi: Added entry for core cortex-r7.
18528 2013-03-27 Walter Lee <walt@tilera.com>
18530 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
18531 double-decrement of next_scratch_regno.
18533 2013-03-27 Walter Lee <walt@tilera.com>
18535 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
18537 (insn_v1mulus): Ditto.
18538 (insn_v2muls): Ditto.
18540 2013-03-27 Walter Lee <walt@tilera.com>
18542 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18543 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18545 2013-03-27 Walter Lee <walt@tilera.com>
18547 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18548 (*sibcall_value): Ditto.
18550 2013-03-27 Walter Lee <walt@tilera.com>
18552 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18553 (insn_mnz_v8qi): ... this ...
18554 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18556 (insn_v<n>mnz): Replaced by ...
18557 (insn_v1mnz): ... this ...
18558 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18560 (insn_mz_<mode>): Replaced by ...
18561 (insn_mz_v8qi): ... this ...
18562 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18564 (insn_v<n>mz): Replaced by ...
18565 (insn_v1mz): ... this ...
18566 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18569 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18571 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18573 2013-03-26 Roland McGrath <mcgrathr@google.com>
18575 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18576 than fprintf with a non-constant, non-format string.
18578 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18580 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18581 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18582 operand 0 predicate.
18583 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18584 attribute. Use general_x64nomem_operand as operand 1 predicate.
18585 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18586 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18587 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18588 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18589 (mov<mode>_insv_1): Remove expander. Merge insn with
18590 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18591 Use general_x64nomem_operand as operand 1 predicate.
18592 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18593 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18594 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18595 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18596 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18597 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18598 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18599 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18600 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18601 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18602 (general_x64nomem_operand): Ditto.
18604 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18606 * config/rtems.opt: Add -pthread option.
18608 2013-03-26 Richard Biener <rguenther@suse.de>
18610 * alias.c (find_base_term): Avoid redundant and not used recursion.
18611 (base_alias_check): Get the initial base term from the caller.
18612 (true_dependence_1): Compute and pass base terms to base_alias_check.
18613 (write_dependence_p): Likewise.
18614 (may_alias_p): Likewise.
18616 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18618 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18619 PC-relative load in SI modes and above only.
18621 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18623 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18624 * config/arm/iwmmxt.md (WCGR0): Update.
18625 (WCGR1, WCGR2, WCGR3): Likewise.
18627 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18629 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18630 Use x64 and nox64 isa attributes.
18632 2013-03-26 Richard Biener <rguenther@suse.de>
18634 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18635 alignment computations and rely on get_object_alignment_1
18636 for the !TYPE_P case.
18637 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18639 2013-03-26 Walter Lee <walt@tilera.com>
18641 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18642 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18644 2013-03-25 Jeff Law <law@redhat.com>
18646 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18647 check for INTEGRAL_TYPE_P that was missing due to checking in
18648 wrong version of prior patch.
18650 2013-03-25 Walter Lee <walt@tilera.com>
18652 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18653 TILEGX_INSN_SHUFFLEBYTES1.
18654 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18656 (tilegx_builtins): Ditto.
18657 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18659 2013-03-25 Walter Lee <walt@tilera.com>
18661 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18662 (floatunssisf2): New pattern.
18663 (floatsidf2): New pattern.
18664 (floatunssidf2): New pattern.
18666 2013-03-25 Walter Lee <walt@tilera.com>
18668 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18669 tests for constraint J, K, N, P.
18671 2013-03-25 Walter Lee <walt@tilera.com>
18673 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18674 Use indirect/pcrel encoding.
18675 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18678 2013-03-25 Steve Ellcey <sellcey@mips.com>
18680 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18681 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18682 * config/mips/mips.c (mips_option_override): Set IMADD default.
18683 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18684 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18685 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18686 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18687 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18689 2013-03-25 Jeff Law <law@redhat.com>
18691 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18692 slightly to avoid creating and folding useless trees. Simplify
18693 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18695 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18697 * config/i386/i386.md (*zero_extendsidi2): Merge with
18698 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18699 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18700 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18701 targets. Clarify comment.
18703 2013-03-25 Martin Jambor <mjambor@suse.cz>
18705 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18706 pass-through jump functions differently.
18707 (ipa_read_jump_function): Likewise. Also use setter functions to set
18710 2013-03-25 Martin Jambor <mjambor@suse.cz>
18712 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18713 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18715 (ipa_get_indirect_edge_target): New function.
18716 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18717 ipa_get_indirect_edge_target. Update all callers.
18718 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18719 ipa_get_indirect_edge_target_1 instead of calling
18720 ipa_get_indirect_edge_target.
18721 (create_specialized_node): Pass aggvlas to
18722 ipcp_discover_new_direct_edges.
18724 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18726 * config/arm/arm.md (f_sels, f_seld): New types.
18727 (*cmov<mode>): New pattern.
18728 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18731 2013-03-25 Kai Tietz <ktietz@redhat.com>
18733 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18734 POSIX-printf for mingw-hosted builds.
18736 2013-03-25 Richard Biener <rguenther@suse.de>
18738 PR middle-end/56694
18739 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18740 must-not-throw stmt location.
18742 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18744 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18745 Emit load-acquire versions when acq is true.
18746 (arm_emit_store_exclusive): Add rel parameter.
18747 Emit store-release versions when rel is true.
18748 (arm_split_compare_and_swap): Use acquire-release instructions
18750 of barriers when appropriate.
18751 (arm_split_atomic_op): Likewise.
18752 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18753 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18754 (VUNSPEC_SLX): Likewise.
18755 (VUNSPEC_LDA): Likewise.
18756 (VUNSPEC_STL): Likewise.
18757 * config/arm/sync.md (atomic_load<mode>): New pattern.
18758 (atomic_store<mode>): Likewise.
18759 (arm_load_acquire_exclusive<mode>): Likewise.
18760 (arm_load_acquire_exclusivesi): Likewise.
18761 (arm_load_acquire_exclusivedi): Likewise.
18762 (arm_store_release_exclusive<mode>): Likewise.
18764 2013-03-25 Catherine Moore <clm@codesourcery.com>
18766 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18767 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18768 * config/mip/predicates.md (lwsp_swsp_operand,
18769 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18770 sb16_operand, db4_operand, db7_operand, ib3_operand,
18771 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18772 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18773 andi16_operand): New predicates.
18774 * config/mips/mips.md (compression): New attribute.
18775 (enabled): New attribute.
18776 (length): Consider compression in computing length.
18777 (shift_compression): New code attribute.
18778 (*add<mode>3): New operands. Record compression.
18779 (sub<mode>3): Likewise.
18780 (one_cmpl<mode>2): Likewise.
18781 (*and<mode>3): Likewise.
18782 (*ior<mode>3): Likewise.
18783 (unnamed pattern for xor): Likewise.
18784 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18785 (*<optab><mode>3): Likewise.
18786 (*mov<mode>_internal: Likewise.
18787 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18788 (mips_unsigned_immediate_p): New.
18789 (umips_lwsp_swsp_address_p): New.
18790 (m16_based_address_p): New.
18791 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18792 (mips_unsigned_immediate_p): New prototype.
18793 (lwsp_swsp_address_p): New prototype.
18794 (m16_based_address_p): New prototype.
18795 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18796 (mips_signed_immediate_p): New function.
18797 (m16_based_address_p): New function.
18798 (lwsp_swsp_address_p): New function.
18799 (mips_print_operand_punctuation): Recognize short delay slot insns
18800 for microMIPS.add<mode>3"
18802 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18805 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18806 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18808 2013-03-25 Richard Biener <rguenther@suse.de>
18810 PR tree-optimization/56689
18811 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18814 2013-03-25 Richard Biener <rguenther@suse.de>
18816 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18818 (memory_references): Likewise.
18819 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18820 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18821 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18822 (gather_mem_refs_in_loops): Fold into ...
18823 (analyze_memory_references): ... this. Move initialization
18824 to tree_ssa_lim_initialize.
18825 (fill_always_executed_in): Rename to ...
18826 (fill_always_executed_in_1): ... this.
18827 (fill_always_executed_in): Move contains_call computation to
18828 this new function from ...
18829 (tree_ssa_lim_initialize): ... here.
18830 (tree_ssa_lim): Call fill_always_executed_in.
18832 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18834 * postreload.c (reload_combine): Fix code detecting returns.
18836 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18838 * function.c (emit_use_return_register_into_block): On cc0 targets,
18839 do not emit the sequence between cc0 setter and user.
18841 2013-03-25 Kai Tietz <ktietz@redhat.com>
18843 * config/i386/predicates.md (local_symbolic_operand): Interpret
18844 dll-imported symbols as none-local.
18846 2013-03-25 Richard Biener <rguenther@suse.de>
18848 * tree-ssa-loop-im.c (struct depend): Remove.
18849 (struct lim_aux_data): Make depends a vec of gimples.
18850 (free_lim_aux_data): Adjust.
18851 (add_dependency): Likewise.
18852 (set_level): Likewise.
18854 2013-03-25 Richard Biener <rguenther@suse.de>
18856 PR middle-end/56434
18857 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18858 the pointer returned by calls with ECF_MALLOC set.
18860 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18862 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18864 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18866 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18867 using MMXMODE mode iterator.
18868 (*move<mode>_internal): Merge with *movv2sf_internal and
18869 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18871 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18873 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18874 (record_last_mem_set_info): Likewise.
18876 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18877 of XNEWVEC followed by memset.
18878 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18880 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18882 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18883 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18884 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18885 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18886 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18887 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18888 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18889 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18890 BARRIER_P instead of GET_CODE.
18892 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18894 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18895 inaccuracy in the probing code.
18897 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18898 (ctrapdi4): Likewise.
18900 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18902 * calls.c (expand_call): Add missing guard to code handling return
18903 of non-BLKmode structures in MSB.
18904 * function.c (expand_function_end): Likewise.
18906 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18908 * combine.c (try_combine): Adjust comment. Do not add the set of
18909 insn #0 if the destination indirectly is set or dies in insn #2.
18910 Tidy up code to distribute a new note.
18912 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18914 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18915 also for alternatives 16 and 17.
18917 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18919 * config/i386/sse.md (*mov<mode>_internal): Merge with
18920 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18921 Emit insn template depending on type attribute. Use
18922 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18923 movd instead of movq mnemonic for interunit moves. Rewrite mode
18924 attribute calculation. Remove unit attribute calculation.
18925 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18926 Set prefix_data16 attribute for DImode ssemov types.
18927 Use Ym instead of y for SSE-MMX conversion alternatives.
18928 Reorder operand constraints.
18930 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18932 * df.h (df_insn_delete): Adjust prototype.
18933 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18934 and let it decide whether mark the basic block dirty.
18935 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18936 * df-scan.c (df_insn_info_delete): New helper function, split
18937 off from df_insn_delete.
18938 (df_scan_free_bb_info): Use it.
18939 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18941 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18942 that the insn is actually an insn and it has a non-NULL basic block.
18943 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18945 2013-03-22 Richard Biener <rguenther@suse.de>
18947 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18949 (mem_ref_alloc): Do not allocate them.
18950 (refs_independent_p): Do not query or maintain a cache.
18952 2013-03-22 Richard Biener <rguenther@suse.de>
18954 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18955 (gather_mem_refs_in_loops): Do not compute it.
18956 (analyze_memory_references): Do not allocate it.
18957 (tree_ssa_lim_finalize): Do not free it.
18958 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18960 2013-03-22 Richard Biener <rguenther@suse.de>
18962 * is-a.h (as_a): Use gcc_checking_assert.
18964 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18966 * config/aarch64/aarch64.c (aarch64_print_operand): New
18967 format specifier for printing a constant in hex.
18968 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18969 format specifier for printing second operand.
18971 2013-03-22 Richard Biener <rguenther@suse.de>
18973 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18975 (gather_mem_refs_in_loops): Perform store accumulation here.
18976 (create_vop_ref_mapping_loop): Remove.
18977 (create_vop_ref_mapping): Likewise.
18978 (analyze_memory_references): Initialize refs_stored_in_loop.
18979 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18980 (record_indep_loop): Remove.
18981 (record_dep_loop): New function.
18982 (ref_indep_loop_p_1): Adjust to only walk over references
18983 in the loop, not its subloops.
18984 (ref_indep_loop_p): Rename to ...
18985 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18986 maintaining a more fine-grained cache.
18987 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18988 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18990 2013-03-22 Richard Biener <rguenther@suse.de>
18992 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18993 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18994 aggregate mem_ref_loc.
18995 (free_mem_ref_locs): Inline into ...
18996 (memref_free): ... this and adjust.
18997 (mem_ref_alloc): Adjust.
18998 (mem_ref_locs_alloc): Remove.
18999 (record_mem_ref_loc): Adjust.
19000 (get_all_locs_in_loop): Rewrite into ...
19001 (for_all_locs_in_loop): ... this iterator.
19002 (rewrite_mem_ref_loc): New functor.
19003 (rewrite_mem_refs): Use for_all_locs_in_loop.
19004 (sm_set_flag_if_changed): New functor.
19005 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
19006 (ref_always_accessed): New functor.
19007 (ref_always_accessed_p): Use for_all_locs_in_loop.
19009 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19011 * tree-pass.h (PROP_gimple_lvec): New.
19012 * passes.c (dump_properties): Handle PROP_gimple_lvec.
19013 (init_optimization_passes): Move pass_lower_vector.
19014 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19016 (pass_lower_vector): Provide PROP_gimple_lvec.
19017 (pass_lower_vector_ssa): Likewise.
19018 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19020 2013-03-21 Mark Wielaard <mjw@redhat.com>
19022 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19024 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19026 * config/i386/i386.md (*movdi_internal): Disparage slightly
19027 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19028 conversion alternatives.
19030 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19032 PR middle-end/48087
19033 * diagnostic.def (DK_WERROR): New kind.
19034 * diagnostic.h (werrorcount): Define.
19035 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19036 promoted to DK_ERROR, increment DK_WERROR counter instead of
19038 * toplev.c (toplev_main): Call print_ignored_options even if
19039 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19040 even if just werrorcount is non-zero.
19043 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19045 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19046 (string_cst_pool_decl): New function.
19047 (optimize_one_addr_into_implicit_ptr): New function.
19048 (resolve_addr_in_expr): Optimize DWARF location expression
19049 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19050 which doesn't live in memory, but has DW_AT_location or
19051 DW_AT_const_value, or refers to a string literal, into
19052 DW_OP_GNU_implicit_pointer.
19053 (optimize_location_into_implicit_ptr): New function.
19054 (resolve_addr): If removing DW_AT_location of a variable because
19055 it was DW_OP_addr of address of the variable, but the variable doesn't
19056 live in memory, try to emit const value attribute for the initializer.
19058 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19060 * tree.h (VECTOR_TYPE_P): New macro.
19061 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19062 TYPE_MODE): Use it.
19063 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19064 VEC_COND_EXPR cannot be lvalues.
19065 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19067 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19069 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19070 Restrict the transformation to equal modes.
19072 2013-03-21 Richard Biener <rguenther@suse.de>
19074 PR tree-optimization/39326
19075 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19076 (MEM_ANALYZABLE): Adjust.
19077 (record_mem_ref_loc): Move bitmap ops ...
19078 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19079 unanalyzable refs, do not record locations for it.
19080 (analyze_memory_references): Allocate ref zero as shared
19082 (refs_independent_p): Do not test for unanalyzed mems here.
19083 (ref_indep_loop_p_1): Special-case disambiguation against
19084 the unanalyzed ref.
19085 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19087 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19089 * config/arm/arm-protos.h (tune_params): Add
19090 prefer_neon_for_64bits field.
19091 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19092 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19093 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19094 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19095 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19096 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19097 (arm_option_override): Handle -mneon-for-64bits new option.
19098 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19099 (prefer_neon_for_64bits): Declare new variable.
19100 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19101 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19102 (arch_enabled): Handle new arch types. Remove support for onlya8
19104 (one_cmpldi2): Use new arch names.
19105 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19106 * config/arm/arm.opt (mneon-for-64bits): Add option.
19107 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19108 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19109 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19111 * doc/invoke.texi (-mneon-for-64bits): Document.
19113 2013-03-21 Richard Biener <rguenther@suse.de>
19115 PR tree-optimization/39326
19116 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19117 (sort_bbs_in_loop_postorder_cmp): New function.
19118 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19120 2013-03-21 Richard Biener <rguenther@suse.de>
19122 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19123 (vect_insert_into_interleaving_chain): Likewise.
19124 (vect_drs_dependent_in_basic_block): Inline ...
19125 (vect_slp_analyze_data_ref_dependence): ... here. New function,
19127 (vect_analyze_data_ref_dependence): ... here. Simplify.
19128 (vect_check_interleaving): Simplify.
19129 (vect_analyze_data_ref_dependences): Likewise. Split out ...
19130 (vect_slp_analyze_data_ref_dependences): ... this new function.
19131 (dr_group_sort_cmp): New function.
19132 (vect_analyze_data_ref_accesses): Compute data-reference groups
19133 here instead of in vect_analyze_data_ref_dependence. Use
19134 a more efficient algorithm.
19135 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
19136 vect_slp_analyze_data_ref_dependences. Call
19137 vect_analyze_data_ref_accesses earlier.
19138 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
19139 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
19140 (vect_slp_analyze_data_ref_dependences): New prototype.
19142 2013-03-21 Richard Biener <rguenther@suse.de>
19144 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
19145 ref is stored in the loop.
19146 (find_refs_for_sm): Walk only over all stores.
19147 (store_motion_loop): Allocate from lim_bitmap_obstack.
19148 (store_motion): Likewise.
19150 2013-03-21 Richard Biener <rguenther@suse.de>
19152 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
19153 Update virtual SSA form.
19155 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19157 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
19158 * configure: Regenerate.
19159 * config.in: Regenerate.
19160 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
19161 if !HAVE_LD_EH_FRAME_CIEV3.
19163 2013-03-21 Richard Biener <rguenther@suse.de>
19165 * tree-cfg.c (verify_expr_no_block): New function.
19166 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
19167 nor DECL_VALUE_EXPR have locations with associated blocks.
19168 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
19169 (clear_unused_block_pointer): Remove code dealing with
19170 blocks in DECL_DEBUG_EXPR locations.
19172 2013-03-21 Richard Biener <rguenther@suse.de>
19174 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
19175 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
19176 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
19177 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
19178 instead of DECL_DEBUG_EXPR_IS_FROM.
19179 * gimplify.c (gimplify_modify_expr): Likewise.
19180 * tree-cfg.c (verify_expr_location_1): Likewise.
19181 * tree-complex.c (create_one_component_var): Likewise.
19182 * tree-sra.c (create_access_replacement): Likewise.
19183 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
19184 (clear_unused_block_pointer): Likewise.
19185 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
19186 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
19187 * var-tracking.c (var_debug_decl): Likewise.
19188 (track_expr_p): Likewise.
19189 * tree-inline.c (add_local_variables): Likewise. Set
19190 DECL_HAS_DEBUG_EXPR_P after copying it.
19191 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
19192 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
19194 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19197 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
19198 * configure: Regenerate.
19199 * config.in: Regenerate.
19200 * config/i386/i386.md (*movdf_internal): Use
19201 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19202 movd instead of movq mnemonic for interunit moves.
19203 (*movdi_internal): Ditto.
19205 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19207 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
19208 (abd<mode>_3): New pattern.
19209 (aba<mode>_3): New pattern.
19210 (fabd<mode>_3): New pattern.
19212 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19214 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
19215 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
19216 occurrence of REGISTER_PREFIX as its empty string.
19218 2013-03-20 Jeff Law <law@redhat.com>
19220 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
19221 addititional equivalences for equality comparisons between an SSA_NAME
19222 and a constant where the SSA_NAME was set from a widening conversion.
19224 2013-03-20 Walter Lee <walt@tilera.com>
19226 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
19228 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19230 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
19231 depending on type attribute.
19232 (*movti_internal): Ditto.
19233 (*movtf_internal): Ditto.
19234 (*movxf_internal): Ditto.
19235 (*movdf_internal): Ditto.
19236 (*movsf_internal): Ditto.
19238 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19240 * config/i386/i386.md (*movti_internal): Set prefix attribute to
19241 maybe_vex for sselog1 and ssemov types.
19242 (*movdi_internal): Reorder operand constraints.
19243 (*movsi_internal): Ditto. Set prefix attribute to
19244 maybe_vex for sselog1 and ssemov types.
19245 (*movtf_internal): Set prefix attribute to maybe_vex
19246 for sselog1 and ssemov types.
19247 (*movdf_internal): Ditto. Set prefix_data16 attribute for
19248 DImode ssemov types. Reorder operand constraints.
19249 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19250 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19251 attribute for SImode ssemov types. Reorder operand constraints.
19253 2013-03-20 Martin Jambor <mjambor@suse.cz>
19255 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19256 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19258 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19260 * config/rs6000/predicates.md (indexed_address, update_address_mem
19261 update_indexed_address_mem): New predicates.
19262 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19263 attribute for load/store instructions.
19264 * config/rs6000/dfp.md (movsd_store): Likewise.
19265 (movsd_load): Likewise.
19266 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19267 (unnamed HI->DI extend define_insn): Likewise.
19268 (unnamed SI->DI extend define_insn): Likewise.
19269 (unnamed QI->SI extend define_insn): Likewise.
19270 (unnamed QI->HI extend define_insn): Likewise.
19271 (unnamed HI->SI extend define_insn): Likewise.
19272 (unnamed HI->SI extend define_insn): Likewise.
19273 (extendsfdf2_fpr): Likewise.
19274 (movsi_internal1): Likewise.
19275 (movsi_internal1_single): Likewise.
19276 (movhi_internal): Likewise.
19277 (movqi_internal): Likewise.
19278 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
19279 attribute for load/store instructions.
19280 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
19282 (mov<mode>_softfloat): Likewise.
19283 (mov<mode>_hardfloat32): Likewise.
19284 (mov<mode>_hardfloat64): Likewise.
19285 (mov<mode>_softfloat64): Likewise.
19286 (movdi_internal32): Likewise.
19287 (movdi_internal64): Likewise.
19288 (probe_stack_<mode>): Likewise.
19290 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
19292 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
19293 floating point, and decimal floating point to reload iterator.
19295 * config/rs6000/constraints.md (wl constraint): New constraints to
19296 return FLOAT_REGS if certain options are used to reduce the number
19297 of separate patterns that exist in the file.
19298 (wx constraint): Likewise.
19299 (wz constraint): Likewise.
19301 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19302 -mdebug=reg, print wg, wl, wx, and wz constraints.
19303 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19304 Initialize the reload functions for 64-bit binary/decimal floating
19306 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19307 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19308 create the buffer on the stack to overcome not having a 32-bit
19310 (rs6000_emit_move): Likewise.
19311 (rs6000_secondary_memory_needed_rtx): Likewise.
19312 (rs6000_alloc_sdmode_stack_slot): Likewise.
19313 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19314 via xxlxor, just like DFmode 0.0.
19316 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
19317 define as 1 if we are running on a power7 or newer.
19318 (enum r6000_reg_class_enum): Add new constraints.
19320 * config/rs6000/dfp.md (movsd): Delete, combine with binary
19321 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19322 with other moves by using conditional constraits (wg). Use LFIWZX
19323 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
19324 (movsd splitter): Likewise.
19325 (movsd_hardfloat): Likewise.
19326 (movsd_softfloat): Likewise.
19328 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
19329 binary and decimal floating point moves.
19330 (fmove_ok): New attributes to combine binary and decimal floating
19331 point moves, and to combine power6x (mfpgpr) moves along normal
19333 (real_value_to_target): Likewise.
19334 (f32_lr): Likewise.
19335 (f32_lm): Likewise.
19336 (f32_li): Likewise.
19337 (f32_sr): Likewise.
19338 (f32_sm): Likewise.
19339 (f32_si): Likewise.
19340 (movsf): Combine binary and decimal floating point moves. Combine
19341 power6x (mfpgpr) moves with other moves by using conditional
19342 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
19343 (mov<mode> for SFmode/SDmode); Likewise.
19344 (SFmode/SDmode splitters): Likewise.
19345 (movsf_hardfloat): Likewise.
19346 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
19347 (movsf_softfloat): Likewise.
19348 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
19350 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
19351 wx and wz constraints.
19353 * config/rs6000/constraints.md (wg constraint): New constraint to
19354 return FLOAT_REGS if -mmfpgpr (power6x) was used.
19356 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
19359 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19360 -mdebug=reg, print wg, wl, wx, and wz constraints.
19361 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19362 Initialize the reload functions for 64-bit binary/decimal floating
19364 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19365 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19366 create the buffer on the stack to overcome not having a 32-bit
19368 (rs6000_emit_move): Likewise.
19369 (rs6000_secondary_memory_needed_rtx): Likewise.
19370 (rs6000_alloc_sdmode_stack_slot): Likewise.
19371 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19372 via xxlxor, just like DFmode 0.0.
19374 * config/rs6000/dfp.md (movdd): Delete, combine with binary
19375 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19376 with other moves by using conditional constraits (wg). Use LFIWZX
19377 and STFIWX for loading SDmode on power7.
19378 (movdd splitters): Likewise.
19379 (movdd_hardfloat32): Likewise.
19380 (movdd_softfloat32): Likewise.
19381 (movdd_hardfloat64_mfpgpr): Likewise.
19382 (movdd_hardfloat64): Likewise.
19383 (movdd_softfloat64): Likewise.
19385 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
19386 64-bit binary and decimal floating point moves.
19387 (FMOVE64X): Likewise.
19388 (movdf): Combine 64-bit binary and decimal floating point moves.
19389 Combine power6x (mfpgpr) moves with other moves by using
19390 conditional constraits (wg).
19391 (mov<mode> for DFmode/DDmode): Likewise.
19392 (DFmode/DDmode splitters): Likewise.
19393 (movdf_hardfloat32): Likewise.
19394 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
19395 (movdf_softfloat32): Likewise.
19396 (movdf_hardfloat64_mfpgpr): Likewise.
19397 (movdf_hardfloat64): Likewise.
19398 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
19399 (movdf_softfloat64): Likewise.
19400 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
19401 (reload_<mode>_load): Move to later in the file so they aren't in
19402 the middle of the floating point move insns.
19403 (reload_<mode>_store): Likewise.
19405 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
19408 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
19409 constraint if -mdebug=reg.
19410 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
19411 Enable using dd reload support if needed.
19413 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
19414 binary and decimal floating point moves in rs6000.md.
19415 (movtd_internal): Likewise.
19417 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
19418 decimal floating point moves.
19420 (movtf_internal): Likewise.
19421 (mov<mode>_internal, TDmode/TFmode): Likewise.
19422 (movtf_softfloat): Likewise.
19423 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
19425 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
19426 movdi_internal64, using wg constraint for move direct operations.
19427 (movdi_internal64): Likewise.
19429 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
19430 MODES_TIEABLE_P for selected modes. Print the numerical value of
19431 the various virtual registers. Use GPR/FPR first/last values,
19432 instead of hard coding the register numbers. Print which modes
19433 have reload functions registered.
19434 (rs6000_option_override_internal): If -mdebug=reg, trace the options
19435 settings before/after setting cpu, target and subtarget settings.
19436 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
19437 and for secondary reload failures in rs6000_secondary_reload_inner.
19438 (rs6000_secondary_reload_fail): Likewise.
19439 (rs6000_secondary_reload_inner): Likewise.
19441 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
19442 macros for first/last GPR and FPR registers.
19443 (LAST_GPR_REGNO): Likewise.
19444 (FIRST_FPR_REGNO): Likewise.
19445 (LAST_FPR_REGNO): Likewise.
19447 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
19448 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
19449 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
19450 (vcond<mode><mode>): Likewise.
19451 (vcondu<mode><mode>): Likewise.
19452 (vector_gtu<mode>): Likewise.
19453 (vector_gte<mode>): Likewise.
19454 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
19455 to prevent the compiler from converting DImode operations to TImode.
19456 (ior<mode>3): Likewise.
19457 (and<mode>3): Likewise.
19458 (one_cmpl<mode>2): Likewise.
19459 (nor<mode>3): Likewise.
19460 (andc<mode>3): Likewise.
19462 * config/rs6000/constraints.md (wt constraint): New constraint
19463 that returns VSX_REGS if TImode is allowed in VSX registers.
19465 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
19466 constant under VSX.
19468 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
19469 similar to TImode, but it is restricted to being in the GPRs.
19471 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
19472 TImode to occupy a single VSX register.
19474 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
19475 -mvsx-timode for power7/power8.
19476 (power7 cpu): Likewise.
19477 (power8 cpu): Likewise.
19479 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
19480 sure that TFmode/TDmode take up two registers if they are ever
19481 allowed in the upper VSX registers.
19482 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
19484 (rs6000_init_hard_regno_mode_ok): Likewise.
19485 (rs6000_debug_reg_global): Add debugging for PTImode and wt
19486 constraint. Print if LRA is turned on.
19487 (rs6000_option_override_internal): Give an error if -mvsx-timode
19488 and VSX is not enabled.
19489 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
19490 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
19491 to reg+offset addressing. Use PTImode when checking offset
19492 addresses for validity.
19493 (reg_offset_addressing_ok_p): Likewise.
19494 (rs6000_legitimate_offset_address_p): Likewise.
19495 (rs6000_legitimize_address): Likewise.
19496 (rs6000_legitimize_reload_address): Likewise.
19497 (rs6000_legitimate_address_p): Likewise.
19498 (rs6000_eliminate_indexed_memrefs): Likewise.
19499 (rs6000_emit_move): Likewise.
19500 (rs6000_secondary_reload): Likewise.
19501 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
19502 reloads to fpr registers to continue to use reg+offset addressing,
19503 but 64-bit reloads to altivec registers need reg+reg addressing.
19504 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
19505 it. Treat LO_SUM like a PLUS operation.
19506 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
19507 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
19508 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
19509 registers to share a register with a smaller sized type, since VSX
19510 puts scalars in the upper 64-bits.
19511 (print_operand): Add support for PTImode.
19512 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
19513 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
19514 registers, but don't have arithmetic support.
19515 (rs6000_memory_move_cost): Add test for VSX.
19516 (rs6000_opt_masks): Add -mvsx-timode.
19518 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
19521 (VSr): Use wt constraint for TImode.
19522 (VSv): Drop TImode support.
19523 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
19524 (vsx_movti_64bit): Likewise.
19525 (vsx_movti_32bit): Likewise.
19526 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
19527 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
19528 one '?' on the appropriate output constraint. Do not allow TImode
19529 logical operations on 32-bit systems.
19530 (vsx_ior<mode>3): Likewise.
19531 (vsx_xor<mode>3): Likewise.
19532 (vsx_one_cmpl<mode>2): Likewise.
19533 (vsx_nor<mode>3): Likewise.
19534 (vsx_andc<mode>3): Likewise.
19535 (vsx_concat_<mode>): Likewise.
19536 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
19538 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19539 OPTION_MASK_VSX_TIMODE.
19540 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19541 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19543 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19544 (TI2 iterator): New iterator for TImode, PTImode.
19545 (wd mode attribute): Add values for vector types.
19546 (movti_string): Replace TI move operations with operations for TImode
19547 and PTImode. Add support for TImode being allowed in VSX registers.
19548 (mov<mode>_string, TImode/PTImode): Likewise.
19549 (movti_ppc64): Likewise.
19550 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19551 (TI mode splitters): Likewise.
19553 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19556 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19558 PR tree-optimization/56355
19559 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19560 Also handle integers with undefined overflow.
19562 2013-03-20 Catherine Moore <clm@codesourcery.com>
19563 Maciej W. Rozycki <macro@codesourcery.com>
19564 Tom de Vries <tom@codesourcery.com>
19565 Nathan Sidwell <nathan@codesourcery.com>
19566 Iain Sandoe <iain@codesourcery.com>
19567 Nathan Froyd <froydnj@codesourcery.com>
19568 Chao-ying Fu <fu@mips.com>
19570 * doc/extend.texi: (micromips, nomicromips, nocompression):
19571 Document new function attributes.
19572 * doc/invoke.texi (minterlink-compressed, mmicromips,
19573 m14k, m14ke, m14kec): Document new options.
19574 (minterlink-mips16): Update documentation.
19575 * doc/md.texi (ZC, ZD): Document new constraints.
19576 * configure.ac (gcc_cv_as_micromips): Check if linker
19577 supports the .set micromips directive.
19578 * configure: Regenerate.
19579 * config.in: Regenerate.
19580 * config/mips/mips-tables.opt: Regenerate.
19581 * config/mips/micromips.md: New file.
19582 * constraints.md (ZC, ZD): New constraints.
19583 * config/mips/predicates.md (movep_src_register): New predicate.
19584 (movep_src_operand): New predicate.
19585 (non_volatile_mem_operand): New predicate.
19586 * config/mips/mips.md (multimem): New type.
19587 (length): Differentiate between 17-bit and 18-bit branch offsets.
19588 (MOVEP1, MOVEP2): New mode iterator.
19589 (mov_<load>l): Use ZC constraint.
19590 (mov_<load>r): Likewise.
19591 (mov_<store>l): Likewise.
19592 (mov_<store>r): Likewise.
19593 (*branch_equality<mode>_inverted): Add microMIPS support.
19594 (*branch_equality<mode>): Likewise.
19595 (*jump_absolute): Likewise.
19596 (indirect_jump_<mode>): Likewise.
19597 (tablejump_<mode>): Likewise.
19598 (<optab>_internal): Likewise.
19599 (sibcall_internal): Likewise.
19600 (sibcall_value_internal): Likewise.
19601 (prefetch): Use constraint ZD.
19602 * config/mips/mips.opt (minterlink-compressed): New option.
19603 (minterlink-mips16): Now an alias for minterlink-compressed.
19604 (mmicromips): New option.
19605 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19606 (compare_and_swap_12): Likewise.
19607 (sync_add<mode>): Likewise.
19608 (sync_<optab>_12): Likewise.
19609 (sync_old_<optab>_12): Likewise.
19610 (sync_new_<optab>_12): Likewise.
19611 (sync_nand_12): Likewise.
19612 (sync_old_nand_12): Likewise.
19613 (sync_new_nand_12): Likewise.
19614 (sync_sub<mode>): Likewise.
19615 (sync_old_add<mode>): Likewise.
19616 (sync_old_sub<mode>): Likewise.
19617 (sync_new_add<mode>): Likewise.
19618 (sync_new_sub<mode>): Likewise.
19619 (sync_<optab><mode>): Likewise.
19620 (sync_old_<optab><mode>): Likewise.
19621 (sync_new_<optab><mode>): Likewise.
19622 (sync_nand<mode>): Likewise.
19623 (sync_old_nand<mode>): Likewise.
19624 (sync_new_nand<mode>): Likewise.
19625 (sync_lock_test_and_set<mode>): Likewise.
19626 (test_and_set_12): Likewise.
19627 (atomic_compare_and_swap<mode>): Likewise.
19628 (atomic_exchange<mode>_llsc): Likewise.
19629 (atomic_fetch_add<mode>_llsc): Likewise.
19630 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19631 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19632 (umips_save_restore_pattern_p): Likewise.
19633 (umips_load_store_pair_p): Likewise.
19634 (umips_output_load_store_pair): Likewise.
19635 (umips_movep_target_p): Likewise.
19636 (umips_12bit_offset_address_p): Likewise.
19637 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19638 (mips_base_mips16): Rename this...
19639 (mips_base_compression_flags): ...to this. Update all uses.
19640 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19641 (mips_mips16_decl_p): Delete.
19642 (mips_nomips16_decl_p): Delete.
19643 (mips_get_compress_on_flags): New function.
19644 (mips_get_compress_off_flags): New function.
19645 (mips_get_compress_mode): New function.
19646 (mips_get_compress_on_name): New function.
19647 (mips_get_compress_off_name): New function.
19648 (mips_insert_attributes): Support multiple compression types.
19649 (mips_merge_decl_attributes): Likewise.
19650 (umips_12bit_offset_address_p): New function.
19651 (mips_start_function_definition): Emit .set micromips directive.
19652 (mips_call_may_need_jalx_p): New function.
19653 (mips_function_ok_for_sibcall): Add microMIPS support.
19654 (mips_print_operand_punctuation): Support short delay slots and
19656 (umips_swm_mask, umips_swm_encoding): New.
19657 (umips_build_save_restore): New function.
19658 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19659 (was_mips16_p): Remove.
19660 (old_compression_mode): New.
19661 (mips_set_compression_mode): New function.
19662 (mips_set_current_function): Add microMIPS support.
19663 (mips_option_override): Likewise.
19664 (umips_save_restore_pattern_p): New function.
19665 (umips_output_save_restore): New function.
19666 (umips_load_store_pair_p_1): New function.
19667 (umips_load_store_pair_p): New function.
19668 (umips_output_load_store_pair_1): New function.
19669 (umips_output_load_store_pair): New function.
19670 (umips_movep_target_p) New function.
19671 (mips_prepare_pch_save): Add microMIPS support.
19672 * config/mips/mips.h (TARGET_COMPRESSION): New.
19673 (TARGET_CPU_CPP_BUILTINS): Update macro
19674 to use new compression flags and to support microMIPS.
19675 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19676 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19677 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19678 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19679 (ASM_SPEC): Support mmicromips and mno-micromips.
19680 (M16STORE_REG_P): New macro.
19681 (MIPS_CALL): Support TARGET_MICROMIPS.
19682 (MICROMIPS_J): New macro.
19683 (mips_base_mips16): Rename this...
19684 (mips_base_compression_flags): ...to this.
19685 (UMIPS_12BIT_OFFSET_P): New macro.
19686 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19687 (MULTILIB_DIRNAMES): Likewise.
19688 2013-03-20 Richard Biener <rguenther@suse.de>
19690 PR tree-optimization/56661
19691 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19692 the result does not have to be distinct.
19694 2013-03-20 Richard Biener <rguenther@suse.de>
19696 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19699 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19700 Steven Bosscher <steven@gcc.gnu.org>
19702 PR rtl-optimization/56605
19703 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19705 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19708 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19709 that require movd instead of movq.
19711 2013-03-20 Richard Biener <rguenther@suse.de>
19713 * tree-ssa-structalias.c (struct variable_info): Add pointer
19714 to the first field of an aggregate with sub-vars. Make
19715 this and the pointer to the next subfield its ID.
19716 (vi_next): New function.
19717 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19718 storedanything_id, integer_id): Increment by one.
19719 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19720 get_call_clobber_vi): Adjust.
19721 (solution_set_expand): Simplify and speedup.
19722 (solution_set_add): Inline into ...
19723 (set_union_with_increment): ... this. Adjust accordingly.
19724 (do_sd_constraint): Likewise.
19725 (do_ds_constraint): Likewise.
19726 (do_complex_constraint): Simplify.
19727 (build_pred_graph): Adjust.
19728 (solve_graph): Likewise. Simplify and speedup.
19729 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19730 get_constraint_for_component_ref, get_constraint_for_1,
19731 first_vi_for_offset, first_or_preceding_vi_for_offset,
19732 create_function_info_for, create_variable_info_for_1,
19733 create_variable_info_for, intra_create_variable_infos): Adjust.
19734 (init_base_vars): Push NULL for ID zero.
19735 (compute_points_to_sets): Adjust.
19737 2013-03-20 Richard Biener <rguenther@suse.de>
19739 * cfgloop.c (verify_loop_structure): Streamline and avoid
19740 ICEing on corrupt loop tree.
19741 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19744 2013-03-20 Richard Biener <rguenther@suse.de>
19746 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19747 check whether an SSA update is needed.
19749 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19751 * config/mips/constraints.md (T): Rename to...
19755 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19756 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19758 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19760 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19761 (*subsi3_carryin_uxtw): Likewise.
19763 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19765 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19766 (*rorsi3_insn_uxtw): Likewise.
19768 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19770 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19771 (*extrsi5_insn_uxtw): Likewise.
19773 2013-03-19 Richard Biener <rguenther@suse.de>
19775 PR tree-optimization/56273
19776 * passes.c (init_optimization_passes): Move second VRP after DOM.
19778 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19780 * config/i386/i386.md (*movti_internal): Merge from
19781 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19782 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19783 nox64 isa attributes.
19785 2013-03-18 Richard Biener <rguenther@suse.de>
19787 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19789 (merge_node_constraints): Likewise.
19790 (build_succ_graph): Likewise.
19791 (valid_graph_edge): Inline into single caller.
19792 (unify_nodes): Likewise. Use bitmap_set_bit return value
19794 (scc_visit): Fix formatting and variable use.
19795 (do_sd_constraint): Use gcc_checking_assert.
19796 (do_ds_constraint): Likewise.
19797 (do_complex_constraint): Likewise.
19798 (condense_visit): Likewise. Cleanup.
19799 (dump_pred_graph): New function.
19800 (perform_var_substitution): Dump the pred-graph before
19801 variable substitution.
19802 (find_equivalent_node): Use gcc_checking_assert.
19803 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19805 2013-03-18 Richard Biener <rguenther@suse.de>
19807 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19808 Remove cond_expr_stmt_list argument and do not gimplify the
19810 (vect_loop_versioning): Adjust.
19811 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19812 Cleanup to use less temporaries.
19813 (vect_create_data_ref_ptr): Cleanup.
19815 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19817 PR tree-optimization/56635
19818 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19819 require types_compatible_p types.
19821 2013-03-18 Nick Clifton <nickc@redhat.com>
19823 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19824 spurious backslash.
19826 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19827 Add missing line to comment describing function.
19829 2013-03-18 Richard Biener <rguenther@suse.de>
19831 PR tree-optimization/56210
19832 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19833 Handle string / character search functions.
19834 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19836 2013-03-18 Richard Biener <rguenther@suse.de>
19838 PR middle-end/56483
19839 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19840 and implement properly.
19841 * gimple.h (gimple_cond_single_var_p): Remove.
19843 2013-03-18 Richard Biener <rguenther@suse.de>
19845 * tree-data-ref.h (find_data_references_in_loop): Declare.
19846 * tree-data-ref.c (get_references_in_stmt): Use a stack
19847 vector pre-allocated in the callers.
19848 (find_data_references_in_stmt): Adjust.
19849 (graphite_find_data_references_in_stmt): Likewise.
19850 (create_rdg_vertices): Likewise.
19851 (find_data_references_in_loop): Export.
19852 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19853 Compute dependences here...
19854 (vect_analyze_data_refs): ...not here. When we encounter
19855 a non-vectorizable data reference in basic-block vectorization
19856 truncate the data reference vector. Do not bother to
19857 fixup data-dependence information for gather loads.
19858 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19859 of data references, as reported.
19861 2013-03-18 Richard Biener <rguenther@suse.de>
19863 PR tree-optimization/3713
19864 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19865 has_constants and expr.
19866 (stmt_has_constants): Properly valueize SSA names when deciding
19867 whether the stmt has constants.
19869 2013-03-18 Richard Biener <rguenther@suse.de>
19871 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19872 whole function when there is nothing to do.
19873 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19874 * tree-vectorizer.c (vectorize_loops): Update virtual and
19875 loop-closed SSA once.
19876 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19878 2013-03-18 Richard Biener <rguenther@suse.de>
19880 PR middle-end/56113
19881 * domwalk.c (bb_postorder): New global static.
19882 (cmp_bb_postorder): New function.
19883 (walk_dominator_tree): Replace scheme imposing an order for
19884 visiting dominator sons by one sorting them at the time they
19885 are pushed on the stack.
19887 2013-03-18 Richard Biener <rguenther@suse.de>
19889 PR tree-optimization/39326
19890 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19891 (struct mem_ref): Replace mem member with ao_ref typed member.
19892 (MEM_ANALYZABLE): Adjust.
19893 (memref_eq): Likewise.
19894 (mem_ref_alloc): Likewise.
19895 (gather_mem_refs_stmt): Likewise.
19896 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19897 (execute_sm_if_changed_flag_set): Adjust.
19898 (execute_sm): Likewise.
19899 (ref_always_accessed_p): Likewise.
19900 (refs_independent_p): Likewise.
19901 (can_sm_ref_p): Likewise.
19903 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19906 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19907 return 1 even for !unsignedp.
19909 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19911 * config/i386/i386.md (isa): Add x64 and nox64.
19912 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19913 (*pushtf): Enable *roF alternative for x64 isa only.
19914 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19915 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19916 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19917 (*movtf_internal): Merge from *movtf_internal_rex64 and
19918 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19919 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19920 nox64 isa attributes.
19921 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19922 nox64 isa attributes.
19923 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19925 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19927 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19929 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19932 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19933 class names. Remove trailing comma after "ALL_REGS".
19935 2013-03-16 Jan Hubicka <jh@suse.cz>
19937 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19938 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19939 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19940 of cgraph_get_create_node.
19941 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19943 2013-03-16 Jason Merrill <jason@redhat.com>
19946 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19947 with DW_AT_default_value.
19949 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19951 * BASE-VER: Set to 4.9.0.
19953 2013-03-14 Andi Kleen <ak@linux.intel.com>
19956 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19957 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19958 Document _x* TSX intrinsics.
19960 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19961 David Holsgrove <david.holsgrove@xilinx.com>
19963 * configure.ac: Add MicroBlaze TLS support detection.
19964 * configure: Regenerate.
19965 * config/microblaze/microblaze-protos.h
19966 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19967 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19968 * config/microblaze/microblaze.c (microblaze_address_type): Add
19969 ADDRESS_TLS and tls_reloc address types.
19970 (microblaze_address_info): Add tls_reloc.
19971 (TARGET_HAVE_TLS): Define.
19972 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19973 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19974 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19975 load_tls_operand, microblaze_call_tls_get_addr,
19976 microblaze_legitimize_tls_address): New functions.
19977 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19978 (get_base_reg): Use microblaze_tls_symbol_p.
19979 (microblaze_classify_address): Handle TLS.
19980 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19981 label_mentioned_p and microblaze_tls_referenced_p.
19982 (microblaze_legitimize_address): Handle TLS.
19983 (microblaze_address_insns): Handle ADDRESS_TLS.
19984 (pic_address_needs_scratch): Handle TLS.
19985 (print_operand_address): Handle TLS.
19986 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19987 (microblaze_expand_move): Handle TLS.
19988 (microblaze_legitimate_constant_p): Check
19989 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19990 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19991 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19992 (PIC_OFFSET_TABLE_REGNUM): Set.
19993 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19994 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19995 (addsi3, movsi_internal2, movdf_internal): Update constraints
19996 * config/microblaze/predicates.md (arith_plus_operand): Define
19997 (move_operand): Redefine as move_src_operand,
19998 check microblaze_tls_referenced_p.
20000 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20002 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
20003 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
20005 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20007 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
20010 2013-03-14 Jakub Jelinek <jakub@redhat.com>
20012 PR tree-optimization/53265
20013 * common.opt (Waggressive-loop-optimizations): New option.
20014 * tree-ssa-loop-niter.c: Include tree-pass.h.
20015 (do_warn_aggressive_loop_optimizations): New function.
20016 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20017 if number_of_latch_executions returned constant.
20018 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20019 early. If number_of_latch_executions returned constant, set
20020 nb_iterations_upper_bound back to it.
20021 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20023 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20024 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20026 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20027 (MULTILIB_OSDIRNAMES): Set.
20028 * genmultilib: If defaultosdirname doesn't start with :: , set
20029 defaultosdirname2 instead, clear it and emit two . multilib_raw
20030 entries instead of just one.
20032 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20034 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20035 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20036 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20037 (SUBTARGET_OVERRIDE_OPTIONS): New.
20039 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20042 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20043 (musermode): Convert to Var(TARGET_USERMODE).
20044 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20045 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20046 * config/sh/sh.c (sh_option_override): Use
20047 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20048 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20050 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20052 (udivsi3_i4_single, divsi3_i4_single): Use
20053 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20055 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20057 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20060 2013-03-13 Richard Biener <rguenther@suse.de>
20062 PR tree-optimization/56608
20063 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20064 calls when vectorizing basic-blocks.
20066 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20069 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20072 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20074 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20076 2013-03-11 Jan Hubicka <jh@suse.cz>
20079 * lto-streamer-out.c (output_symbol_p): Skip references from
20080 constructors of external variables.
20082 2013-03-11 Jan Hubicka <jh@suse.cz>
20084 PR middle-end/56571
20085 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20087 * emit-rtl.c (verify_rtx_sharing): Likewise.
20088 (copy_insn_1): Likewise.
20089 * rtl.c (copy_rtx): Likewise.
20091 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20094 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20095 output_operand_lossage message.
20097 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20100 * arm.c (shift_op): Validate RTL pattern on the fly.
20101 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20104 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20107 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20108 new call patterns that clobber %fr12.
20109 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20110 split and postreload patterns.
20111 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20112 registers %fr12 and %fr12R as call used.
20114 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20116 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20117 (canon_address, record_store, replace_read, check_mem_read_rtx,
20118 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20119 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20120 rest_of_handle_dse): Likewise.
20122 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
20124 PR middle-end/56524
20125 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
20127 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
20128 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
20129 (save_optabs_if_changed): Replace with...
20130 (init_tree_optimization_optabs): ...this.
20131 * optabs.c (save_optabs_if_changed): Rename to...
20132 (init_tree_optimization_optabs): ...this. Take the optimization node
20133 as argument. Do nothing if the base optabs are already correct.
20134 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
20135 to recompute optabs.
20136 * function.h (function): Remove optabs field.
20137 * function.c (invoke_set_current_function_hook): Call
20138 init_tree_optimization_optabs. Use the result to initialize
20141 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
20143 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
20144 if GTMA_HAS_NO_INSTRUMENTATION.
20145 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
20146 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
20147 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
20148 * gimple-pretty-print.c (dump_gimple_transaction): Handle
20149 GTMA_HAS_NO_INSTRUMENTATION.
20151 2013-03-08 Jakub Jelinek <jakub@redhat.com>
20153 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
20156 2013-03-08 Marek Polacek <polacek@redhat.com>
20157 Jakub Jelinek <jakub@redhat.com>
20159 PR tree-optimization/56478
20160 * predict.c (is_comparison_with_loop_invariant_p): Change the
20161 type of loop_step to tree.
20162 (predict_loops): Adjust.
20163 (predict_iv_comparison): Perform the computations on double_ints.
20165 2013-03-08 Richard Biener <rguenther@suse.de>
20167 PR tree-optimization/56570
20168 * tree-cfg.c (verify_expr_location_1): Verify locations for
20170 * tree-sra.c (create_access_replacement): Strip locations
20171 from DECL_DEBUG_EXPRs.
20173 2013-03-08 Richard Biener <rguenther@suse.de>
20175 * tree-inline.c (expand_call_inline): Do not associate
20176 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
20177 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
20179 2013-03-08 Richard Biener <rguenther@suse.de>
20181 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
20182 or block changes with -Og. Fix for location / block encoding
20183 changes and PHI arguments with locations.
20185 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
20187 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
20189 (struct output_info): Likewise.
20190 (register_overhead): Remove bad gcc_assert.
20191 (bitmap_find_bit): If there is only a single bitmap element, do not
20192 count a miss as a search.
20193 (print_statistics): Update for counter type changes.
20194 (dump_bitmap_statistics): Likewise. Print headers such that they
20195 are properly lined up with the printed counters.
20197 2013-03-07 Jakub Jelinek <jakub@redhat.com>
20199 PR tree-optimization/56559
20200 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
20201 check that it has only a single use.
20203 2013-03-07 Richard Biener <rguenther@suse.de>
20205 * doc/invoke.texi (fwhole-program): Discourage use in combination
20208 2013-03-06 Jakub Jelinek <jakub@redhat.com>
20210 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
20212 PR tree-optimization/56539
20213 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
20214 instead of GSI_CONTINUE_LINKING as last argument to
20215 force_gimple_operand_gsi. Adjust function comment.
20217 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
20220 PR middle-end/56548
20221 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
20222 promoted mode, convert the result back to the original mode.
20224 2013-03-06 Richard Biener <rguenther@suse.de>
20226 PR middle-end/56294
20227 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
20228 (insert_updated_phi_nodes_compare_uids): New function.
20229 (update_ssa): Sort symbols_to_rename after UID before
20230 traversing it to insert PHI nodes.
20232 2013-03-06 Richard Biener <rguenther@suse.de>
20234 PR middle-end/50494
20235 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
20236 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
20239 2013-02-13 Richard Biener <rguenther@suse.de>
20242 * varasm.c (output_constant_def_1): Get the decl representing
20243 the constant as argument.
20244 (output_constant_def): Wrap output_constant_def_1.
20245 (make_decl_rtl): Use output_constant_def_1 with the decl
20246 representing the constant.
20247 (build_constant_desc): Optionally re-use a decl already
20248 representing the constant.
20249 (tree_output_constant_def): Adjust.
20251 2013-03-06 Joey Ye <joey.ye@arm.com>
20254 * gcc.c (convert_white_space): New function.
20255 (main): Handles white space in function name.
20257 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20260 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20261 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20262 to SH_DIV_CALL_TABLE for TARGET_SH2.
20263 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20265 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20266 call-table options.
20268 2013-03-05 Sterling Augustine <saugustine@google.com>
20269 Cary Coutant <ccoutant@google.com>
20272 * dwarf2out.c (resolve_addr): Don't call
20273 remove_loc_list_addr_table_entries a second time for the same
20276 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20279 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
20280 (avoid_complex_debug_insns): New function.
20281 (expand_debug_locations): Call it.
20283 PR rtl-optimization/56484
20284 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
20285 lifetimes of hard registers on small register class machines.
20287 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20289 * config/microblaze/microblaze-protos.h: Rename
20290 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
20291 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
20293 (microblaze_fast_interrupt_function_p): New function.
20294 (microblaze_is_interrupt_handler): Rename to
20295 microblaze_is_interrupt_variant and add fast_interrupt check.
20296 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
20297 (save_restore_insns): Likewise.
20298 (compute_frame_size): Likewise.
20299 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
20300 (microblaze_globalize_label): Likewise.
20301 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
20302 * config/microblaze/microblaze.md: Use wrapper
20303 microblaze_is_interrupt_variant.
20305 2013-03-05 Kai Tietz <ktietz@redhat.com>
20307 * sdbout.c (sdbout_one_type): Switch to current function's section
20308 supporting cold/hot.
20310 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20312 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
20315 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20317 PR middle-end/56461
20318 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
20319 if VALGRIND_GET_VBITS is defined, temporarily make object
20320 memory all defined, and restore previous valgrind addressability
20321 and definability afterwards. Free this_object at the end.
20323 PR middle-end/56461
20324 * lra.c (lra): Call lra_clear_live_ranges if live_p,
20325 right before calling lra_create_live_ranges, also call it
20326 when clearing live_p. Only call lra_clear_live_ranges
20327 at the end if live_p.
20329 PR middle-end/56461
20330 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
20332 2013-03-05 Richard Biener <rguenther@suse.de>
20334 PR tree-optimization/56521
20335 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
20338 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
20341 * except.h (remove_unreachable_eh_regions): New prototype.
20342 * except.c (remove_eh_handler_splicer): New function, split out
20343 of remove_eh_handler.
20344 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
20345 warning about running it on many EH regions one at a time.
20346 (remove_unreachable_eh_regions_worker): New function, walk the
20347 EH tree in depth-first order and remove non-marked regions.
20348 (remove_unreachable_eh_regions): New function.
20349 * tree-eh.c (mark_reachable_handlers): New function, split out
20350 from remove_unreachable_handlers.
20351 (remove_unreachable_handlers): Use mark_reachable_handlers and
20352 remove_unreachable_eh_regions.
20353 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
20354 and remove_unreachable_eh_regions.
20356 2013-03-05 Richard Biener <rguenther@suse.de>
20358 PR middle-end/56525
20359 * loop-init.c (fix_loop_structure): Remove loops in two stages,
20360 not freeing them until the end.
20362 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20364 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
20366 2013-03-05 Richard Biener <rguenther@suse.de>
20368 PR tree-optimization/56270
20369 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
20370 of loads after scheduling an SLP instance.
20372 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20374 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
20376 (check_gcc_parallelize): Run guality.exp as a separate job from
20377 vect.exp with unsorted.exp and $(dg_target_exps) separately from
20378 struct-layout-1.exp with stackalign.exp.
20380 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
20382 PR middle-end/56461
20383 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
20384 load_index sbitmap even if some bit in it isn't set.
20386 PR middle-end/56461
20387 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
20388 (discover_iteration_bound_by_body_walk): Change queues to
20389 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
20390 spelling in comment. Call safe_push on queues[bound_index] directly.
20391 Release queues[queue_index] in every iteration unconditionally.
20392 Release bounds vector.
20394 PR middle-end/56461
20395 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
20396 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
20397 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
20400 2013-03-05 Richard Biener <rguenther@suse.de>
20403 * tree-inline.c (remap_blocks_to_null): New function.
20404 (expand_call_inline): When expanding a call stmt without
20405 an associated block inline remap all callee blocks to NULL.
20407 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20409 PR rtl-optimization/56494
20410 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
20411 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
20412 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
20414 PR middle-end/56461
20415 * sel-sched-ir.c (free_sched_pools): Release
20416 succs_info_pool.stack[succs_info_pool.max_top] vectors too
20417 if succs_info_pool.max_top isn't -1.
20420 * opts.c (opts_obstack, opts_concat): Moved to...
20421 * opts-common.c (opts_obstack, opts_concat): ... here.
20423 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20425 PR middle-end/56461
20426 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
20428 2013-03-04 Martin Jambor <mjambor@suse.cz>
20430 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
20431 all appropriate places.
20433 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
20435 PR tree-optimization/56424
20436 * ipa-split.c (split_function): Do not set the RSO flag if result is
20437 not by reference and its type is a register type.
20439 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
20441 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
20442 (microblaze_legitimate_pic_operand): Likewise
20443 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
20444 new function microblaze_legitimate_pic_operand
20445 * config/microblaze/microblaze-protos.h
20446 (microblaze_legitimate_pic_operand): Declare.
20448 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20450 * config/microblaze/predicates.md (call_insn_simple_operand):
20451 New predicate for supported rtx code types.
20452 * config/microblaze/microblaze.md (call_internal1): Use
20453 call_insn_simple_operand predicate.
20455 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20457 PR middle-end/56461
20458 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
20459 partitions.ordered_remove.
20461 PR middle-end/56461
20462 * tree-vect-stmts.c (vectorizable_conversion): Don't call
20463 vec_oprnds0.create (1) for modifier == NONE.
20465 PR middle-end/56461
20466 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
20467 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
20468 vec_oprnds1 right before pushing anything to it for
20471 PR middle-end/56461
20472 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
20473 set nbbs to 0 instead of having separate code path.
20474 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
20475 instead of false as last argument if returning NULL.
20477 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
20479 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
20480 the attribute is now called "target" instead of "option".
20481 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
20482 * doc/tm.texi.in (Target Attributes): Likewise document the correct
20483 attribute/pragma name for TARGET_OPTION_VALID_P and
20484 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
20485 * doc/tm.texi: Regenerated.
20487 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
20489 * config/microblaze/microblaze.c:
20490 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
20491 * config/microblaze/microblaze.h: Add -mxl-reorder to
20493 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
20494 instructions emitted if TARGET_REORDER.
20495 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
20496 or 0 for -m/-mno case, but initialises as 2 to detect default use case
20499 2013-03-01 Xinliang David Li <davidxl@google.com>
20501 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
20504 2013-03-01 Jakub Jelinek <jakub@redhat.com>
20506 PR middle-end/56461
20507 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
20508 vector even when returning true. Fix up function comment formatting.
20510 PR middle-end/56461
20511 * ira-build.c (ira_loop_nodes_count): New variable.
20512 (create_loop_tree_nodes): Initialize it.
20513 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
20515 PR middle-end/56461
20516 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
20517 method on dr_chain and result_chain.
20518 * tree-vect-stmts.c (vectorizable_store): Only call
20519 result_chain.create if j == 0.
20521 PR middle-end/56461
20522 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
20523 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
20524 before overwriting it.
20526 2013-03-01 Tobias Burnus <burnus@net-b.de>
20528 * doc/extended.texi (C Extensions): Change order in @menu
20530 (Other MIPS Built-in Functions): Move last MIPS entry before
20531 "picoChip Built-in Functions".
20532 (SH Built-in Functions): Move after RX Built-in Functions.
20533 * doc/gcc.texi (Introduction): Change order in @menu
20535 * doc/md.texi (Constraints): Ditto.
20536 * gty.texi (Type Information): Ditto.
20537 (User-provided marking routines for template types): Make
20539 * doc/invoke.texi (AArch64 Options): Move before
20540 "Adapteva Epiphany Options".
20542 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20543 Jakub Jelinek <jakub@redhat.com>
20546 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20547 no_address_safety_analysis attribute.
20548 * doc/extend.texi (no_address_safety_attribute): Rename to
20549 no_sanitize_address attribute, mention no_address_safety_analysis
20550 attribute as deprecated alias.
20552 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20554 PR middle-end/56461
20555 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20556 type to vec<vec<tree> > *.
20557 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20558 to be vec<tree> instead of vec<tree> *, set vec_defs
20559 to vNULL and call vec_defs.create (number_of_vects), adjust other
20561 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20562 vectorizable_condition): Adjust vect_get_slp_defs callers.
20564 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20566 * config/aarch64/aarch64.c
20567 (aarch64_float_const_representable): Remove unused variable.
20569 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20571 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20573 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20575 * config/aarch64/aarch64-builtins.c
20576 (aarch64_init_simd_builtins): Make static.
20578 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20580 * config/aarch64/aarch64.c
20581 (aarch64_simd_make_constant): Make static.
20583 2013-02-28 Martin Jambor <mjambor@suse.cz>
20585 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20586 with no initialization to the RHS of debug statements.
20588 2013-02-28 Martin Jambor <mjambor@suse.cz>
20590 PR tree-optimization/56294
20591 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20593 (get_access_replacement): Do not call create_access_replacement.
20594 Assert a replacement exists.
20595 (get_repl_default_def_ssa_name): Create the replacement declaration
20598 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20600 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20601 final_end_function.
20603 2013-02-28 Marek Polacek <polacek@redhat.com>
20605 PR rtl-optimization/56466
20606 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20607 if we're changing a loop.
20608 (peel_loops_completely): Likewise.
20610 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20613 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20615 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20618 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20619 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20620 INTX_FTYPE_FX, FX_FTYPE_INTX.
20621 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20623 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20625 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20626 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20627 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20628 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20629 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20630 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20631 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20632 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20633 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20634 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20635 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20636 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20637 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20638 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20639 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20640 (avrxmega6): Increase max flash segments from 5 to 6.
20641 * config/avr/t-multilib: Regenerate.
20642 * config/avr/avr-tables.opt: Regenerate.
20643 * doc/avr-mmcu.texi: Regenerate.
20645 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20647 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20648 (avr_device_to_arch): Rename to avr_device_to_ld.
20649 (avr_device_to_as): New prototype.
20650 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20651 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20652 * config/avr/driver-avr.c (avr_device_to_as): New.
20653 (avr_device_to_arch): Rename to avr_device_to_ld.
20655 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20657 PR middle-end/56461
20658 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20659 method on dr_chain and result_chain.
20661 PR middle-end/56461
20662 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20663 pointer_set_destroy on not_executed_last_iteration.
20665 PR middle-end/56461
20666 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20668 PR middle-end/56461
20669 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20670 FOR_EACH_DEFINED_FUNCTION when freeing state.
20672 PR middle-end/56461
20673 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20675 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20678 PR middle-end/56461
20679 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20680 known_aggs[i].items and release known_aggs vector.
20682 PR middle-end/56461
20683 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20685 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20687 * config/microblaze/microblaze.c (microblaze_emit_compare):
20688 Use xor for EQ/NE comparisions.
20689 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20690 (cbranchsf4): Adjust operator to comparison_operator.
20692 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20694 PR middle-end/56461
20695 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20697 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20698 vec_safe_push, always update *slot.
20699 (redirect_edge_var_map_clear): Use vec_free.
20700 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20701 (free_var_map_entry): Use vec_free.
20702 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20703 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20705 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20707 PR middle-end/45472
20708 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20709 when the may_trap_p bit of the exprs being merged differs.
20710 Reorder tests for speculativeness in the logical and operator.
20712 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20714 * incpath.c (add_standard_paths): Use reconcat instead of concat
20715 where appropriate and avoid leaking memory.
20717 * opts.h: Include obstack.h.
20718 (opts_concat): New prototype.
20719 (opts_obstack): New declaration.
20720 * opts.c (opts_concat): New function.
20721 (opts_obstack): New variable.
20722 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20723 (finish_options): Use opts_concat instead of concat
20724 and XOBNEWVEC instead of XNEWVEC.
20725 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20726 generate_option): Likewise.
20727 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20728 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20731 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20732 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20734 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20736 PR middle-end/56461
20737 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20739 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20741 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20742 (arm_block_move_unaligned_straight): Likewise.
20743 (arm_adjust_block_mem): Likewise.
20745 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20748 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20749 temp, cond and label.
20750 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20753 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20754 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20757 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20758 of prologue/epilogue functions.
20761 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20762 Remove unused variable cfa_offset.
20763 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20766 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20767 type promotion to unsigned.
20770 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20771 for HOST_WIDE_INT of 32 bit / same size as int.
20772 (arm_block_move_unaligned_straight): Likewise.
20773 (arm_adjust_block_mem): Likewise.
20776 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20779 2013-02-26 Marek Polacek <polacek@redhat.com>
20781 PR tree-optimization/56426
20782 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20784 2013-02-26 Richard Biener <rguenther@suse.de>
20787 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20788 unused variable loops.
20790 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20792 PR tree-optimization/56448
20793 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20794 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20795 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20796 later operands of the references, or even first operand for
20797 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20799 PR tree-optimization/56443
20800 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20801 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20802 to type_for_mode langhook.
20804 2013-02-25 Matt Turner <mattst88@gmail.com>
20806 * doc/invoke.texi: Document r4700.
20808 2013-02-25 Richard Biener <rguenther@suse.de>
20810 PR tree-optimization/56175
20811 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20813 (simplify_bitwise_binary): ... here. Also guard the conversion
20814 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20816 2013-02-25 Catherine Moore <clm@codesourcery.com>
20819 2013-02-24 Catherine Moore <clm@codesourcery.com>
20820 Maciej W. Rozycki <macro@codesourcery.com>
20821 Tom de Vries <tom@codesourcery.com>
20822 Nathan Sidwell <nathan@codesourcery.com>
20823 Iain Sandoe <iain@codesourcery.com>
20824 Nathan Froyd <froydnj@codesourcery.com>
20825 Chao-ying Fu <fu@mips.com>
20827 * doc/extend.texi: (micromips, nomicromips, nocompression):
20828 Document new function attributes.
20829 * doc/invoke.texi (minterlink-compressed, mmicromips,
20830 m14k, m14ke, m14kec): Document new options.
20831 (minterlink-mips16): Update documentation.
20832 * doc/md.texi (ZC, ZD): Document new constraints.
20833 * configure.ac (gcc_cv_as_micromips): Check if linker
20834 supports the .set micromips directive.
20835 * configure: Regenerate.
20836 * config.in: Regenerate.
20837 * config/mips/mips-tables.opt: Regenerate.
20838 * config/mips/micromips.md: New file.
20839 * constraints.md (ZC, AD): New constraints.
20840 * config/mips/predicates.md (movep_src_register): New predicate.
20841 (movep_src_operand): New predicate.
20842 (non_volatile_mem_operand): New predicate.
20843 * config/mips/mips.md (multimem): New type.
20844 (length): Differentiate between 17-bit and 18-bit branch offsets.
20845 (MOVEP1, MOVEP2): New mode iterator.
20846 (mov_<load>l): Use ZC constraint.
20847 (mov_<load>r): Likewise.
20848 (mov_<store>l): Likewise.
20849 (mov_<store>r): Likewise.
20850 (*branch_equality<mode>_inverted): Add microMIPS support.
20851 (*branch_equality<mode>): Likewise.
20852 (*jump_absolute): Likewise.
20853 (indirect_jump_<mode>): Likewise.
20854 (tablejump_<mode>): Likewise.
20855 (<optab>_internal): Likewise.
20856 (sibcall_internal): Likewise.
20857 (sibcall_value_internal): Likewise.
20858 (prefetch): Use constraint ZD.
20859 * config/mips/mips.opt (minterlink-compressed): New option.
20860 (minterlink-mips16): Now an alias for minterlink-compressed.
20861 (mmicromips): New option.
20862 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20863 (compare_and_swap_12): Likewise.
20864 (sync_add<mode>): Likewise.
20865 (sync_<optab>_12): Likewise.
20866 (sync_old_<optab>_12): Likewise.
20867 (sync_new_<optab>_12): Likewise.
20868 (sync_nand_12): Likewise.
20869 (sync_old_nand_12): Likewise.
20870 (sync_new_nand_12): Likewise.
20871 (sync_sub<mode>): Likewise.
20872 (sync_old_add<mode>): Likewise.
20873 (sync_old_sub<mode>): Likewise.
20874 (sync_new_add<mode>): Likewise.
20875 (sync_new_sub<mode>): Likewise.
20876 (sync_<optab><mode>): Likewise.
20877 (sync_old_<optab><mode>): Likewise.
20878 (sync_new_<optab><mode>): Likewise.
20879 (sync_nand<mode>): Likewise.
20880 (sync_old_nand<mode>): Likewise.
20881 (sync_new_nand<mode>): Likewise.
20882 (sync_lock_test_and_set<mode>): Likewise.
20883 (test_and_set_12): Likewise.
20884 (atomic_compare_and_swap<mode>): Likewise.
20885 (atomic_exchange<mode>_llsc): Likewise.
20886 (atomic_fetch_add<mode>_llsc): Likewise.
20887 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20888 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20889 (umips_save_restore_pattern_p): Likewise.
20890 (umips_load_store_pair_p): Likewise.
20891 (umips_output_load_store_pair): Likewise.
20892 (umips_movep_target_p): Likewise.
20893 (umips_12bit_offset_address_p): Likewise.
20894 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20895 (mips_base_mips16): Rename this...
20896 (mips_base_compression_flags): ...to this. Update all uses.
20897 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20898 (mips_mips16_decl_p): Delete.
20899 (mips_nomips16_decl_p): Delete.
20900 (mips_get_compress_on_flags): New function.
20901 (mips_get_compress_off_flags): New function.
20902 (mips_get_compress_mode): New function.
20903 (mips_get_compress_on_name): New function.
20904 (mips_get_compress_off_name): New function.
20905 (mips_insert_attributes): Support multiple compression types.
20906 (mips_merge_decl_attributes): Likewise.
20907 (umips_12bit_offset_address_p): New function.
20908 (mips_start_function_definition): Emit .set micromips directive.
20909 (mips_call_may_need_jalx_p): New function.
20910 (mips_function_ok_for_sibcall): Add microMIPS support.
20911 (mips_print_operand_punctuation): Support short delay slots and
20913 (umips_swm_mask, umips_swm_encoding): New.
20914 (umips_build_save_restore): New function.
20915 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20916 (was_mips16_p): Remove.
20917 (old_compression_mode): New.
20918 (mips_set_compression_mode): New function.
20919 (mips_set_current_function): Add microMIPS support.
20920 (mips_option_override): Likewise.
20921 (umips_save_restore_pattern_p): New function.
20922 (umips_output_save_restore): New function.
20923 (umips_load_store_pair_p_1): New function.
20924 (umips_load_store_pair_p): New function.
20925 (umips_output_load_store_pair_1): New function.
20926 (umips_output_load_store_pair): New function.
20927 (umips_movep_target_p) New function.
20928 (mips_prepare_pch_save): Add microMIPS support.
20929 * config/mips/mips.h (TARGET_COMPRESSION): New.
20930 (TARGET_CPU_CPP_BUILTINS): Update macro
20931 to use new compression flags and to support microMIPS.
20932 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20933 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20934 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20935 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20936 (ASM_SPEC): Support mmicromips and mno-micromips.
20937 (M16STORE_REG_P): New macro.
20938 (MIPS_CALL): Support TARGET_MICROMIPS.
20939 (MICROMIPS_J): New macro.
20940 (mips_base_mips16): Rename this...
20941 (mips_base_compression_flags): ...to this.
20942 (UMIPS_12BIT_OFFSET_P): New macro.
20943 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20944 (MULTILIB_DIRNAMES): Likewise.
20946 2013-02-25 Tom de Vries <tom@codesourcery.com>
20948 PR rtl-optimization/56131
20949 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20950 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20951 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20953 2013-02-25 Tobias Burnus <burnus@net-b.de>
20955 * doc/invoke.texi (-fsanitize=): Move from optimization
20956 to debugging options.
20958 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20960 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20962 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20963 Alexander Monakov <amonakov@ispras.ru>
20965 PR middle-end/56077
20966 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20967 flush pending lists also on non-jumps. Adjust comment.
20969 2013-02-24 Catherine Moore <clm@codesourcery.com>
20970 Maciej W. Rozycki <macro@codesourcery.com>
20971 Tom de Vries <tom@codesourcery.com>
20972 Nathan Sidwell <nathan@codesourcery.com>
20973 Iain Sandoe <iain@codesourcery.com>
20974 Nathan Froyd <froydnj@codesourcery.com>
20975 Chao-ying Fu <fu@mips.com>
20977 * doc/extend.texi: (micromips, nomicromips, nocompression):
20978 Document new function attributes.
20979 * doc/invoke.texi (minterlink-compressed, mmicromips,
20980 m14k, m14ke, m14kec): Document new options.
20981 (minterlink-mips16): Update documentation.
20982 * doc/md.texi (ZC, ZD): Document new constraints.
20983 * configure.ac (gcc_cv_as_micromips): Check if linker
20984 supports the .set micromips directive.
20985 * configure: Regenerate.
20986 * config.in: Regenerate.
20987 * config/mips/mips-tables.opt: Regenerate.
20988 * config/mips/micromips.md: New file.
20989 * constraints.md (ZC, AD): New constraints.
20990 * config/mips/predicates.md (movep_src_register): New predicate.
20991 (movep_src_operand): New predicate.
20992 (non_volatile_mem_operand): New predicate.
20993 * config/mips/mips.md (multimem): New type.
20994 (length): Differentiate between 17-bit and 18-bit branch offsets.
20995 (MOVEP1, MOVEP2): New mode iterator.
20996 (mov_<load>l): Use ZC constraint.
20997 (mov_<load>r): Likewise.
20998 (mov_<store>l): Likewise.
20999 (mov_<store>r): Likewise.
21000 (*branch_equality<mode>_inverted): Add microMIPS support.
21001 (*branch_equality<mode>): Likewise.
21002 (*jump_absolute): Likewise.
21003 (indirect_jump_<mode>): Likewise.
21004 (tablejump_<mode>): Likewise.
21005 (<optab>_internal): Likewise.
21006 (sibcall_internal): Likewise.
21007 (sibcall_value_internal): Likewise.
21008 (prefetch): Use constraint ZD.
21009 * config/mips/mips.opt (minterlink-compressed): New option.
21010 (minterlink-mips16): Now an alias for minterlink-compressed.
21011 (mmicromips): New option.
21012 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21013 (compare_and_swap_12): Likewise.
21014 (sync_add<mode>): Likewise.
21015 (sync_<optab>_12): Likewise.
21016 (sync_old_<optab>_12): Likewise.
21017 (sync_new_<optab>_12): Likewise.
21018 (sync_nand_12): Likewise.
21019 (sync_old_nand_12): Likewise.
21020 (sync_new_nand_12): Likewise.
21021 (sync_sub<mode>): Likewise.
21022 (sync_old_add<mode>): Likewise.
21023 (sync_old_sub<mode>): Likewise.
21024 (sync_new_add<mode>): Likewise.
21025 (sync_new_sub<mode>): Likewise.
21026 (sync_<optab><mode>): Likewise.
21027 (sync_old_<optab><mode>): Likewise.
21028 (sync_new_<optab><mode>): Likewise.
21029 (sync_nand<mode>): Likewise.
21030 (sync_old_nand<mode>): Likewise.
21031 (sync_new_nand<mode>): Likewise.
21032 (sync_lock_test_and_set<mode>): Likewise.
21033 (test_and_set_12): Likewise.
21034 (atomic_compare_and_swap<mode>): Likewise.
21035 (atomic_exchange<mode>_llsc): Likewise.
21036 (atomic_fetch_add<mode>_llsc): Likewise.
21037 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21038 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21039 (umips_save_restore_pattern_p): Likewise.
21040 (umips_load_store_pair_p): Likewise.
21041 (umips_output_load_store_pair): Likewise.
21042 (umips_movep_target_p): Likewise.
21043 (umips_12bit_offset_address_p): Likewise.
21044 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21045 (mips_base_mips16): Rename this...
21046 (mips_base_compression_flags): ...to this. Update all uses.
21047 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21048 (mips_mips16_decl_p): Delete.
21049 (mips_nomips16_decl_p): Delete.
21050 (mips_get_compress_on_flags): New function.
21051 (mips_get_compress_off_flags): New function.
21052 (mips_get_compress_mode): New function.
21053 (mips_get_compress_on_name): New function.
21054 (mips_get_compress_off_name): New function.
21055 (mips_insert_attributes): Support multiple compression types.
21056 (mips_merge_decl_attributes): Likewise.
21057 (umips_12bit_offset_address_p): New function.
21058 (mips_start_function_definition): Emit .set micromips directive.
21059 (mips_call_may_need_jalx_p): New function.
21060 (mips_function_ok_for_sibcall): Add microMIPS support.
21061 (mips_print_operand_punctuation): Support short delay slots and
21063 (umips_swm_mask, umips_swm_encoding): New.
21064 (umips_build_save_restore): New function.
21065 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21066 (was_mips16_p): Remove.
21067 (old_compression_mode): New.
21068 (mips_set_compression_mode): New function.
21069 (mips_set_current_function): Add microMIPS support.
21070 (mips_option_override): Likewise.
21071 (umips_save_restore_pattern_p): New function.
21072 (umips_output_save_restore): New function.
21073 (umips_load_store_pair_p_1): New function.
21074 (umips_load_store_pair_p): New function.
21075 (umips_output_load_store_pair_1): New function.
21076 (umips_output_load_store_pair): New function.
21077 (umips_movep_target_p) New function.
21078 (mips_prepare_pch_save): Add microMIPS support.
21079 * config/mips/mips.h (TARGET_COMPRESSION): New.
21080 (TARGET_CPU_CPP_BUILTINS): Update macro
21081 to use new compression flags and to support microMIPS.
21082 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21083 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21084 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21085 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21086 (ASM_SPEC): Support mmicromips and mno-micromips.
21087 (M16STORE_REG_P): New macro.
21088 (MIPS_CALL): Support TARGET_MICROMIPS.
21089 (MICROMIPS_J): New macro.
21090 (mips_base_mips16): Rename this...
21091 (mips_base_compression_flags): ...to this.
21092 (UMIPS_12BIT_OFFSET_P): New macro.
21093 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21094 (MULTILIB_DIRNAMES): Likewise.
21096 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21099 * target-globals.c (save_target_globals): For init_reg_sets and
21100 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21102 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21104 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21105 * config/aarch64/t-aarch64
21106 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21108 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21110 PR inline-asm/56148
21111 * lra-constraints.c (process_alt_operands): Reload operand
21112 conflicting with earlier clobber only if no more other conflicting
21115 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21118 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21119 if not linking a shared library.
21121 2013-02-22 Seth LaForge <sethml@google.com>
21123 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
21125 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
21127 * config/arm/arm.md (split for extendsidi): Update condition.
21128 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
21129 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
21130 (qhs_zextenddi_cstr): Likewise.
21132 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21134 PR middle-end/56420
21135 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
21136 avoid signed wrapping.
21137 (expand_mult): Handle properly multiplication by
21138 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
21139 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
21140 in the compiler if coeff is HOST_WIDE_INT_MIN.
21141 (expand_divmod): Don't make ext_op1 static, change it's type to
21142 uhwi. Avoid undefined behavior in -INTVAL (op1).
21144 PR rtl-optimization/50339
21145 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
21147 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
21148 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
21149 into splitting_ashiftrt field.
21150 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
21152 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
21155 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
21157 PR middle-end/56108
21158 * trans-mem.c (execute_tm_mark): Do not expand transactions that
21159 are sure to go irrevocable.
21161 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
21163 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
21164 scalars are valid operands.
21166 2013-02-21 Martin Jambor <mjambor@suse.cz>
21168 PR tree-optimization/56310
21169 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
21170 only matching indices and non-negative final offsets.
21171 (intersect_aggregates_with_edge): Pass src_idx to
21172 agg_replacements_to_vector. Pass src_idx insstead of index to
21173 intersect_with_agg_replacements.
21175 2013-02-21 Martin Jambor <mjambor@suse.cz>
21177 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
21178 instead of hard-wired defaults.
21180 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
21182 * doc/invoke.texi (MIPS Options): Update documentation of the
21183 floating-point multiply-accumulate instruction restrictions.
21185 2013-02-21 Kostya Serebryany <kcc@google.com>
21187 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
21188 asan_shadow_offset on x86_64 linux.
21190 2013-02-21 Richard Biener <rguenther@suse.de>
21192 PR tree-optimization/56415
21194 2013-02-11 Richard Biener <rguenther@suse.de>
21196 PR tree-optimization/56273
21197 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21200 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21203 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
21206 PR inline-asm/56405
21207 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
21208 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
21210 2013-02-20 Jan Hubicka <jh@suse.cz>
21212 PR tree-optimization/56265
21213 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
21214 when target is referenced for first time.
21216 2013-02-20 Richard Biener <rguenther@suse.de>
21218 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
21219 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
21220 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
21221 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
21222 not return anything.
21223 (rename_ssa_copies): Do not remove unused locals.
21224 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
21225 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
21226 * passes.c (execute_function_todo): Do not schedule unused locals
21227 removal if cleanup_tree_cfg did something.
21228 * tree-ssa-live.c (remove_unused_locals): Dump statistics
21229 about the number of removed locals.
21231 2013-02-20 Richard Biener <rguenther@suse.de>
21233 PR tree-optimization/56398
21234 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
21236 2013-02-20 Martin Jambor <mjambor@suse.cz>
21238 PR tree-optimization/55334
21239 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
21240 restricted pointers to arrays.
21242 2013-02-20 Richard Biener <rguenther@suse.de>
21243 Jakub Jelinek <jakub@redhat.com>
21245 PR tree-optimization/56396
21246 * tree-ssa-ccp.c (n_const_val): New static variable.
21247 (get_value): Return NULL for SSA names we don't have a lattice
21249 (ccp_initialize): Initialize n_const_val.
21250 * tree-ssa-copy.c (n_copy_of): New static variable.
21251 (init_copy_prop): Initialize n_copy_of.
21252 (get_value): Return NULL_TREE for SSA names we don't have a
21255 2013-02-20 Martin Jambor <mjambor@suse.cz>
21257 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21259 2013-02-20 Richard Biener <rguenther@suse.de>
21261 * genpreds.c (write_lookup_constraint): Do not compare first
21262 letter of the constraint again.
21264 2013-02-20 Richard Biener <rguenther@suse.de>
21266 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21268 (get_use_iv_cost): Terminate hashtable walk when coming across
21271 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
21273 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
21274 reassociation for avx2 targets.
21276 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21278 * config/microblaze/microblaze.c: microblaze_has_clz = 0
21279 Add version check for v8.10.a to enable microblaze_has_clz
21280 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
21281 version and TARGET_PATTERN_COMPARE check
21282 * config/microblaze/microblaze.md: New clzsi2 instruction
21284 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21286 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
21287 function before branching.
21289 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21291 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
21293 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
21295 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21297 PR middle-end/55889
21298 * sel-sched.c: Include ira.h.
21299 (implicit_clobber_conflict_p): New function.
21300 (moveup_expr): Use it.
21301 * Makefile.in (sel-sched.o): Depend on ira.h.
21303 2013-02-19 Richard Biener <rguenther@suse.de>
21305 PR tree-optimization/56384
21306 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
21307 (vn_hash_type): Split out from ...
21308 (vn_hash_constant_with_type): ... here.
21309 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
21310 (vn_phi_eq): Compare types from vn_phi_s structure.
21311 (vn_phi_lookup): Populate vn_phi_s type.
21312 (vn_phi_insert): Likewise.
21314 2013-02-19 Jakub Jelinek <jakub@redhat.com>
21316 PR tree-optimization/56350
21317 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
21318 if haven't found reduction or nested cycle operand, rather than
21319 asserting we must find it.
21321 PR tree-optimization/56381
21322 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
21325 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
21326 Jakub Jelinek <jakub@redhat.com>
21329 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
21330 (swap_optab_enable): Same.
21331 (init_all_optabs): Use argument instead of global.
21332 * tree.h (struct tree_optimization_option): New field target_optabs.
21333 * expr.h (init_all_optabs): Add argument to prototype.
21334 (TREE_OPTIMIZATION_OPTABS): New.
21335 (save_optabs_if_changed): Protoize.
21336 * optabs.h: Declare this_fn_optabs.
21337 * optabs.c (save_optabs_if_changed): New.
21338 Declare this_fn_optabs.
21339 (init_optabs): Add argument to init_all_optabs() call.
21340 * function.c (invoke_set_current_function_hook): Handle per
21342 * function.h (struct function): New field optabs.
21343 * config/mips/mips.c (mips_set_mips16_mode): Handle when
21344 optimization_current_node has changed.
21345 * target-globals.h (save_target_globals_default_opts): Protoize.
21346 * target-globals.c (save_target_globals_default_opts): New.
21348 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21351 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
21352 registers %fr12 and %fr12R as call used.
21355 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
21356 and HImode, require all displacements to be an integer multiple of
21358 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
21359 only allow QImode and HImode when reload is in progress and strict is
21360 true. Likewise for symbolic addresses. Use base14_operand to check
21361 displacements in REG+BASE addresses.
21363 2013-02-18 Richard Biener <rguenther@suse.de>
21365 PR tree-optimization/56366
21366 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21367 handle sign-conversion of outer-loop initial induction value.
21369 2013-02-18 Richard Biener <rguenther@suse.de>
21371 PR middle-end/56349
21372 * cfghooks.c (merge_blocks): If we merge a latch into another
21373 block adjust references to it.
21374 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
21375 (verify_loop_structure): Verify that a recorded latch is in fact
21378 2013-02-18 Richard Biener <rguenther@suse.de>
21380 PR tree-optimization/56321
21381 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
21382 order SSA name release and virtual operand unlinking.
21384 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21386 * config/microblaze/microblaze.md (save_stack_block): Define.
21387 (restore_stack_block): Likewise.
21389 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21391 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
21392 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
21393 * config/microblaze/microblaze.c (microblaze_option_override):
21394 Bail out early for PIC modes when target does not support PIC.
21396 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21398 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
21399 Replace with a microblaze version.
21400 (microblaze_trampoline_init): Adapt for microblaze.
21401 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
21404 2013-02-16 Jakub Jelinek <jakub@redhat.com>
21405 Dodji Seketeli <dodji@redhat.com>
21408 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
21409 (instrument_mem_region_access): Do not forget to always put
21410 instrumentation of the of 'base' and 'base + len' in a "if (len !=
21411 0) statement, even for cases where either 'base' or 'base + len'
21412 are not instrumented -- because they have been previously
21413 instrumented. Simplify the logic by putting all the statements
21414 instrument 'base + len' inside a sequence, and then insert that
21415 sequence right before the current insertion point. Then, to
21416 instrument 'base + len', just get an iterator on that statement.
21417 And do not forget to update the pointer to iterator the function
21418 received as argument.
21420 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
21422 PR rtl-optimization/56348
21423 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
21425 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
21427 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
21428 (clean_graph_dump_file): Pass base to start_graph_dump.
21430 2013-02-14 Richard Henderson <rth@redhat.com>
21433 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
21435 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
21437 * collect2-aix.h: Define F_LOADONLY.
21439 2013-02-14 Richard Biener <rguenther@suse.de>
21442 * varasm.c (output_constant_def_1): Get the decl representing
21443 the constant as argument.
21444 (output_constant_def): Wrap output_constant_def_1.
21445 (make_decl_rtl): Use output_constant_def_1 with the decl
21446 representing the constant.
21447 (build_constant_desc): Optionally re-use a decl already
21448 representing the constant.
21449 (tree_output_constant_def): Adjust.
21451 2013-02-14 Dodji Seketeli <dodji@redhat.com>
21454 * asan.c (instrument_builtin_call): Really put the length of the
21455 second source argument into src1_len.
21457 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21459 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
21460 argument. If it is false, don't create edge from then_bb to
21462 (insert_if_then_before_iter): Pass true to it.
21463 (build_check_stmt): Pass false to it.
21464 (transform_statements): Flush hash table only on extended basic
21465 block boundaries, rather than at the beginning of every bb.
21466 Don't flush hash table on nonfreeing_call_p calls.
21467 * tree-flow.h (nonfreeing_call_p): New prototype.
21468 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
21470 2013-02-13 David S. Miller <davem@davemloft.net>
21472 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
21474 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
21477 * ira.c (max_regno_before_ira): Move from ...
21479 (fix_reg_equiv_init): Use max_regno_before_ira instead of
21482 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21484 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
21486 2013-02-13 Richard Biener <rguenther@suse.de>
21489 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
21490 globals in MEM_REFs.
21492 2013-02-13 Richard Biener <rguenther@suse.de>
21494 * loop-init.c (loop_optimizer_init): Clear loop state when
21495 re-initializing preserved loops.
21496 * loop-unswitch.c (unswitch_single_loop): Return whether
21497 we unswitched the loop. Do not verify loop state here.
21498 (unswitch_loops): When we unswitched a loop discover new loops.
21500 2013-02-13 Kostya Serebryany <kcc@google.com>
21502 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
21504 * sanitizer.def: Rename __asan_init to __asan_init_v1.
21506 2013-02-12 Dodji Seketeli <dodji@redhat.com>
21508 Avoid instrumenting duplicated memory access in the same basic block
21509 * Makefile.in (asan.o): Add new dependency on hash-table.h
21510 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
21511 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
21512 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
21513 (free_mem_ref_resources, has_mem_ref_been_instrumented)
21514 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
21515 (get_mem_ref_of_assignment): New functions.
21516 (get_mem_refs_of_builtin_call): Extract from
21517 instrument_builtin_call and tweak a little bit to make it fit with
21519 (instrument_builtin_call): Use the new
21520 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
21521 of is_gimple_builtin_call.
21522 (instrument_derefs, instrument_mem_region_access): Insert the
21523 instrumented memory reference into the hash table.
21524 (maybe_instrument_assignment): Renamed instrument_assignment into
21525 this, and change it to advance the iterator when instrumentation
21526 actually happened and return true in that case. This makes it
21527 homogeneous with maybe_instrument_assignment, and thus give a
21528 chance to callers to be more 'regular'.
21529 (transform_statements): Clear the memory reference hash table
21530 whenever we enter a new BB, when we cross a function call, or when
21531 we are done transforming statements. Use
21532 maybe_instrument_assignment instead of instrumentation. No more
21533 need to special case maybe_instrument_assignment and advance the
21534 iterator after calling it; it's now handled just like
21535 maybe_instrument_call. Update comment.
21537 2013-02-13 Richard Biener <rguenther@suse.de>
21539 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21540 Fix loop discovery code.
21542 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21544 PR inline-asm/56148
21545 * lra-constraints.c (process_alt_operands): Match early clobber
21546 operand with itself. Check conflicts with earlyclobber only if
21547 the operand is not reloaded. Prefer to reload conflicting operand
21548 if earlyclobber and matching operands are the same.
21550 2013-02-12 Richard Biener <rguenther@suse.de>
21553 * lto-streamer-out.c (write_symbol): Do not output symbols
21554 for hard register variables.
21556 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21559 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21560 (umulsidi3_insn, mulsidi3_insn): New insns.
21562 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21564 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21565 (struct tune_params): Add vec_costs field.
21566 * config/arm/arm.c (arm_builtin_vectorization_cost)
21567 (arm_add_stmt_cost): New functions.
21568 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21569 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21570 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21571 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21572 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21573 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21574 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21576 2013-02-12 Richard Biener <rguenther@suse.de>
21579 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21580 decls again if possible.
21582 2013-02-12 Richard Biener <rguenther@suse.de>
21584 PR middle-end/56288
21585 * tree-ssa.c (verify_ssa_name): Fix check, move
21586 SSA_NAME_IN_FREE_LIST check up.
21588 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21589 Steven Bosscher <steven@gcc.gnu.org>
21591 PR rtl-optimization/56151
21592 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21593 equal to op0 or op1, and last_insn pattern is CODE operation
21594 with MEM dest and one of the operands matches that MEM.
21596 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21598 * doc/extend.texi: Document Function Multiversioning and "default"
21599 parameter string to target attribute.
21600 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21601 target attribute parameter is "default".
21602 (ix86_compare_version_priority): Remove checks for target attribute.
21603 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21604 Remove check for target attribute equal to NULL. Add assert.
21605 (ix86_generate_version_dispatcher_body): Change error to sorry.
21607 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21608 Jack Howarth <howarth@bromo.med.uc.edu>
21609 Patrick Marlier <patrick.marlier@gmail.com>
21612 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21613 define ENDFILE_SPEC as TM_DESTRUCTOR.
21614 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21616 2013-02-11 Alexander Potapenko <glider@google.com>
21617 Jack Howarth <howarth@bromo.med.uc.edu>
21618 Jakub Jelinek <jakub@redhat.com>
21621 * config/darwin.c (cdtor_record): Rename ctor_record.
21622 (sort_cdtor_records): Rename sort_ctor_records.
21623 (finalize_dtors): New routine to sort destructors by
21624 priority before use in assemble_integer.
21625 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21627 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21629 PR rtl-optimization/56275
21630 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21631 offset is non-negative and less than cmode size before
21632 calling simplify_subreg.
21634 2013-02-11 Richard Biener <rguenther@suse.de>
21636 PR tree-optimization/56264
21637 * cfgloop.h (fix_loop_structure): Adjust prototype.
21638 * loop-init.c (fix_loop_structure): Return the number of
21639 newly discovered loops.
21640 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21641 are discovered, do a full loop-closed SSA rewrite.
21643 2013-02-11 Richard Biener <rguenther@suse.de>
21645 PR tree-optimization/56273
21646 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21648 (check_array_ref): Fix missing newline in dumps.
21649 (search_for_addr_array): Likewise.
21651 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21653 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21655 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21658 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21660 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21662 PR rtl-optimization/56246
21663 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21665 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21666 constraints are satisfied.
21668 2013-02-08 Jeff Law <law@redhat.com>
21671 * emit-rtl.c (reg_is_parm_p): New function.
21672 * regs.h (reg_is_parm_p): New prototype.
21673 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21674 callee-clobbered registers.
21676 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21679 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21680 If there is no implicit builtin declaration, just return NULL.
21682 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21684 * config/i386/sse.md (FMAMODEM): New mode iterator.
21685 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21686 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21688 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21690 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21691 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21692 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21694 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21696 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21697 (microblaze*-*-elf): Likewise.
21698 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21700 * config/microblaze/microblaze-c.c: Add builtin defines for
21701 _LITTLE_ENDIAN and _BIG_ENDIAN.
21702 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21703 add to TARGET_DEFAULT flags.
21704 Expand ASM_SPEC and LINK_SPEC.
21705 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21706 * config/microblaze/microblaze.md: Update extendsidi2 and
21707 movdi_internal instructions to use low-order / high-order reg
21709 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21710 options and inversemask / mask of LITTLE_ENDIAN.
21711 * config/microblaze/t-microblaze: Expand multilib options to
21712 include mlittle-endian (le) and update exceptions patterns.
21714 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21716 PR rtl-optimization/56195
21717 * lra-constraints.c (get_reload_reg): Don't reuse regs
21718 if they have smaller mode than requested, if they have
21719 wider mode than requested, try to return a SUBREG.
21721 PR tree-optimization/56250
21722 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21723 if type is unsigned and code isn't MULT_EXPR.
21725 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21727 PR tree-optimization/56064
21728 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21729 bits according to mode.
21730 * fixed-value.h (fixed_from_double_int)
21731 (const_fixed_from_double_int): Adjust comments.
21733 2013-02-08 Richard Biener <rguenther@suse.de>
21736 * lto-streamer.h (struct data_in): Remove current_file, current_line
21737 and current_col members.
21738 * lto-streamer-out.c (lto_output_location): Stream changed bits
21739 en-block for efficiency.
21740 * lto-streamer-in.c (clear_line_info): Remove.
21741 (lto_input_location): Cache current file, line and column
21742 globally via local statics. Read changed bits en-block.
21743 (input_function): Do not call clear_line_info.
21744 (lto_read_body): Likewise.
21745 (lto_input_toplevel_asms): Likewise.
21747 2013-02-08 Michael Matz <matz@suse.de>
21749 PR tree-optimization/52448
21750 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21751 (nt_call_phase): New static.
21752 (add_or_mark_expr): Only mark accesses with newer phase than any
21754 (nonfreeing_call_p): New.
21755 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21756 (nt_fini_block): Keep blocks marked as visited.
21757 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21759 2013-02-08 Richard Biener <rguenther@suse.de>
21761 * ira.c (ira): Free broken dominator information.
21763 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21765 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21767 2013-02-08 Marek Polacek <polacek@redhat.com>
21769 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21771 2013-02-08 Richard Biener <rguenther@suse.de>
21773 PR middle-end/56181
21774 * cfgloop.h (flow_loops_find): Adjust.
21775 (bb_loop_header_p): Declare.
21776 * cfgloop.c (bb_loop_header_p): New function split out from ...
21777 (flow_loops_find): ... here. Adjust function signature,
21778 support incremental loop structure update.
21779 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21780 * cfgloopmanip.c (fix_loop_structure): Move ...
21781 * loop-init.c (fix_loop_structure): ... here.
21782 (apply_loop_flags): Split out from ...
21783 (loop_optimizer_init): ... here.
21784 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21785 in incremental mode, only remove dead loops here.
21787 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21790 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21791 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21792 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21793 (*round<mode>3.libgcc): New insns for fixed-modes.
21794 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21795 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21796 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21797 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21798 implementations. Define to __builtin_avr_absFX,
21799 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21800 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21801 __builtin_avr_countlsFX, respectively.
21802 * config/avr/avr-c.c (target.h): Include it.
21803 (enum avr_builtin_id): New enum.
21804 (avr_resolve_overloaded_builtin): New static function.
21805 (avr_register_target_pragmas): Use it to set
21806 targetm.resolve_overloaded_builtin.
21807 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21808 tree nodes used by DEF_BUILTIN.
21809 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21810 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21811 <AVR_BUILTIN_xxBITS>: Same.
21813 2013-02-08 Richard Biener <rguenther@suse.de>
21815 * cfgloop.c (verify_loop_structure): Properly handle
21816 a loop exiting to another loop header.
21817 * ira-int.h (ira_loops): Remove.
21818 * ira.c (ira_loops): Remove.
21819 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21820 (do_reload): Use loop_optimizer_finalize.
21821 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21822 number_of_loops to access the loop tree.
21823 (more_one_region_p): Likewise.
21824 (finish_loop_tree_nodes): Likewise.
21825 (rebuild_regno_allocno_maps): Likewise.
21826 (mark_loops_for_removal): Likewise.
21827 (mark_all_loops_for_removal): Likewise.
21828 (remove_unnecessary_regions): Likewise.
21829 (ira_build): Likewise.
21830 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21832 2013-02-08 Richard Biener <rguenther@suse.de>
21834 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21835 * ipa-pure-const.c (analyze_function): Avoid calling
21836 mark_irreducible_loops twice.
21837 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21839 2013-02-07 David S. Miller <davem@davemloft.net>
21841 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21843 * var-tracking.c (vt_add_function_parameter): Test the presence of
21844 HAVE_window_save properly and do not remap argument registers when
21845 we have a leaf function.
21847 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21850 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21852 * config/i386/i386.c (ix86_print_operand): Ditto.
21854 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21856 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21858 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21860 PR rtl-optimization/56225
21861 * lra-constraints.c (process_alt_operands): Check that reload hard
21862 reg can hold value for strict_low_part.
21864 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21867 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21868 dwarf2out_end_function.
21869 (in_first_function_p, maybe_at_text_label_p,
21870 first_loclabel_num_not_at_text_label): New variables.
21871 (dwarf2out_var_location): In the first function find out
21872 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21873 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21876 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21878 PR rtl-optimization/56178
21879 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21880 SUBREG of a register. Tidy up related block of code.
21881 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21882 note if the source is a register or a SUBREG of a register.
21884 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21887 * config/rs6000/rs6000.md (ptrm): New mode attr.
21888 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21889 call_value_indirect_aix<pttrsize>,
21890 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21893 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21895 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21896 if -bnortl. Convert to strcmp and strncmp.
21898 2013-02-07 Alan Modra <amodra@gmail.com>
21901 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21902 addresses won't wrap when offsetting.
21903 (rs6000_secondary_reload): Provide secondary reloads needed for
21904 wrapping LO_SUM addresses.
21906 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21908 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21909 MACH, just __MACH__.
21911 2013-02-06 Richard Biener <rguenther@suse.de>
21913 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21914 instead of calling fix_loop_structure.
21916 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21918 PR middle-end/56217
21919 * omp-low.c (use_pointer_for_field): Return false if
21920 lower_send_shared_vars doesn't generate any copy-out code.
21922 2013-02-06 Tom de Vries <tom@codesourcery.com>
21924 PR rtl-optimization/56131
21925 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21926 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21927 of the label is NULL. Add comment.
21929 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21931 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21934 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21935 (STATIC_LIBTSAN_LIBS): Likewise.
21936 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21937 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21938 is defined, don't add anything else beyond that.
21939 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21940 (LINK_COMMAND_SPEC): Use them.
21942 PR tree-optimization/56205
21943 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21944 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21945 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21947 2013-02-05 Richard Biener <rguenther@suse.de>
21949 PR tree-optimization/53342
21950 PR tree-optimization/53185
21951 * tree-vectorizer.h (vect_check_strided_load): Remove.
21952 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21953 not disallow peeling for vectorized strided loads.
21954 (vect_check_strided_load): Make static and simplify.
21955 (vect_analyze_data_refs): Adjust.
21956 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21957 correctly when vectorizing strided loads.
21959 2013-02-05 Richard Biener <rguenther@suse.de>
21961 * doc/install.texi: Refer to ISL, not PPL.
21963 2013-02-05 Jan Hubicka <jh@suse.cz>
21965 PR tree-optimization/55789
21966 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21968 2013-02-05 Jan Hubicka <jh@suse.cz>
21970 PR tree-optimization/55789
21971 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21972 the dead call anyway.
21974 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21977 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21979 2013-02-04 Alexander Potapenko <glider@google.com>
21980 Jack Howarth <howarth@bromo.med.uc.edu>
21981 Jakub Jelinek <jakub@redhat.com>
21984 * config/darwin.c (sort_ctor_records): Stabilized qsort
21985 on constructor priority by using original position.
21986 (finalize_ctors): New routine to sort constructors by
21987 priority before use in assemble_integer.
21988 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21990 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21993 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21994 about visibility on artificial decls.
21995 * config/sol2.c (solaris_assemble_visibility): Likewise.
21997 2013-02-04 Kai Tietz <ktietz@redhat.com>
22000 * config/i386/i386.c (function_value_ms_64): Add additional valtype
22001 argument and improve checking of return-argument types for 16-byte
22003 (ix86_function_value_1): Add additional valtype argument on call
22004 of function_value_64.
22005 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
22006 handling infunction_value_64 function.
22008 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
22010 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
22012 2013-02-04 Richard Biener <rguenther@suse.de>
22014 PR tree-optimization/56188
22015 * tree-ssa-structalias.c (label_visit): Consider case with
22016 initially non-empty points-to set.
22017 (perform_var_substitution): Dump node mapping and clean up.
22019 2013-02-04 Richard Guenther <rguenther@suse.de>
22022 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22023 node prevail as last resort.
22024 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22025 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22027 2013-02-04 Richard Biener <rguenther@suse.de>
22029 PR tree-optimization/56113
22030 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22032 (equiv_class_lookup_or_add): ... this.
22033 (label_visit): Adjust and fix error in previous patch.
22034 (perform_var_substitution): Adjust.
22036 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22038 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22039 * config/sh/sh4-300.md: Likewise.
22040 * config/sh/sh4a.md: Likewise.
22041 * config/sh/constraints.md: Likewise.
22042 * config/sh/sh.md: Likewise.
22043 * config/sh/netbsd-elf.h: Likewise.
22044 * config/sh/predicates.md: Likewise.
22045 * config/sh/sh-protos.h: Likewise.
22046 * config/sh/ushmedia.h: Likewise.
22047 * config/sh/linux.h: Likewise.
22048 * config/sh/sh.c: Likewise.
22049 * config/sh/superh.h: Likewise.
22050 * config/sh/elf.h: Likewise.
22051 * config/sh/sh4.md: Likewise.
22052 * config/sh/sh.h: Likewise.
22054 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22056 * config/pa/constraints.md: Adjust unused letters. Change "T"
22057 constraint to match_test floating_point_store_memory_operand().
22058 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22059 (base14_operand): New.
22060 (floating_point_store_memory_operand): New.
22061 (integer_store_memory_operand): Revise to use base14_operand and
22062 reg_plus_base_memory_operand.
22063 (move_dest_operand): Allow symbolic_memory_operands.
22064 (symbolic_memory_operand): Check for LO_SOM.
22065 (symbolic_operand): Change default case to break.
22066 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22067 CONST_DOUBLE values to be reloaded by putting them into memory when
22068 the destination is a floating point register.
22069 (movdf): Remove code to handle CONST_DOUBLE.
22071 (reload_indf_r1): New.
22072 (reload_insf_r1): New.
22073 Consistently use "Q" and "T" constraints with integer and floating
22074 point move instructions, respectively.
22075 (movdi): Remove FAIL.
22076 Change predicate for source operand unamed DImode move from
22077 general_operand to move_src_operand.
22078 (umulsidi3): Change predicate for destination operand to
22080 Likewise for similar unamed patterns.
22081 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22082 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22083 (hppa_legitimize_address): Simplify mask calculation.
22084 (pa_emit_move_sequence): Revised handling of secondary reloads from
22085 REG+D addresses for floating point loads and stores. Directly handle
22086 loading CONST0_RTX (mode) to a floating point register.
22087 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22088 to floating point registers. Don't restrict secondary reloads to
22089 floating point registers to integer modes. Revise some comments and
22091 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22092 (pa_legitimate_address_p): New.
22093 (pa_legitimize_reload_address): New.
22094 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22095 (STRICT_REG_OK_FOR_BASE_P): New.
22096 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22097 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22099 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22100 Andrew Dixie <andrewd@gentrack.com>
22102 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22105 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22107 * expmed.c (extract_bit_field_1): Pass the full width of the
22108 structure to get_best_reg_extraction_insn.
22110 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22113 * configure.ac (use_cxa_atexit): Add AIX.
22114 * configure: Regenerate.
22116 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22118 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22121 * final.c (need_profile_function): New variable.
22122 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22123 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
22124 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
22125 notes, targetm.asm_out.function_prologue doesn't emit anything,
22126 HAVE_prologue and profiler should be emitted before prologue,
22127 set need_profile_function instead of emitting it.
22128 (final_scan_insn): If need_profile_function, emit
22129 profile_function on the first NOTE_INSN_BASIC_BLOCK or
22130 NOTE_INSN_FUNCTION_BEG note.
22132 2013-02-01 Richard Henderson <rth@redhat.com>
22134 * config/rs6000/rs6000.md (smulditi3): New.
22137 * config/alpha/alpha.md (umulditi3): New.
22139 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22141 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
22142 (ASM_OUTPUT_ALIGNED_LOCAL): New.
22144 2013-02-01 Richard Biener <rguenther@suse.de>
22146 PR tree-optimization/56113
22147 * tree-ssa-structalias.c (label_visit): Reduce work for
22148 single-predecessor nodes.
22150 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
22152 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
22153 range isn't testing for zero.
22155 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
22157 PR middle-end/56113
22158 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
22160 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
22161 Nick Clifton <nickc@redhat.com>
22163 * config/v850/constraints.md (Q): Define as a memory constraint.
22164 * config/v850/predicates.md (label_ref_operand): New predicate.
22165 (e3v5_shift_operand): New predicate.
22166 (ior_operator): New predicate.
22167 * config/v850/t-v850: Add e3v5 multilib.
22168 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
22169 (v850_gen_movdi): Prototype.
22170 * config/v850/v850.c: Add support for e3v5 architecture.
22171 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
22173 (construct_save_jarl): Add e3v5 long JARL support.
22174 (v850_adjust_insn_length): New function. Adjust length of call
22175 insns when using e3v5 instructions.
22176 (v850_gen_movdi): New function: Generate instructions to move a
22178 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
22179 (CPP_SPEC): Define __v850e3v5__ as appropriate.
22180 (TARGET_USE_FPU): Enable for e3v5.
22181 (CONST_OK_FOR_W): New macro.
22182 (ADJUST_INSN_LENGTH): Define.
22183 * config/v850/v850.md (UNSPEC_LOOP): Define.
22184 (attr cpu): Add v850e3v5.
22185 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
22186 (movdi): New pattern.
22187 (movdi_internal): New pattern.
22188 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
22189 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
22190 (cstoresf4): Likewise.
22191 (cstoredf4): Likewise.
22192 (insv): New pattern.
22193 (rotlso3_a): New pattern.
22194 (rotlsi3_b): New pattern
22195 (rotlsi3_v850e3v5): New pattern.
22196 (doloop_begin): New pattern.
22197 (fix_loop_counter): New pattern.
22198 (doloop_end): New pattern.
22199 (branch_normal): Add e3v5 long branch support.
22200 (branch_invert): Likewise.
22201 (branch_z_normal): Likewise.
22202 (branch_z_invert): Likewise.
22203 (branch_nz_normal): Likewise.
22204 (branch_nz_invert): Likewise.
22205 (call_internal_short): Add e3v5 register-indirect JARL support.
22206 (call_internal_long): Likewise.
22207 (call_value_internal_short): Likewise.
22208 (call_value_internal_long): Likewise.
22209 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
22210 (mloop): New option.
22211 * config.gcc: Add support for configuring v840e3v5 target.
22212 * doc/invoke.texi: Document new v850 specific command line options.
22214 2013-01-31 Paul Koning <ni1d@arrl.net>
22218 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
22219 children if parent is a class.
22220 (prune_unused_types_prune): Don't add DW_AT_declaration.
22222 2013-01-31 Richard Biener <rguenther@suse.de>
22224 PR tree-optimization/56157
22225 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
22226 match up operand with SLP child.
22228 2013-01-31 Jason Merrill <jason@redhat.com>
22231 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
22232 parameters the first time.
22233 (gen_scheduled_generic_parms_dies): Check completeness here.
22235 2013-01-31 Richard Biener <rguenther@suse.de>
22237 PR middle-end/53073
22238 * common.opt (faggressive-loop-optimizations): New flag,
22239 enabled by default.
22240 * doc/invoke.texi (faggressive-loop-optimizations): Document.
22241 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
22242 infer_loop_bounds_from_undefined by it.
22244 2013-01-31 Richard Biener <rguenther@suse.de>
22246 PR tree-optimization/56150
22247 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
22248 visit virtual operands.
22249 (find_uses_to_rename_bb): Likewise.
22251 2013-01-31 Richard Biener <rguenther@suse.de>
22253 PR tree-optimization/56150
22254 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22255 mixed store non-store stmts.
22257 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22260 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22261 LIBASAN_EARLY_SPEC is defined.
22262 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22263 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22265 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22268 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
22269 invalid args instead of ICEing on it.
22270 (ix86_valid_target_attribute_tree): Return error_mark_node if
22271 ix86_valid_target_attribute_inner_p failed.
22272 (ix86_valid_target_attribute_p): Return false only if
22273 ix86_valid_target_attribute_tree returned error_mark_node. Allow
22274 target("default") attribute.
22275 (sorted_attr_string): Change argument from const char * to tree,
22276 merge in all target attribute arguments rather than just one.
22277 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
22278 instead of free. Avoid using strcat.
22279 (ix86_mangle_function_version_assembler_name): Mangle
22280 target("default") as if no target attribute is present. Adjust
22281 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
22282 instead of xmalloc and XDELETEVEC instead of free.
22283 (ix86_function_versions): Don't return true if one of the decls
22284 doesn't have target attribute. If they don't and one of the decls
22285 is DECL_FUNCTION_VERSIONED, report an error. Adjust
22286 sorted_attr_string caller. Use XDELETEVEC instead of free.
22287 (ix86_supports_function_versions): Remove.
22288 (make_name): Fix up formatting.
22289 (make_dispatcher_decl): Remove resolver_name and its initialization.
22290 Avoid leaking memory.
22291 (is_function_default_version): Return true if there is
22292 target("default") attribute rather than no target attribute at all.
22293 (make_resolver_func): Avoid leaking memory.
22294 (ix86_generate_version_dispatcher_body): Likewise.
22295 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
22296 * target.def (supports_function_versions): Remove.
22297 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
22298 * doc/tm.texi: Regenerated.
22300 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
22302 PR rtl-optimization/56144
22303 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
22304 for values with side effects.
22306 2013-01-30 Richard Biener <rguenther@suse.de>
22308 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
22309 (sparseset_pop): Likewise.
22310 * cfganal.c (compute_idf): Likewise. Increase work-stack size
22311 to be able to use quick_push in the worker loop.
22313 2013-01-30 Marek Polacek <polacek@redhat.com>
22315 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
22317 2013-01-30 Richard Biener <rguenther@suse.de>
22320 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
22322 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
22324 PR tree-optimization/56064
22325 * fixed-value.c (fixed_from_double_int): New function.
22326 * fixed-value.h (fixed_from_double_int): New prototype.
22327 (const_fixed_from_double_int): New static inline function.
22328 * fold-const.c (native_interpret_fixed): New static function.
22329 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
22330 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
22331 (native_encode_fixed): New static function.
22332 (native_encode_expr) <FIXED_CST>: Use it.
22333 (native_interpret_int): Move double_int worker code to...
22334 * double-int.c (double_int::from_buffer): ...this new static method.
22335 * double-int.h (double_int::from_buffer): Prototype it.
22337 2013-01-30 Richard Biener <rguenther@suse.de>
22339 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
22340 New pointer-map and obstack.
22341 (init_alias_vars): Allocate pointer-map and obstack.
22342 (delete_points_to_sets): Free them.
22343 (find_what_var_points_to): Cache result.
22344 (find_what_p_points_to): Adjust for changed interface of
22345 find_what_var_points_to.
22346 (compute_points_to_sets): Likewise.
22347 (ipa_pta_execute): Likewise.
22349 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22351 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
22352 * configure: Regenerate.
22353 * config.in: Regenerate.
22354 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
22355 #nobits/#progbits if supported.
22357 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
22360 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
22361 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
22362 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
22364 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22366 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
22367 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
22369 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22371 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
22372 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
22374 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22376 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
22378 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
22379 * config/arm/cortex-a7.md: New bypasses using
22380 arm_mac_accumulator_is_result.
22382 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22384 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
22385 (cortex_a7_neon_mla): Likewise.
22386 (cortex_a7_fpfmad): New reservation.
22387 (cortex_a7_fpmacs): Use ffmas and update required units.
22388 (cortex_a7_fpmuld): Update required units and latency.
22389 (cortex_a7_fpmacd): Likewise.
22390 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
22391 (cortex_a7_neon). Likewise.
22392 (bypass) Update participating units.
22394 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22396 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
22397 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
22399 * config/arm/vfp11.md (vfp_farith): Use ffmas.
22400 (vfp_fmul): Use ffmad.
22401 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
22402 (cortex_r4_fmacd): Use ffmad.
22403 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
22404 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
22405 (cortex_a9_fmacd): Use ffmad.
22406 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
22407 (cortex_a8_vfp_macd): Use ffmad.
22408 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
22409 (cortex_a5_fpmacd): Use ffmad.
22410 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
22411 (cortex_a15_vfp_macd): Use ffmad.
22412 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
22414 2013-01-29 Jason Merrill <jason@redhat.com>
22417 * varasm.c (default_assemble_visibility): Don't warn about
22418 visibility on artificial decls.
22420 2013-01-29 Richard Biener <rguenther@suse.de>
22422 PR tree-optimization/56113
22423 * tree-ssa-structalias.c (equiv_class_lookup): Also return
22425 (label_visit): Free duplicate bitmaps and record the leader instead.
22426 (perform_var_substitution): Adjust.
22428 2013-01-29 Richard Biener <rguenther@suse.de>
22430 PR tree-optimization/55270
22431 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
22432 the CFG, schedule loops for fixup.
22434 2013-01-29 Nick Clifton <nickc@redhat.com>
22436 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
22439 2013-01-28 Leif Ekblad <leif@rdos.net>
22441 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
22442 * config/i386/i386.h (TARGET_RDOS): New macro.
22443 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
22444 * config/i386/i386.c (ix86_option_override_internal): For 64bit
22445 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
22446 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
22447 DEFAULT_LARGE_SECTION_THRESHOLD.
22448 * config/i386/i386.md (R14_REG, R15_REG): New constants.
22449 * config/i386/rdos.h: New file.
22450 * config/i386/rdos64.h: New file.
22452 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
22455 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
22458 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22460 PR rtl-optimization/56117
22461 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
22462 call cselib_lookup_from_insn on the MEM before calling
22463 add_insn_mem_dependence.
22465 2013-01-28 Richard Biener <rguenther@suse.de>
22467 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
22468 to a stmt that didn't have one.
22469 (copy_phis_for_bb): Likewise for PHI arguments.
22470 (copy_debug_stmt): Likewise for debug stmts.
22472 2013-01-28 Richard Biener <rguenther@suse.de>
22474 PR tree-optimization/56034
22475 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
22476 (partition_builtin_p): Adjust.
22477 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
22478 it is the last partition.
22479 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
22480 up the vertex for the definition.
22481 (classify_partition): Classify whether a partition is a
22482 PKIND_REDUCTION, thus has uses outside of the loop.
22483 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
22484 Merge all PKIND_REDUCTION partitions into the last partition.
22485 (tree_loop_distribution): Seed partitions from reductions as well.
22487 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22489 PR tree-optimization/56125
22490 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
22491 pow(x,c) into sqrt(x) * powi(x, n/2) or
22492 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
22493 optimizing for size.
22494 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
22495 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
22498 PR tree-optimization/56094
22499 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
22500 to UNKNOWN_LOCATION while gimplifying expr.
22502 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
22505 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
22506 operand 0 in movabs insn template for -masm=intel asm alternative.
22507 (*movabs<mode>_2): Ditto for operand 1.
22509 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
22512 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
22515 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22517 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
22520 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22522 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
22523 Undef to avoid warning.
22525 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22527 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
22528 * configure: Regenerate.
22530 2013-01-25 Jakub Jelinek <jakub@redhat.com>
22532 PR tree-optimization/56098
22533 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
22534 for stmts with volatile ops.
22535 (cond_store_replacement): Don't optimize if assign has volatile ops.
22536 (cond_if_else_store_replacement_1): Don't optimize if either
22537 then_assign or else_assign have volatile ops.
22538 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22541 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22543 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22545 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22547 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22548 missing ':' in asm example.
22550 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22552 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22553 entries into lane and laneq entries.
22554 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22555 Remove AdvSIMD scalar modes.
22556 (aarch64_sq<r>dmulh_laneq<mode>): New.
22557 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22559 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22560 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22561 * config/aarch64/iterators.md (VCOND): New.
22564 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22567 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22568 Add NULL LIBNAME argument to existing definitions.
22569 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22570 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22571 * config/avr/avr.c (DEF_BUILTIN): Same.
22572 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22573 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22574 implementation is available (DECL_ASSEMBLER_NAME is set).
22575 (avr_fold_absfx): New static function.
22576 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22577 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22578 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22579 AVR_BUILTIN_ABSLLK.
22580 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22581 (abshk, absk, abslk, absllk): Provide as static inline functions.
22583 2013-01-25 Marek Polacek <polacek@redhat.com>
22585 PR tree-optimization/56035
22586 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22588 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22590 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22591 (*movtf_internal_rex64): Add (!o,C) alternative
22592 (*movxf_internal_rex64): Ditto.
22593 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22595 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22597 * doc/invoke.texi: fix typo.
22598 * doc/objc.texi: fix typo.
22600 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22602 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22603 for the first two alternatives.
22605 2013-01-24 Diego Novillo <dnovillo@google.com>
22607 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22608 (ggc-zone.o): Remove.
22609 * configure.ac: Remove option --with-gc.
22610 * configure: Re-generate.
22611 * doc/install.texi: Remove documentation for --with-gc.
22612 * gengtype.c (write_enum_defn): Remove. Update all users.
22613 (write_Types_process_field): Remove generation of gt_e_* argument.
22614 (output_type_enum): Remove. Update all users.
22615 (write_enum_defn): Remove. Update all users.
22616 (enum alloc_zone): Remove. Update all users.
22617 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22618 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22619 Update all callers.
22620 (struct ptr_data): Remove field TYPE. Update all users.
22621 (gt_pch_note_object): Remove argument TYPE. Update all users.
22622 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22624 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22625 (struct alloc_zone): Remove.
22626 (ggc_internal_alloc_zone_stat): Remove.
22627 (ggc_internal_cleared_alloc_zone_stat): Remove.
22628 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22629 (ggc_pch_count_object): Remove last argument. Update all users.
22630 (ggc_pch_alloc_object): Remove last argument. Update all users.
22631 (struct alloc_zone): Remove.
22632 * ggc-zone.c: Remove.
22633 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22634 (struct alloc_zone): Remove.
22635 (ggc_alloc_typed_stat): Remove.
22636 (ggc_alloc_typed): Remove.
22637 (ggc_splay_alloc): Remove first argument.
22638 (rtl_zone): Remove. Update all users.
22639 (tree_zone): Remove. Update all users.
22640 (tree_id_zone): Remove. Update all users.
22641 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22642 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22643 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22644 * tree-ssanames.c: Remove references to zone allocator in comments.
22646 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22648 * config/avr/avr.c (avr_out_fract): Make register numbers that
22649 might be outside of source operand signed.
22651 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22653 * config/i386/constraints.md (Yf): New constraint.
22654 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22655 of f constraint to conditionaly disable x87 register preferences.
22656 (*movdf_internal): Ditto.
22657 (*movsf_internal): Ditto.
22659 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22661 PR inline-asm/55934
22662 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22663 that have operands with impossible constraints.
22664 Add a FIXME for a speed-up opportunity.
22665 * lra-constraints.c (process_alt_operands): Verify that a class
22666 selected from constraints on asms is valid for the operand mode.
22667 (curr_insn_transform): Remove incorrect comment.
22669 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22671 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22672 TOC operand is a valid symbol ref in the constant pool.
22674 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22676 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22678 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22681 * config/avr/stdfix.h: New file.
22682 * t-avr (stdfix-gcc.h): New rule to build it.
22683 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22685 2013-01-23 Kostya Serebryany <kcc@google.com>
22687 * config/darwin.h: remove dependency on
22688 CoreFoundation (asan on Mac OS).
22690 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22693 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22694 instead of cmpdi_operand for first comparison operand.
22695 Don't assert that comparison operands aren't both constants.
22697 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22699 * doc/install.texi (Downloading the Source): Update references to
22700 downloading separate components.
22702 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22704 * doc/extend.texi (__int128): Improve grammar.
22706 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22709 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22710 alternative to (o,r).
22711 (*movdi_internal_rex64): Remove (!o,n) alternative.
22712 (DImode immediate->memory splitter): Remove.
22713 (DImode immediate->memory peephole2): Remove.
22714 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22715 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22716 alternative to (!o,*r).
22717 (*movtf_internal_sse): New pattern.
22718 (*movxf_internal_rex64): New pattern.
22719 (*movxf_internal): Disable for TARGET_64BIT.
22720 (*movdf_internal_rex64): Remove (!o,F) alternative.
22722 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22724 PR middle-end/56074
22725 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22726 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22727 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22728 stmt locations where LOCATION_LOCUS of the stmt location is
22729 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22732 * config/i386/i386.md (UNSPEC_STOS): New.
22733 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22734 *strsetqi_1): Add UNSPEC_STOS.
22736 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22739 * doc/invoke.texi: Remove left over -Wsynth example.
22741 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22743 PR tree-optimization/56051
22744 * fold-const.c (fold_binary_loc): Don't fold
22745 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22746 a narrowing conversion, or widening conversion from signed
22749 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22751 PR rtl-optimization/56023
22752 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22753 dependent on debug instruction.
22755 2013-01-21 Martin Jambor <mjambor@suse.cz>
22757 PR middle-end/56022
22758 * function.c (allocate_struct_function): Call
22759 invoke_set_current_function_hook earlier.
22761 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22763 * reload1.c (init_reload): Only initialize reload_obstack
22764 during the first call.
22766 2013-01-21 Marek Polacek <polacek@redhat.com>
22768 * cfgloop.c (verify_loop_structure): Fix up grammar.
22770 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22772 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22773 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22775 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22778 * config/arm/marvell-pj4.md: Update copyright year.
22779 Fix up use of alu to alu_reg and simple_alu_imm.
22781 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22783 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22785 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22788 * lra-constraints.c (curr_insn_transform): Don't reuse original
22789 insn for secondary memory move when memory mode should be different.
22791 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22793 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22794 atomic_storedi_1): New patterns.
22796 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22798 btver2 pipeline descriptions.
22799 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22801 * config/i386/i386.md (btver2_decode): New type attributes.
22802 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22804 * config/i386/btver2.md: New file describing btver2 pipelines.
22806 2013-01-19 Andrew Pinski <apinski@cavium.com>
22808 PR tree-optimization/52631
22809 * tree-ssa-sccvn (visit_use): Before looking up the original
22810 statement, try looking up the simplified expression.
22812 2013-01-19 Anthony Green <green@moxielogic.com>
22814 * config/moxie/moxie.c (moxie_expand_prologue): Set
22815 current_function_static_stack_size.
22817 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22819 PR tree-optimization/56029
22820 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22821 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22823 2013-01-18 Sharad Singhai <singhai@google.com>
22825 PR tree-optimization/55995
22826 * dumpfile.c (dump_loc): Print location only if available.
22827 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22829 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22832 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22833 secondary memory move.
22834 (inherit_reload_reg): Use rclass instead of cl for
22835 check_secondary_memory_needed_p.
22837 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22839 PR middle-end/56015
22840 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22841 the case where writing real complex part of target modifies op1.
22843 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22845 * config/aarch64/aarch64-simd.md
22846 (aarch64_vcond_internal<mode>): Handle unordered cases.
22847 * config/aarch64/iterators.md (v_cmp_result): New.
22849 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22850 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22852 * config/arm/marvell-pj4.md: New file.
22853 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22854 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22855 (generic_vfp): Likewise.
22856 * config/arm/arm-cores.def: Add marvell-pj4.
22857 * config/arm/arm-tune.md: Regenerate.
22858 * config/arm/arm-tables.opt: Regenerate.
22859 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22860 * doc/invoke.texi: Document marvell-pj4.
22862 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22864 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22866 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22871 * var-tracking.c (negative_power_of_two_p): New.
22872 (global_get_addr_cache, local_get_addr_cache): New.
22873 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22874 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22876 (vt_stack_offset_p): Remove.
22877 (vt_canon_true_dep): Always canonicalize loc's address.
22878 (clobber_overlapping_mems): Make sure we have a MEM.
22879 (local_get_addr_clear_given_value): New.
22880 (val_reset): Clear local cached entries.
22881 (compute_bb_dataflow): Create and release the local cache.
22882 Disable duplicate MEMs clobbering.
22883 (emit_notes_in_bb): Clobber MEMs likewise.
22884 (vt_emit_notes): Create and release the local cache.
22885 (vt_initialize, vt_finalize): Create and release the global
22886 cache, respectively.
22887 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22889 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22891 PR libmudflap/53359
22892 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22893 not found in the symtab.
22895 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22898 PR rtl-optimization/55547
22899 PR rtl-optimization/53827
22902 * alias.c (offset_overlap_p): New, factored out of...
22903 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22904 the conservative special case for symbolic constants. Don't
22905 adjust zero sizes on alignment.
22907 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22909 PR rtl-optimization/52573
22910 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22911 REG_UNUSED for the same register.
22913 2013-01-17 Richard Biener <rguenther@suse.de>
22914 Marek Polacek <polacek@redhat.com>
22916 PR rtl-optimization/55833
22917 * loop-unswitch.c (unswitch_loops): Move loop verification...
22918 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22919 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22920 Set it to true when we're removing a loop from hierarchy tree in
22921 an irreducible region.
22922 (fix_bb_placements): Adjust caller.
22923 (fix_loop_placements): Likewise.
22925 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22927 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22928 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22929 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22930 Remove ID. Adjust comments.
22931 * config/avr/avr-c.c (avr_builtin_name): Remove.
22932 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22933 * config/avr/avr.c (avr_tolower): New static function.
22934 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22935 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22936 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22939 2013-01-17 Jan Hubicka <jh@suse.cz>
22941 PR tree-optimization/55273
22942 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22944 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22947 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22948 store through atomic_store<mode>_1.
22949 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22951 2013-01-17 Martin Jambor <mjambor@suse.cz>
22953 PR tree-optimizations/55264
22954 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22955 for virtual methods.
22956 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22957 virtual methods before inlining is over.
22958 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22960 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22963 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22965 PR rtl-optimization/56005
22966 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22967 pending reads for prefetch.
22969 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22971 * config/aarch64/aarch64.md
22972 (*cstoresi_neg_uxtw): New pattern.
22973 (*cmovsi_insn_uxtw): New pattern.
22974 (*<optab>si3_uxtw): New pattern.
22975 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22976 (*<optab>si3_insn_uxtw): New pattern.
22977 (*bswapsi2_uxtw): New pattern.
22979 2013-01-16 Richard Biener <rguenther@suse.de>
22981 * tree-inline.c (tree_function_versioning): Remove set but
22982 never used variable.
22984 2013-01-16 Richard Biener <rguenther@suse.de>
22986 PR tree-optimization/55964
22987 * tree-flow.h (rename_variables_in_loop): Remove.
22988 (rename_variables_in_bb): Likewise.
22989 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22990 (copy_loop_before): Adjust and delete update-ssa status.
22991 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22992 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22993 (rename_variables_in_loop): Remove.
22994 (slpeel_update_phis_for_duplicate_loop): Likewise.
22995 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22996 use available cfg machinery instead of duplicating it.
22997 Update PHI nodes and perform poor-mans SSA update here.
22998 (slpeel_tree_peel_loop_to_edge): Adjust.
23000 2013-01-16 Richard Biener <rguenther@suse.de>
23002 PR tree-optimization/54767
23003 PR tree-optimization/53465
23004 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
23005 (vrp_visit_phi_node): For PHI arguments coming via backedges
23006 drop all symbolical range information.
23007 (execute_vrp): Compute backedges.
23009 2013-01-16 Richard Biener <rguenther@suse.de>
23011 * doc/install.texi: Update CLooG and ISL requirements to
23014 2013-01-16 Christian Bruel <christian.bruel@st.com>
23017 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23018 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23019 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23021 2013-01-16 DJ Delorie <dj@redhat.com>
23023 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23024 (UNSPECV_SP_SWITCH_E): New.
23025 (sp_switch_1): Change to an unspec.
23026 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23029 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23031 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23032 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23033 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23034 (expand_mem_thread_fence): Ditto.
23035 (expand_mem_signal_fence): Ditto.
23036 (expand_atomic_load): Ditto.
23037 (expand_atomic_store): Ditto.
23039 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23041 PR rtl-optimization/55547
23042 PR rtl-optimization/53827
23045 * alias.c (memrefs_conflict_p): Set sizes to negative after
23048 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23051 * function.c (thread_prologue_and_epilogue_insns): Always
23052 add crtl->drap_reg to set_up_by_prologue.set, even if
23053 stack_realign_drap is false.
23055 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23057 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23058 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23059 *call): Fix indention.
23061 2013-01-15 Tom de Vries <tom@codesourcery.com>
23064 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23067 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23069 PR rtl-optimization/55153
23070 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23072 2013-01-15 Martin Jambor <mjambor@suse.cz>
23074 PR tree-optimization/55920
23075 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23076 accesses as grp_to_be_debug_replaced.
23078 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23080 PR tree-optimization/55920
23081 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23082 there is non-useless type conversion needed from debug rhs to lhs,
23083 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23085 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23086 Mikael Pettersson <mikpe@it.uu.se>
23089 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23091 (ASM_OUTPUT_CASE_LABEL): Remove.
23092 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23093 * final.c (shorten_branches): Update alignment of labels before
23094 jump tables if CASE_VECTOR_SHORTEN_MODE.
23096 2013-01-15 Richard Biener <rguenther@suse.de>
23099 * system.h: Do not include gmp.h for building host tools.
23101 2013-01-15 Richard Biener <rguenther@suse.de>
23103 PR middle-end/55882
23104 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23105 account for bitpos when computing alignment.
23107 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23109 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23110 (ix86_target_macros_internal): Likewise.
23112 * config/i386/i386.c (m_CORE2I7): Removed.
23113 (m_CORE_HASWELL): New macro.
23114 (m_CORE_ALL): Likewise.
23115 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23116 (initial_ix86_arch_features): Likewise.
23117 (processor_target_table): Initializations for Core avx2.
23118 (cpu_names): New names "core-avx2".
23119 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23120 PROCESSOR_CORE_HASWELL.
23121 (ix86_issue_rate): New case.
23122 (ia32_multipass_dfa_lookahead): Likewise.
23123 (ix86_sched_init_global): Likewise.
23125 * config/i386/i386.h (TARGET_HASWELL): New macro.
23126 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
23127 (processor_type): New PROCESSOR_HASWELL.
23129 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23131 PR tree-optimization/55955
23132 * tree-vect-loop.c (vectorizable_reduction): Give up early on
23133 *SHIFT_EXPR and *ROTATE_EXPR codes.
23135 PR tree-optimization/48766
23136 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
23137 -ftrapv disable -fwrapv.
23139 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23142 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
23143 etc. to 1 and not to __flash.
23144 Use LL suffix for __INT24_MAX__ with -mint8.
23145 Use ULL suffix for __UINT24_MAX__ with -mint8.
23147 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23149 * config/avr/avr-arch.h
23150 (struct base_arch_s): Use typedef avr_arch_t instead.
23151 (struct arch_info_s): Use typedef avr_arch_info_t instead.
23152 (struct mcu_type_s): Use typedef avr_mcu_t instead.
23153 * config/avr/avr.c: Same.
23154 * config/avr/avr-devices.c: Same.
23155 * config/avr/driver-avr.c: Same.
23156 * config/avr/gen-avr-mmcu-texi.c: Same.
23157 * config/avr/avr-mcus.def: Adjust comment.
23159 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
23161 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
23162 * config/aarch64/iterators.md (VALLDI): New.
23164 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
23165 Andi Kleen <ak@linux.intel.com>
23168 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
23169 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
23172 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23174 * config/avr/avr-stdint.h: Remove trailing blanks.
23175 * config/avr/avr-log.h: Same.
23176 * config/avr/avr-arch.h: Same.
23177 * config/avr/avr-devices.c: Same.
23178 * config/avr/avr-dimode.md: Same.
23179 * config/avr/predicates.md: Same.
23180 * config/avr/avr-c.c: Same. And fix typo.
23182 * config/avr/avr-protos.h: Same. And:
23183 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
23184 (init_cumulative_args): Rename to avr_init_cumulative_args.
23185 (expand_prologue): Rename to avr_expand_prologue.
23186 (expand_epilogue): Rename to avr_expand_epilogue.
23187 (adjust_insn_length): Rename to avr_adjust_insn_length.
23188 (notice_update_cc): Rename to avr_notice_update_cc.
23189 (final_prescan_insn): Rename to avr_final_prescan_insn.
23190 * config/avr/avr.c: Same.
23191 * config/avr/avr.h: Same.
23192 * config/avr/avr.md: Remove trailing blanks.
23193 (prologue): Use avr_expand_prologue.
23194 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
23196 2013-01-14 Richard Biener <rguenther@suse.de>
23198 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
23199 verify_location, collect_subblocks): New functions.
23200 (verify_gimple_in_cfg): Verify that locations only reference
23201 BLOCKs in the functions BLOCK tree.
23203 2013-01-14 Richard Biener <rguenther@suse.de>
23205 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
23207 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
23209 (insert_out_of_ssa_copy_on_edge): Likewise.
23210 (rewrite_close_phi_out_of_ssa): Likewise.
23211 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
23213 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
23214 propagated constants.
23215 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
23218 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23220 * config/avr/avr-modes.def: Add GPL copyright notice.
23222 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
23224 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
23225 MEMMODEL_MASK to determine memory model.
23226 (atomic_store<mode>): Ditto from operands[2].
23227 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
23229 2013-01-13 Jakub Jelinek <jakub@redhat.com>
23232 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
23233 (fold_gimple_assign): Don't call unshare_expr here.
23234 (fold_ctor_reference): Call unshare_expr.
23236 2013-01-13 Terry Guo <terry.guo@arm.com>
23238 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
23239 * doc/fragments.texi: Document MULTILIB_REUSE.
23240 * gcc.c (multilib_reuse): New internal spec.
23241 (set_multilib_dir): Also search multilib from multilib_reuse.
23242 * genmultilib (tmpmultilib3): Refactor code.
23243 (tmpmultilib4): Ditto.
23244 (multilib_reuse): New multilib argument.
23246 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
23248 * Makefile.in: Update copyright.
23250 2013-01-12 Tom de Vries <tom@codesourcery.com>
23252 PR middle-end/55890
23253 * calls.c (expand_call): Check if arg_nr is valid.
23255 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23257 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23258 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23259 documentation. Add missing '__' in front of
23260 __builtin_ia32_packssdw256.
23262 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23265 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23266 NO_REGS for larl operands.
23267 (s390_reload_larl_operand): Use s390_load_address instead of
23270 2013-01-11 Richard Biener <rguenther@suse.de>
23272 * tree-cfg.c (verify_node_sharing_1): Split out from ...
23273 (verify_node_sharing): ... here.
23274 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
23276 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
23278 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
23279 Substitute TREECHECKING.
23280 * configure: Regenerate.
23281 * Makefile.in (TREECHECKING): New.
23283 2013-01-11 Richard Guenther <rguenther@suse.de>
23285 PR tree-optimization/44061
23286 * tree-vrp.c (extract_range_basic): Compute zero as
23287 value-range for __builtin_constant_p of function parameters.
23289 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
23291 Update copyright years.
23293 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
23295 PR rtl-optimization/55672
23296 * lra-eliminations.c (mark_not_eliminable): Permit addition with
23297 const to be eliminable.
23299 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
23301 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
23302 * configure: Regenerate.
23304 2013-01-10 Richard Biener <rguenther@suse.de>
23306 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
23308 2013-01-10 Richard Biener <rguenther@suse.de>
23311 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
23312 locations for virtual PHI arguments.
23313 (rewrite_update_phi_arguments): Likewise.
23315 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
23317 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
23320 2013-01-10 Jakub Jelinek <jakub@redhat.com>
23322 PR tree-optimization/55921
23323 * tree-complex.c (expand_complex_asm): New function.
23324 (expand_complex_operations_1): Call it for GIMPLE_ASM.
23326 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23329 * config/s390/s390.c (s390_symref_operand_p)
23330 (s390_loadrelative_operand_p): Merge the two functions.
23331 (s390_check_qrst_address, print_operand_address): Add parameters
23332 to s390_loadrelative_operand_p invokation.
23333 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
23334 (s390_reload_larl_operand, s390_secondary_reload): Use
23335 s390_loadrelative_operand_p instead of s390_symref_operand_p.
23336 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
23338 2013-01-09 Mike Stump <mikestump@comcast.net>
23340 * dse.c (record_store): Remove unnecessary assert.
23342 2013-01-09 Jan Hubicka <jh@suse.cz>
23344 PR tree-optimization/55569
23345 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
23346 * cfgloop.h (scale_loop_profile): Likewise.
23348 2013-01-09 Jan Hubicka <jh@suse.cz>
23351 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
23353 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
23355 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
23357 PR middle-end/55114
23358 * expr.h (maybe_emit_group_store): Declare.
23359 * expr.c (maybe_emit_group_store): New function.
23360 * builtins.c (expand_builtin_int_roundingfn): Call it.
23361 (expand_builtin_int_roundingfn_2): Likewise.
23363 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
23365 PR rtl-optimization/55829
23366 * lra-constraints.c (match_reload): Add code for absent output.
23367 (curr_insn_transform): Add code for reloads of matched inputs
23370 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
23372 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
23373 attribute of movddup insn to DF.
23374 (*vec_interleave_lowv2df): Ditto.
23375 (vec_dupv2df): Ditto.
23377 2013-01-09 Jan Hubicka <jh@suse.cz>
23379 PR tree-optimiation/55875
23380 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
23381 EVERY_ITERATION parameter.
23382 (number_of_iterations_exit): Check if exit is executed every iteration.
23383 (idx_infer_loop_bounds): Similarly here.
23384 (n_of_executions_at_most): Simplify
23385 to only test for cases where statement is dominated by the
23386 particular bound; handle correctly the "postdominance" test.
23387 (scev_probably_wraps_p): Use max loop iterations info
23388 as a global bound first.
23390 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
23391 Nick Clifton <nickc@redhat.com>
23393 * config/v850/v850.md (cbranchsf4): New pattern.
23394 (cstoresf4): New pattern.
23395 (cbranchdf4): New pattern.
23396 (cstoredf4): New pattern.
23397 (movsicc): Disallow floating point comparisons.
23398 (cmpsf_le_insn): Fix order of operators.
23399 (cmpsf_lt_insn): Likewise.
23400 (cmpsf_eq_insn): Likewise.
23401 (cmpdf_le_insn): Likewise.
23402 (cmpdf_lt_insn): Likewise.
23403 (cmpdf_eq_insn): Likewise.
23404 (cmpsf_ge_insn): Use LE comparison.
23405 (cmpdf_ge_insn): Likewise.
23406 (cmpsf_gt_insn): Use LT comparison.
23407 (cmpdf_gt_insn): Likewise.
23408 (cmpsf_ne_insn): Delete pattern.
23409 (cmpdf_ne_insn): Delete pattern.
23410 * config/v850/v850.c (v850_gen_float_compare): Use
23411 gen_cmpdf_eq_insn for NE comparison.
23412 (v850_float_z_comparison_operator)
23413 (v850_float_nz_comparison_operator): Move from here ...
23414 * config/v850/predicates.md: ... to here. Move GT and GE
23415 comparisons into v850_float_z_comparison_operator.
23416 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
23418 (v850_float_nz_comparison_operator): Likewise.
23420 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23422 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
23423 with calls to gen_insvsi/gen_insvdi.
23425 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23427 * config/i386/i386.c (initial_ix86_tune_features): Set up
23428 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
23430 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
23431 Jakub Jelinek <jakub@redhat.com>
23433 PR tree-optimization/48189
23434 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
23435 If nitercst is 0, don't predict the exit edge.
23437 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23439 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
23440 in asm_fprintf with reg_names.
23441 (aarch64_print_operand_address): Likewise.
23442 (aarch64_return_addr): Likewise.
23443 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
23445 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23447 * config/pa/pa.h (VAL_U6_BITS_P): Define.
23448 (INT_U6_BITS): Likewise.
23449 * config/pa/predicates.md (uint6_operand): New predicate.
23450 (shift5_operand, shift6_operand): Likewise.
23451 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
23453 (lshrdi3): Use shift6_operand.
23454 (shrpsi4, shrpdi4): New insn patterns.
23455 (extzv): Delete expander.
23456 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
23457 predicates in unamed zero extract patterns. Tighten common constraint.
23458 (extv): Delete expander.
23459 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
23460 predicates in unamed sign extract patterns. Tighten common constraint.
23461 (insv): Delete expander.
23462 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
23463 predicates in unamed insert patterns. Tighten common constraint.
23464 Change uint32_operand predicate to uint6_operand predicate in unamed
23465 DImode pattern to insert constant values of type 1...1xxxx.
23467 2013-01-04 Jan Hubicka <jh@suse.cz>
23469 PR tree-optimization/55823
23470 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
23473 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23474 Uros Bizjak <ubizjak@gmail.com>
23476 PR rtl-optimization/55845
23477 * df-problems.c (can_move_insns_across): Stop scanning at
23478 volatile_insn_p source instruction or give up if
23479 across_from .. across_to range contains any volatile_insn_p
23482 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23484 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
23485 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
23487 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
23488 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
23490 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23493 * asan.c (asan_clear_shadow): New function.
23494 (asan_emit_stack_protection): Use it.
23496 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23498 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
23499 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
23500 with tab instead of space.
23502 2013-01-08 Nick Clifton <nickc@redhat.com>
23504 * config/rl78/rl78.c (rl78_expand_prologue): Always select
23505 register bank 0 at the start of an interrupt handler.
23506 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
23509 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23511 * config/aarch64/aarch64-simd.md
23512 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
23513 (aarch64_simd_bsl): Likewise.
23514 (aarch64_vcond_internal<mode>): Likewise.
23515 (vcond<mode><mode>): Likewise.
23516 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
23517 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
23519 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23521 * config/aarch64/aarch64-builtins.c
23522 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
23524 2013-01-08 Martin Jambor <mjambor@suse.cz>
23527 * tree-sra.c (analyze_access_subtree): Return true also after
23528 potentially creating a debug-only replacement.
23530 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23532 PR middle-end/55890
23533 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
23535 PR tree-optimization/54120
23536 * tree-vrp.c (range_fits_type_p): Don't allow
23537 src_precision < precision from signed vr to unsigned_p
23538 if vr->min or vr->max is negative.
23539 (simplify_float_conversion_using_ranges): Test can_float_p
23540 against CODE_FOR_nothing.
23542 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23543 Richard Biener <rguenther@suse.de>
23545 PR middle-end/55851
23546 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23547 types instead of just INTEGER_TYPE types.
23549 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23551 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23554 2013-01-07 Steve Ellcey <sellcey@mips.com>
23557 * config/mips/mips.opt: Change mad to mmad to match documentation.
23559 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23562 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23563 .progmemx.data now.
23565 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23568 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23569 (avr_addrspace_t): Add .section_name field.
23570 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23572 (avr_addrspace): Same. Initialize .section_name. Remove last
23573 NULL entry. Put __memx into .progmemx.data.
23574 (progmem_section_prefix): Remove.
23575 (avr_asm_init_sections): No need to initialize progmem_section.
23576 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23577 section name prefix.
23578 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23579 retrieve the progmem section.
23580 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23581 boundary to run over avr_addrspace[].
23582 (avr_register_target_pragmas): Ditto.
23584 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23586 * varasm.c (output_constant_def_contents): For asan_protect_global
23587 protected strings, adjust DECL_ALIGN if needed, before testing for
23589 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23590 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23592 (output_object_block): For asan protected decls, emit asan padding
23593 after their contents.
23594 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23595 (asan_finish_file): Test it here instead.
23597 2013-01-07 Nick Clifton <nickc@redhat.com>
23598 Matthias Klose <doko@debian.org>
23599 Doug Kwan <dougkwan@google.com>
23600 H.J. Lu <hongjiu.lu@intel.com>
23603 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23605 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23607 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23609 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23611 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23613 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23616 * doc/install.texi (Cross-Compiler-Specific Options): Document
23619 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23621 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23622 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23623 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23624 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23625 vqmovun_high_s64): Fix source operand number and update copyright.
23627 2013-01-07 Richard Biener <rguenther@suse.de>
23629 PR middle-end/55890
23630 * gimple.h (gimple_call_builtin_p): New overload.
23631 * gimple.c (validate_call): New function.
23632 (gimple_call_builtin_p): Likewise.
23633 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23634 Use gimple_call_builtin_p.
23635 (find_func_clobbers): Likewise.
23636 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23637 (strlen_optimize_stmt): Likewise.
23639 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23641 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23642 (vld1q_dup_*): Likewise.
23643 (vld1_*): Likewise.
23644 (vld1q_*): Likewise.
23645 (vld1_lane_*): Likewise.
23646 (vld1q_lane_*): Likewise.
23648 2013-01-07 Richard Biener <rguenther@suse.de>
23650 * lto-streamer.h (LTO_minor_version): Bump to 2.
23652 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23654 * config/aarch64/aarch64-protos.h
23655 (aarch64_const_double_zero_rtx_p): Rename to...
23656 (aarch64_float_const_zero_rtx_p): ...this.
23657 (aarch64_float_const_representable_p): New.
23658 (aarch64_output_simd_mov_immediate): Likewise.
23659 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23660 move immediate case.
23661 * config/aarch64/aarch64.c
23662 (aarch64_const_double_zero_rtx_p): Rename to...
23663 (aarch64_float_const_zero_rtx_p): ...this.
23664 (aarch64_print_operand): Allow printing of new constants.
23665 (aarch64_valid_floating_const): New.
23666 (aarch64_legitimate_constant_p): Check for valid floating-point
23668 (aarch64_simd_valid_immediate): Likewise.
23669 (aarch64_vect_float_const_representable_p): New.
23670 (aarch64_float_const_representable_p): Likewise.
23671 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23672 (aarch64_output_simd_mov_immediate): New.
23673 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23674 (*movdf_aarch64): Likewise.
23675 * config/aarch64/constraints.md (Ufc): New.
23676 (Y): call aarch64_float_const_zero_rtx.
23677 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23679 2013-01-07 Richard Biener <rguenther@suse.de>
23681 PR tree-optimization/55888
23682 PR tree-optimization/55862
23683 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23684 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23685 not if it is contained therein.
23687 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23689 * config/avr/t-avr: Typo.
23691 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23694 * config/avr/t-avr: Don't automatically rebuild
23695 $(srcdir)/config/avr/t-multilib
23696 $(srcdir)/config/avr/avr-tables.opt
23697 $(srcdir)/doc/avr-mmcu.texi
23698 (avr-mcus): New phony target to build them on request.
23699 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23700 * avr/avr-mcus.def: Adjust comments.
23702 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23704 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23706 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23708 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23710 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23712 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23714 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23716 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23717 to generate profiling.
23718 * config/rs6000/aix64.h (LIB_SPEC): Same.
23720 2013-01-04 Andrew Pinski <apinski@cavium.com>
23722 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23724 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23726 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23728 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23730 (ix86_expand_move): Ditto.
23731 (ix86_zero_extend_to_Pmode): Ditto.
23732 (ix86_expand_call): Ditto.
23733 (ix86_expand_special_args_builtin): Ditto.
23734 (ix86_expand_builtin): Ditto.
23736 2013-01-04 Richard Biener <rguenther@suse.de>
23738 PR tree-optimization/55862
23739 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23740 translating them through PHI nodes.
23742 2013-01-04 Martin Jambor <mjambor@suse.cz>
23744 PR tree-optimization/55755
23745 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23746 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23748 2013-01-04 Marek Polacek <polacek@redhat.com>
23750 PR middle-end/55859
23751 * opts.c (default_options_optimization): Clarify error message.
23753 2013-01-04 Richard Biener <rguenther@suse.de>
23755 PR middle-end/55863
23756 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23759 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23762 * config/pa/pa.md (movsi): Revert previous change.
23763 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23766 2013-01-03 Richard Henderson <rth@redhat.com>
23768 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23769 after eliminating TLS symbols.
23771 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23774 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23775 * graphite-poly.c (debug_gmp_value): Likewise.
23777 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23780 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23781 selected code model, define __code_mode_small__, __code_model_medium__,
23782 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23783 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23784 xchg temporary register with %k. Declare temporary register as
23786 [__x86_64__]: For medium and large code models, preserve %rbx register.
23788 2013-01-03 Richard Biener <rguenther@suse.de>
23790 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23791 (dump_subscript): Adjust.
23792 (finalize_ddr_dependent): Do not dump redundant info.
23793 (analyze_siv_subscript): Adjust.
23794 (subscript_dependence_tester): Likewise.
23795 (compute_affine_dependence): Likewise.
23797 2013-01-03 Richard Biener <rguenther@suse.de>
23800 2013-01-03 Richard Biener <rguenther@suse.de>
23802 PR tree-optimization/55857
23803 * tree-vect-stmts.c (vectorizable_load): Do not setup
23804 re-alignment for invariant loads.
23806 2013-01-02 Richard Biener <rguenther@suse.de>
23808 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23809 invariant load do not generate a vector load from the scalar location.
23811 2013-01-03 Richard Biener <rguenther@suse.de>
23813 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23814 for not vectorizing.
23815 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23816 not build INDIRECT_REFs, call get_name once only.
23817 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23818 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23820 2013-01-03 Richard Biener <rguenther@suse.de>
23822 PR tree-optimization/55857
23823 * tree-vect-stmts.c (vectorizable_load): Do not setup
23824 re-alignment for invariant loads.
23826 2013-01-03 Richard Biener <rguenther@suse.de>
23829 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23830 prefer a built-in decl.
23832 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23834 * gcc.c (process_command): Update copyright notice dates.
23835 * gcov.c (print_version): Likewise.
23836 * gcov-dump.c (print_version): Likewise.
23838 PR rtl-optimization/55838
23839 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23840 iv0.step, iv1.step and step.
23842 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23843 Marc Glisse <marc.glisse@inria.fr>
23845 PR tree-optimization/55832
23846 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23847 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23848 integer_{one,zero}_node.
23850 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23853 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23854 * var-tracking.c (reverse_op): Don't add reverse ops to
23855 VALUEs that have already
23856 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23858 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23860 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23862 2013-01-02 Teresa Johnson <tejohnson@google.com>
23864 * dumpfile.c (dump_loc): Print filename with location.
23865 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23866 new location_t parameter to emit complete unroll message with
23867 new dump framework.
23868 (canonicalize_loop_induction_variables): Compute loops location
23869 and pass to try_unroll_loop_completely.
23870 * loop-unroll.c (report_unroll_peel): New function.
23871 (peel_loops_completely): Use new dump format with location
23872 for main dumpfile message, and invoke report_unroll_peel on success.
23873 (decide_unrolling_and_peeling): Ditto.
23874 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23875 by report_unroll_peel.
23876 (decide_peel_completely): Ditto.
23877 (decide_unroll_constant_iterations): Ditto.
23878 (decide_unroll_runtime_iterations): Ditto.
23879 (decide_peel_simple): Ditto.
23880 (decide_unroll_stupid): Ditto.
23881 * cfgloop.c (get_loop_location): New function.
23882 * cfgloop.h (get_loop_location): Declare.
23884 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23886 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23889 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23891 PR middle-end/55198
23892 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23893 BLKmode objects when EXPAND_MEMORY is specified.
23895 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23897 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23899 (fold_builtin_cpu): Do not share cpu model decls across statements.
23901 2013-01-02 Jason Merrill <jason@redhat.com>
23904 * tree.c (build_array_type_1): Revert earlier change.
23906 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23908 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23910 * config/aarch64/aarch64-tune.md: Re-generate.
23912 2013-01-02 Richard Biener <rguenther@suse.de>
23914 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23915 invariant load do not generate a vector load from the scalar location.
23917 2013-01-02 Richard Biener <rguenther@suse.de>
23920 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23921 * configure: Regenerate.
23923 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23925 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23926 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23927 (expand_builtin_int_roundingfn_2): Keep the original target around
23928 for the fallback case.
23930 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23932 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23933 to be clear for sign changes.
23935 2013-01-01 Jan Hubicka <jh@suse.cz>
23937 * ipa-inline-analysis.c: Fix formatting.
23939 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23941 PR tree-optimization/55831
23942 * tree-vect-loop.c (get_initial_def_for_induction): Use
23943 gsi_after_labels instead of gsi_start_bb.
23945 Copyright (C) 2013 Free Software Foundation, Inc.
23947 Copying and distribution of this file, with or without modification,
23948 are permitted in any medium without royalty provided the copyright
23949 notice and this notice are preserved.